You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@openoffice.apache.org by Dean Webber <de...@student.uq.edu.au> on 2026/01/01 01:07:13 UTC

Re: WG: There is only one problem in OpenOffice....

Dear Wolfgang,

I am not a developer, but had the same file issue when I upgraded to Debian 13.

One of the developers suggested I use the releases here, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/ , which is an upstream build which I suspect has a fix for this issue.

Specifically for you, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_install-rpm_de.tar.gz and https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_langpack-rpm_de.tar.gz

It worked for me, perhaps it will work for you also.

Happy New Year.

Faithfully,
Dean

Translation en -> german

Lieber Wolfgang,

Ich bin kein Entwickler, hatte aber beim Upgrade auf Debian 13 das gleiche Dateiproblem.

Einer der Entwickler schlug vor, die Releases hier zu verwenden, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de / , Das ist ein Upstream-Build, von dem ich vermute, dass es eine Lösung für dieses Problem gibt.

Speziell für Sie, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_install-rpm_de.tar.gz und https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_langpack-rpm_de.tar.gz

Es hat bei mir funktioniert, vielleicht funktioniert es auch bei dir.

Frohes neues Jahr.

Treu,
Dean

________________________________
From: Wolfgang Hinzmann <Wo...@protonmail.ch>
Sent: 31 December 2025 02:12
To: dev@ <de...@openoffice.apache.org>
Subject: Re: WG: There is only one problem in OpenOffice....

Hello Marcus,

many thanks for these hints again.

I tried both:

As first step I renamed the folder

$HOME/.openoffice.org/4/user/basic"

After that, I started my computer new.
When I started now OpenOffice, I got several messages like
"Cannot load [file...].....".
Despite this fact, I was able to create a new textfile and could save it.
As usual with the "Error message".

Because of this errors I tried the second step:
I renamed the whole folder - as You described:
$HOME/.openoffice.org

Then I started my computer new,
and after that I started OpenOffice.
A window opened, and I was asked to put my whole Name, First name and family name,
and an abbreviation for it. By clicking then the "Coninue-Button"
OpenOffice opened, and I had the choice between the tools like Writer, Calculation...
I opened writer, opened a new document, and then saving it:
"The error message appears again" - as I described it.

In effect, both steps - renaming the basic-folder, and after that,
renaming the "$HOME/.openoffice.org"- folder,
did not change the error away...

So I decided, to change these system-folders in its originial state
by renaming it, and deleting the new created ones.

If You are interested in the error-messages, I have got,
I would repeat the steps.

I had really very great hope. Especially with the second step,
because it started with the "Welcome-Window", and asking me
for my name and its abbreviation.
Strange. Very strange.

Despite of this:
I wish You a very good new year 2026!

Best wishes Wolfgang

Gesendet mit [Proton Mail](https://proton.me/mail/home): Ein sicherer E-Mail-Dienst.

Marcus marcus.mail@wtnet.de schrieb am Montag, 29. Dezember 2025 um 16:15:

> Hi Wolfgang,
>
> first of all you can try to reset your Basic settings. The following
> will reset only the Basic things (e.g., macros) but nothing more:
>
> - close all AOO documents and windows
> - open a terminal
> - rename the Basic directory "$HOME/.openoffice.org/4/user/basic"
> - cd ~
> - mv .openoffice.org/4/user/basic
> .openoffice.org/4/user/basic_maybe_broken
> - start AOO new
> - try if the error message is still showing
>
> If this will not help, try to reset the complete AOO profile:
>
> - close all AOO documents and windows
> - open a terminal
> - rename the complete profile directory "$HOME/.openoffice.org"
> - cd ~
> - mv .openoffice.org .openoffice.org_maybe_broken
> - start AOO new
> - try if the error message is still showing
>
> Please note:
> You should know that all your customized settings will be gone. They all
> need to be setup again.
>
> Please be carefully:
> Rename the profile directory. You can delete it anytime later when it is
> sure that the profile was somehow broken.
> Otherwise if the problem persists you can decide to go back to the old
> profile directory by renaming it back with the following commands in the
> terminal:
>
> cd ~
> rm -rf .openoffice.org
> mv .openoffice.org_maybe_broken .openoffice.org
>
> HTH
>
> Marcus
>
> Am 26.12.25 um 10:40 schrieb Wolfgang Hinzmann:
>
>> as You suggested, I will try OpenOffice, especially Writer,
>> with a new profile.
>> Because I never needed special settings and other useful extensions so far.
>> The only question for me in that wayis:
>> How can I realize that?
>> If You get other hints too, I would be very glad.
>>
>> Marcus marcus.mail@wtnet.de schrieb am Dienstag, 23. Dezember 2025 um 13:45:
>>
>>> Am 21.12.25 um 18:09 schrieb Wolfgang Hinzmann:
>>>
>>>> /as You suggested, I opened the terminal, and /
>>>> /typed then the command "soffice"./
>>>> /After that, the main-menu of OpenOffice started./
>>>> /There, I go to "Open last files", and opened the "Testfile01.odt"./
>>>> /After that I got the message:/
>>>>
>>>> /image.png/
>>>> /
>>>> /In the_terminal-window_, there are the following messages:/
>>>> /wolfganghinzmann@fedora:~$ soffice/
>>>> /javaldx: Could not find a Java Runtime Environment!/
>>>> /
>>>> /That's all./
>>>
>>> than I (you) need help from others as I've no idea at the moment what
>>> could be checked next.
>>>
>>> @All:
>>> Has anyone more hints? Thanks.
>>>
>>> Last resort:
>>> There is always the possibility to start from scratch with a new AOO
>>> profile. But that means you would loose all your settings that you have
>>> done up to know (customized toolbars, menus, keyboard shortcuts, see
>>> menu "Tools - Customize" or in German: "Extras - Anpassen" plus some
>>> more settings in other locations, e.g., "Format - AutoCorrext -
>>> AutoCorrect Optionen").
>>>
>>> So, it depends on you if you want to try it.
>>>
>>>> Marcus marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de schrieb am
>>>> Sonntag, 21. Dezember 2025 um 14:44:
>>>>
>>>> Am 21.12.25 um 14:12 schrieb Wolfgang Hinzmann:
>>>>
>>>> /I don't use macros in my documents./
>>>>
>>>> OK, a possible root cause less.
>>>>
>>>> /Question:/
>>>> /You suggested, I should try to start Writer from Terminal./
>>>> /How can I do this? I am really not so familiar with all these
>>>> commands./
>>>>
>>>> that should be simple. You just open the terminal and enter the
>>>> following command:
>>>>
>>>> soffice
>>>>
>>>> If this results in an unknown command, then please use it with the full
>>>> path. It depends where you have installed it in your system. E.g.,
>>>>
>>>> /opt/openoffice4/program/soffice
>>>>
>>>> The output until the error message appears could be helpful. Fore this
>>>> purpose start OpenOffice and then open a document that shows that error
>>>> message. After that please post the complete text here.
>>>>
>>>> Thanks
>>>>
>>>> Marcus
>>>>
>>>> Gesendet mit Proton Mail https://proton.me/mail/home: <https://
>>>> proton.me/mail/home:> Ein sicherer E-
>>>> Mail-Dienst.
>>>>
>>>> Marcus marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de
>>>> mailto:marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de schrieb am
>>>> Sonntag, 21. Dezember 2025 um 10:05:
>>>>
>>>> Am 20.12.25 um 16:32 schrieb Wolfgang Hinzmann:
>>>>
>>>> great that it is now installing and working.
>>>>
>>>> /There is only one problem with OpenOffice, I had already in the/
>>>> /earlier version too:/
>>>> /By opening an odt-file with Writer I receive always the error
>>>> message/
>>>> / "Allgemeiner Fehler"./
>>>> /image.png
>>>> /
>>>> /I had the hope, that this message will disappear by the
>>>> installation of
>>>> the new version./
>>>> /But it doesn't./
>>>> /
>>>> /But what is the solution for it?/
>>>> /After pressing the button "OK" I am able to work normally with
>>>> the file
>>>> I opened:/
>>>> /Changing it and saving it./
>>>>
>>>> do you have macros inside the document? Please have a look via
>>>> the menu
>>>> "Tools - Macros - Organize Macros - Basic" (German: "Extras -
>>>> Makros -
>>>> Makros verwalten - Basic")?
>>>> If so, it's up to you to decide if you still need them - then
>>>> you have
>>>> to adjust parts to run correctly. Then there is good chance that the
>>>> document will open without error messages. Or to delete them.
>>>>
>>>> Please make sure you have made a backup before deleting things
>>>> that you
>>>> still need.

Re: SELinux [was: WG: There is only one problem in OpenOffice....

Posted by Pe...@apache.org.
Hi Arrigo,

Am 12.06.2026 um 22:02 schrieb Arrigo Marchiori:
> Hello Damjan, All,
>
> On Mon, Jun 08, 2026 at 10:25:07PM +0200, Arrigo Marchiori wrote:
>
>> Dear All,
>>
>> On Mon, Jun 08, 2026 at 04:54:21AM +0200, Damjan Jovanovic wrote:
>>
> [...]
>>> Let's understand this better first.
>>>
>>> This is an excellent blog post on the topic:
>>> https://linuxvox.com/blog/stacks-are-executable-even-with-noexecstack/
>>>
>>> Now as per that article, note how both FreeBSD and Linux already link with
>>> -Wl,-z,noexecstack for libraries, at least with gbuild:
>>>
>>> main/solenv]$ grep execstack * -R
>>> gbuild/platform/freebsd.mk:gb_Library_TARGETTYPEFLAGS := -shared
>>> -Wl,-z,noexecstack
>>> gbuild/platform/linux.mk:gb_Library_TARGETTYPEFLAGS := -shared
>>> -Wl,-z,noexecstack
>>>
>>> but even this seems unnecessary, as that should be the default with GCC >=
>>> 4.1.
>>>
>>> A quick audit on my build tree
>>> (main/instsetoo_native/unxfbsdx/Apache_OpenOffice/installed/install/en-US/openoffice4/program)
>>> shows how nothing has executable stacks:
>>>
>>> find . -type f | while read i; do readelf -l "$i" 2>/dev/null| grep
>>> GNU_STACK|grep E; done
>>> (no output. You can remove the "grep E" to see every GNU_STACK entry has
>>> only "RW" permissions.)
>> I had to add option --wide to readelf(1).
>>
>> Here is an interesting thing:
>>
>>   - the solver directory from my build of AOO41X with openSUSE 16.0,
>>     GCC 13.4.0, has no files with executable stack.
>>
>>   - the solver directory from my build of AOO41X with CentOS 5, GCC 4.1.2,
>>     has the following file with executable stack:
>>      ./4116/unxlngx6.pro/lib/libraptor2.so.0
>>
>> So I was probably addressing the wrong problem: libraptor2.so.0 shall
>> not require executable stack, instead.
> PR #483 now tries to explicitly disable executable stack when
> compiling raptor.
>
> How can we test if AOO is still working as expected?

I think it is used during save or load process for ODF files.

Loading and saving should trigger raptor use.

All the best

Peter

>
> Best regards,

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org


Re: SELinux [was: WG: There is only one problem in OpenOffice....

Posted by Arrigo Marchiori <ar...@apache.org>.
Hello Damjan, All,

On Mon, Jun 08, 2026 at 10:25:07PM +0200, Arrigo Marchiori wrote:

> Dear All,
> 
> On Mon, Jun 08, 2026 at 04:54:21AM +0200, Damjan Jovanovic wrote:
> 
[...]
> > Let's understand this better first.
> > 
> > This is an excellent blog post on the topic:
> > https://linuxvox.com/blog/stacks-are-executable-even-with-noexecstack/
> > 
> > Now as per that article, note how both FreeBSD and Linux already link with
> > -Wl,-z,noexecstack for libraries, at least with gbuild:
> > 
> > main/solenv]$ grep execstack * -R
> > gbuild/platform/freebsd.mk:gb_Library_TARGETTYPEFLAGS := -shared
> > -Wl,-z,noexecstack
> > gbuild/platform/linux.mk:gb_Library_TARGETTYPEFLAGS := -shared
> > -Wl,-z,noexecstack
> > 
> > but even this seems unnecessary, as that should be the default with GCC >=
> > 4.1.
> > 
> > A quick audit on my build tree
> > (main/instsetoo_native/unxfbsdx/Apache_OpenOffice/installed/install/en-US/openoffice4/program)
> > shows how nothing has executable stacks:
> > 
> > find . -type f | while read i; do readelf -l "$i" 2>/dev/null| grep
> > GNU_STACK|grep E; done
> > (no output. You can remove the "grep E" to see every GNU_STACK entry has
> > only "RW" permissions.)
> 
> I had to add option --wide to readelf(1).
> 
> Here is an interesting thing:
> 
>  - the solver directory from my build of AOO41X with openSUSE 16.0,
>    GCC 13.4.0, has no files with executable stack.
> 
>  - the solver directory from my build of AOO41X with CentOS 5, GCC 4.1.2,
>    has the following file with executable stack:
>     ./4116/unxlngx6.pro/lib/libraptor2.so.0
> 
> So I was probably addressing the wrong problem: libraptor2.so.0 shall
> not require executable stack, instead.

PR #483 now tries to explicitly disable executable stack when
compiling raptor.

How can we test if AOO is still working as expected?

Best regards,
-- 
Arrigo

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org


Re: SELinux [was: WG: There is only one problem in OpenOffice....

Posted by Arrigo Marchiori <ar...@apache.org>.
Dear All,

On Mon, Jun 08, 2026 at 04:54:21AM +0200, Damjan Jovanovic wrote:

> On Sun, Jun 7, 2026 at 10:43 PM Arrigo Marchiori <ar...@apache.org> wrote:
> 
> > Hello Damjan, All,
> >
> > On Sun, Jun 07, 2026 at 10:00:37PM +0200, Damjan Jovanovic wrote:
> >
> > > Could this be the problem? I think it must be, OpenGrok shows there's
> > > nowhere else in OpenOffice we use PROT_EXEC.
> > >
> > > At least on lines 105 and 111 it's granting read, write and execute
> > > permissions together:
> > >
> > > File main/bridges/source/cpp_uno/shared/vtablefactory.cxx:
> > >
> > >      77 extern "C" void * SAL_CALL allocExec(rtl_arena_type *, sal_Size *
> > > size) {
> > > ...
> > >      96     sal_Size n = (*size + (pagesize - 1)) & ~(pagesize - 1);
> > >      97     void * p;
> > >      98 #if defined SAL_UNX
> > >      99     p = mmap(
> > >     100         0, n, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANON, -1,
> > >     101         0);
> > >     102     if (p == MAP_FAILED) {
> > >     103         p = 0;
> > >     104     }
> > >     105     else if (mprotect (static_cast<char*>(p), n, PROT_READ |
> > > PROT_WRITE | PROT_EXEC) == -1)
> > >     106     {
> > >     107         munmap (static_cast<char*>(p), n);
> > >     108         p = 0;
> > >     109     }
> > >     110 #elif defined SAL_W32
> > >     111     p = VirtualAlloc(0, n, MEM_COMMIT, PAGE_EXECUTE_READWRITE);
> > >     112 #elif defined(SAL_OS2)
> > >     113     p = 0;
> > >     114     DosAllocMem( &p, n, PAG_COMMIT | PAG_READ | PAG_WRITE |
> > > OBJ_ANY);
> > >     115 #endif
> > >     116     if (p != 0) {
> > >     117         *size = n;
> > >     118     }
> > >     119     return p;
> > >     120 }
> > >
> > > There are already some workarounds for SELinux in this directory, but we
> > > probably need more.
> >
> > I don't know SELinux well, nor those syscalls. So I cannot confirm,
> > unfortunately.
> >
> > Do you think we should put a ``workaround for SELinux'' in place,
> > rather than running execstack as I was suggesting in PR #483?  If so,
> > could you please point me to some documentation? I have access both to
> > an ``old'' build VM and a ``new'' Linux-based system, so I can easily
> > do tests.
> >
> >
> Let's understand this better first.
> 
> This is an excellent blog post on the topic:
> https://linuxvox.com/blog/stacks-are-executable-even-with-noexecstack/
> 
> Now as per that article, note how both FreeBSD and Linux already link with
> -Wl,-z,noexecstack for libraries, at least with gbuild:
> 
> main/solenv]$ grep execstack * -R
> gbuild/platform/freebsd.mk:gb_Library_TARGETTYPEFLAGS := -shared
> -Wl,-z,noexecstack
> gbuild/platform/linux.mk:gb_Library_TARGETTYPEFLAGS := -shared
> -Wl,-z,noexecstack
> 
> but even this seems unnecessary, as that should be the default with GCC >=
> 4.1.
> 
> A quick audit on my build tree
> (main/instsetoo_native/unxfbsdx/Apache_OpenOffice/installed/install/en-US/openoffice4/program)
> shows how nothing has executable stacks:
> 
> find . -type f | while read i; do readelf -l "$i" 2>/dev/null| grep
> GNU_STACK|grep E; done
> (no output. You can remove the "grep E" to see every GNU_STACK entry has
> only "RW" permissions.)

I had to add option --wide to readelf(1).

Here is an interesting thing:

 - the solver directory from my build of AOO41X with openSUSE 16.0,
   GCC 13.4.0, has no files with executable stack.

 - the solver directory from my build of AOO41X with CentOS 5, GCC 4.1.2,
   has the following file with executable stack:
    ./4116/unxlngx6.pro/lib/libraptor2.so.0

So I was probably addressing the wrong problem: libraptor2.so.0 shall
not require executable stack, instead.

I am having a hard time understanding why this is happening, though.

Best regards,
-- 
Arrigo

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org


Re: SELinux [was: WG: There is only one problem in OpenOffice....

Posted by Pe...@apache.org.
Hi Arrigo, all,

Not every Distro is using SE Linux.

Distro 	MAC System 	SELinux?
*RHEL/CentOS/AlmaLinux/Rocky* 	SELinux 	✅ Default enforcing
*Fedora* 	SELinux 	✅ Default enforcing
*Debian* 	AppArmor 	❌
*Ubuntu* 	AppArmor 	❌
*openSUSE/SLES* 	AppArmor 	❌
*Arch Linux* 	None default 	❌
*Gentoo* 	Optional 	❌ by default
*Alpine* 	None 	❌
*Android* 	SELinux 	✅


Am 07.06.2026 um 22:43 schrieb Arrigo Marchiori:
> Hello Damjan, All,
>
> On Sun, Jun 07, 2026 at 10:00:37PM +0200, Damjan Jovanovic wrote:
>
>> Could this be the problem? I think it must be, OpenGrok shows there's
>> nowhere else in OpenOffice we use PROT_EXEC.
>>
>> At least on lines 105 and 111 it's granting read, write and execute
>> permissions together:
>>
>> File main/bridges/source/cpp_uno/shared/vtablefactory.cxx:
>>
>>       77 extern "C" void * SAL_CALL allocExec(rtl_arena_type *, sal_Size *
>> size) {
>> ...
>>       96     sal_Size n = (*size + (pagesize - 1)) & ~(pagesize - 1);
>>       97     void * p;
>>       98 #if defined SAL_UNX
>>       99     p = mmap(
>>      100         0, n, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANON, -1,
>>      101         0);
>>      102     if (p == MAP_FAILED) {
>>      103         p = 0;
>>      104     }
>>      105     else if (mprotect (static_cast<char*>(p), n, PROT_READ |
>> PROT_WRITE | PROT_EXEC) == -1)
>>      106     {
>>      107         munmap (static_cast<char*>(p), n);
>>      108         p = 0;
>>      109     }
>>      110 #elif defined SAL_W32
>>      111     p = VirtualAlloc(0, n, MEM_COMMIT, PAGE_EXECUTE_READWRITE);
>>      112 #elif defined(SAL_OS2)
>>      113     p = 0;
>>      114     DosAllocMem( &p, n, PAG_COMMIT | PAG_READ | PAG_WRITE |
>> OBJ_ANY);
>>      115 #endif
>>      116     if (p != 0) {
>>      117         *size = n;
>>      118     }
>>      119     return p;
>>      120 }
>>
>> There are already some workarounds for SELinux in this directory, but we
>> probably need more.
> I don't know SELinux well, nor those syscalls. So I cannot confirm,
> unfortunately.
>
> Do you think we should put a ``workaround for SELinux'' in place,
> rather than running execstack as I was suggesting in PR #483?  If so,
> could you please point me to some documentation? I have access both to
> an ``old'' build VM and a ``new'' Linux-based system, so I can easily
> do tests.

We could add policies for SELinux to exclude certain parts from the 
protection.

the recommendation is actually to remove execstack.

|# Add to assembler flags in build system -Wa,--noexecstack # Add to 
linker flags -Wl,-z,noexecstack|

however we might still end up with issues where we need then a policy. 
We likely hit issues with:

Component 	Risk 	Reason
*UNO bridge / libffi* 	🔴 High 	Trampolines
*Basic macro engine* 	🔴 High 	Runtime code generation
*Java integration (JNI)* 	🟡 Medium 	Depends on JVM
*Inline assembly* 	🟡 Medium 	Old platform-specific code
*Bundled libraries* 	🟡 Medium 	May have old compile flags


All the best

Peter


>
> Best regards,

Re: SELinux [was: WG: There is only one problem in OpenOffice....

Posted by Damjan Jovanovic <da...@apache.org>.
On Mon, Jun 8, 2026 at 8:03 AM <Pe...@apache.org> wrote:

>
> Am 08.06.2026 um 04:54 schrieb Damjan Jovanovic:
> > On Sun, Jun 7, 2026 at 10:43 PM Arrigo Marchiori<ar...@apache.org>
> wrote:
> >
> >> Hello Damjan, All,
> >>
> >> On Sun, Jun 07, 2026 at 10:00:37PM +0200, Damjan Jovanovic wrote:
> >>
> >>> Could this be the problem? I think it must be, OpenGrok shows there's
> >>> nowhere else in OpenOffice we use PROT_EXEC.
> >>>
> >>> At least on lines 105 and 111 it's granting read, write and execute
> >>> permissions together:
> >>>
> >>> File main/bridges/source/cpp_uno/shared/vtablefactory.cxx:
> >>>
> >>>       77 extern "C" void * SAL_CALL allocExec(rtl_arena_type *,
> sal_Size *
> >>> size) {
> >>> ...
> >>>       96     sal_Size n = (*size + (pagesize - 1)) & ~(pagesize - 1);
> >>>       97     void * p;
> >>>       98 #if defined SAL_UNX
> >>>       99     p = mmap(
> >>>      100         0, n, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANON,
> -1,
> >>>      101         0);
> >>>      102     if (p == MAP_FAILED) {
> >>>      103         p = 0;
> >>>      104     }
> >>>      105     else if (mprotect (static_cast<char*>(p), n, PROT_READ |
> >>> PROT_WRITE | PROT_EXEC) == -1)
> >>>      106     {
> >>>      107         munmap (static_cast<char*>(p), n);
> >>>      108         p = 0;
> >>>      109     }
> >>>      110 #elif defined SAL_W32
> >>>      111     p = VirtualAlloc(0, n, MEM_COMMIT,
> PAGE_EXECUTE_READWRITE);
> >>>      112 #elif defined(SAL_OS2)
> >>>      113     p = 0;
> >>>      114     DosAllocMem( &p, n, PAG_COMMIT | PAG_READ | PAG_WRITE |
> >>> OBJ_ANY);
> >>>      115 #endif
> >>>      116     if (p != 0) {
> >>>      117         *size = n;
> >>>      118     }
> >>>      119     return p;
> >>>      120 }
> >>>
> >>> There are already some workarounds for SELinux in this directory, but
> we
> >>> probably need more.
> >> I don't know SELinux well, nor those syscalls. So I cannot confirm,
> >> unfortunately.
> >>
> >> Do you think we should put a ``workaround for SELinux'' in place,
> >> rather than running execstack as I was suggesting in PR #483?  If so,
> >> could you please point me to some documentation? I have access both to
> >> an ``old'' build VM and a ``new'' Linux-based system, so I can easily
> >> do tests.
> >>
> >>
> > Let's understand this better first.
> >
> > This is an excellent blog post on the topic:
> > https://linuxvox.com/blog/stacks-are-executable-even-with-noexecstack/
> >
> > Now as per that article, note how both FreeBSD and Linux already link
> with
> > -Wl,-z,noexecstack for libraries, at least with gbuild:
> >
> > main/solenv]$ grep execstack * -R
> > gbuild/platform/freebsd.mk:gb_Library_TARGETTYPEFLAGS := -shared
> > -Wl,-z,noexecstack
> > gbuild/platform/linux.mk:gb_Library_TARGETTYPEFLAGS := -shared
> > -Wl,-z,noexecstack
> >
> > but even this seems unnecessary, as that should be the default with GCC
> >=
> > 4.1.
> >
> > A quick audit on my build tree
> >
> (main/instsetoo_native/unxfbsdx/Apache_OpenOffice/installed/install/en-US/openoffice4/program)
> > shows how nothing has executable stacks:
> >
> > find . -type f | while read i; do readelf -l "$i" 2>/dev/null| grep
> > GNU_STACK|grep E; done
> > (no output. You can remove the "grep E" to see every GNU_STACK entry has
> > only "RW" permissions.)
> >
> > It's of course possible that the stack is made executable at run-time,
> > because when libraries needing execstack are loaded, they poison the
> stack
> > of the entire application. Even this can be checked:
> >
> > procstat vm PID |grep rwx
> > (or on Linux, use "cat /proc/PID/map")
> >
> >    PID              START                END PRT  RES PRES REF SHD FLAG
> TP
> > PATH
> >   6460        0x89b9a5000        0x89bc15000 rwx  134  134   1   0 -----
> sw
> >   6460        0x89bf35000        0x89c1a5000 rwx  186  186   1   0 -----
> sw
> >   6460        0x8a346d000        0x8a36dd000 rwx   43   43   1   0 -----
> sw
> >
> > The "sw" node type means swap.
> >
> > Each of those 3 memory mappings is 2.4375 MiB in size, and probably made
> by
> > the bridges code from my earlier post, but isn't a stack.
> >
> > A stack would have flag "D" (grows Down), but none of the mappings with
> > flag "D" have permission "x".
> >
> > So it looks like on FreeBSD noexecstack already works like it should, at
> > build time and run time.
> >
> > Now why is it broken on Linux? I don't have time to check right now,
> please
> > try the above on Linux and post what you get?
>
> maybe this AI code analysis helps (take it with a grain of salt):
>
> SELinux fights AOO because AOO does something SELinux is /specifically
> built to stop/ — and it's not the markings, it's the runtime behavior.
> vtablefactory.cxx:99-109
>
> <vscode-webview://0and1hd7healr0tlpsb02psleiicthqlvde48hdaraebfrgihgpa/main/bridges/source/cpp_uno/shared/vtablefactory.cxx#L99-L109>:
>
> |p = mmap(0, n, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANON, -1, 0);
> ... else if (mprotect(p, n, PROT_READ | PROT_WRITE | PROT_EXEC) == -1)
> // W+X anon |
>
> That's the smoking gun. The UNO C++↔UNO bridge *generates machine code
> at runtime* — the vtable trampolines / code snippets (cpp2uno.cxx:412
> |codeSnippet|
> <vscode-webview://0and1hd7healr0tlpsb02psleiicthqlvde48hdaraebfrgihgpa/main/bridges/source/cpp_uno/gcc3_linux_x86-64/cpp2uno.cxx#L412>)
>
> — into anonymous memory it then marks executable. Those are exactly the
> 2.44 MiB |rwx| pools you saw on FreeBSD. AOO is, in effect, a JIT.
>
>
>     Mapping AOO's behavior onto SELinux's permission classes
>
> Vanilla Linux DAC just allows W+X memory. SELinux doesn't — it gates
> executable memory through distinct permissions, and AOO trips four of them:
>
> SELinux perm    What it guards  Where AOO triggers it
> *|execmem|*     anonymous W+X / making writable anon memory executable
> |allocExec|mmap+mprotect to RWX
> <vscode-webview://0and1hd7healr0tlpsb02psleiicthqlvde48hdaraebfrgihgpa/main/bridges/source/cpp_uno/shared/vtablefactory.cxx#L100-L105>—
>
> the bridge snippet pools
> *|execstack|*   |mprotect|-ing the process's own stack to executable
> glibc|_dl_make_stack_executable|when an exec-stack DSO is|dlopen|'d (our
> earlier discussion)
> *|execmod|*     making a/file-backed, relocation-modified/page executable
> *text relocations (TEXTREL)*in shipped|.so|s from non-PIC asm — the
> loader writes into|.text|, so the page is W+X
> *|execheap|*    executable heap (|brk|/|sbrk|region)    if exec memory is
> satisfied from the heap rather than a fresh|mmap|
>
> So everything we've discussed lands here: the unmarked/legacy asm →
> |execstack|/|execmod|; the runtime codegen → |execmem|.
>
>
>     Why it's/structural/, not just a markings bug
>
> This is the key distinction from the previous two turns. You can fix
> markings (|.note.GNU-stack|, |-z noexecstack|) and you can fix TEXTRELs
> (build everything |-fPIC|). But |execmem| is *unavoidable by design*: a
> UNO bridge /must/ synthesize trampolines at runtime to glue the C++
> vtable ABI to UNO's dispatch. Even a perfectly noexecstack-clean,
> fully-PIC AOO still calls |allocExec|. So an SELinux-confined domain
> will deny it regardless — the only resolutions are a policy allow rule,
> the |allow_execmem|/|deny_execmem| boolean, or a labeled exception for
> the soffice domain. That's why SELinux singles AOO out where most
> desktop apps sail through: most apps never allocate executable memory;
> AOO does it on every component load.
>
> (There's a partial escape hatch in the same file: the MAP_SHARED fd path
> at line 287
> <vscode-webview://0and1hd7healr0tlpsb02psleiicthqlvde48hdaraebfrgihgpa/main/bridges/source/cpp_uno/shared/vtablefactory.cxx#L287>
>
> uses a *double mapping* — one |RW| view and a separate |RX| view of the
> same shm object — which is W^X-clean and dodges |execmem|-on-anon. But
> it's conditional/fallback, not the default |allocExec| path, so the W+X
> |mprotect| is what actually runs.)
>
>
>     And why FreeBSD didn't fight it
>
> FreeBSD has no SELinux/Type-Enforcement layer in the default install —
> its MAC framework isn't enforcing |execmem|/|execstack|/|execmod|
> equivalents. Combined with rtld not poisoning the stack (prior turn),
> the /same/ JIT-ing bridge code that draws AVC denials on an enforcing
> SELinux box just runs silently on FreeBSD. The behavior is identical;
> only the enforcement regime differs.
>
> So the one-line answer: *SELinux fights AOO because AOO is a JIT* — the
> UNO bridges allocate W+X memory (|execmem|) on every C++ component
> bridge, on top of any residual exec-stack/TEXTREL
> (|execstack|/|execmod|) from legacy asm — and SELinux's whole purpose is
> to deny exactly those operations unless explicitly permitted.
>
> Here is the recommended action by claude (dont know if this is rubbish,
> but maybe it gives more insight):
>
> can we redisign uno in a SELinux save way?
>
> Short answer: yes, and the cleanest fix is well-trodden ground — but
> it's a /source/ redesign of the bridges, which puts it in the same
> bucket as your VCL-lifetime track: a post-migration design effort,
> explicitly outside the "don't touch source" Bazel scope. Let me lay out
> the design space honestly, because there are three very different tiers
> with very different cost/payoff.
>
>
>     What exactly has to go
>
> Recall the SELinux triggers ranked by how fundamental they are:
>
>   * *|execmem|*— the W+X|mprotect|in|allocExec|
>
> <vscode-webview://0and1hd7healr0tlpsb02psleiicthqlvde48hdaraebfrgihgpa/main/bridges/source/cpp_uno/shared/vtablefactory.cxx#L100-L105>./This
>     is the architectural one./Every C++ component load synthesizes
>     machine code.
>   * |execstack|/|execmod|— legacy-asm side effects (unmarked|.s|,
>     TEXTRELs). Fixable with notes/|-fPIC|,*independent*of the redesign.
>
> So "SELinux-safe UNO" = "eliminate runtime-generated executable code
> from the C++ bridge." Everything else is hygiene.
>
>
>     Why the codegen exists (the constraint you have to design around)
>
> The C++↔UNO bridge fakes a C++ vtable. At a virtual call site the /only/
> thing distinguishing slot N from slot M is *the address jumped to* — the
> callee gets no "which slot am I" parameter. So the bridge emits N tiny
> thunks, each baking its own |(vtableOffset, functionOffset)| as an
> immediate, then jumping to the shared |cpp_vtable_call| (cpp2uno.cxx:412
> <vscode-webview://0and1hd7healr0tlpsb02psleiicthqlvde48hdaraebfrgihgpa/main/bridges/source/cpp_uno/gcc3_linux_x86-64/cpp2uno.cxx#L412>).
>
> N distinct entry addresses ⇒ N distinct code blobs. That's the whole
> reason for the JIT.
>
> The key insight: *those thunks are pure functions of their index.*
> Nothing in them depends on runtime data except an integer that's
> bounded. That's exactly the property that lets you precompute them.
>
>
>     Tier 1 — static trampoline table (recommended pragmatic fix)
>
> Generate the thunks at *compile time*: a fixed array of trampolines
> |thunk[0..N_MAX]|, each hardcoded with its own index, living in the
> bridge |.so|'s normal read-only |.text|. |vtablefactory| stops calling
> |mmap|/|mprotect| and just hands out |&thunk[i]|. No anonymous memory,
> no |mprotect|, *no |execmem|* — the code executes from a file-backed,
> never-writable segment, which is precisely what SELinux is happy with.
>
>   * *Precedent:*this is exactly what libffi did in 2021
>     (|FFI_EXEC_STATIC_TRAMP|) to become W^X / SELinux / PaX-clean — same
>     problem, same solution. You're not inventing anything.
>   * *Cost:*pick an|N_MAX|upper bound on slots-per-proxy (UNO interfaces
>     are large but bounded; even 4096 thunks × ~24 B ≈ 100 KB, trivial).
>     The per-slot variable data (|this|-relative routing) is already
>     recoverable from the proxy object, so the thunk only needs to inject
>     the index — which is what makes static precompute work.
>   * *Catch:*the thunks are hand asm/per architecture/(x86, x86-64, arm,
>     …). Each|gcc3_*|bridge needs its static table. That's mechanical but
>     it's N ports.
>
> This is the highest payoff-per-effort: it kills the structural |execmem|
> denial without changing the UNO API at all.
>
>
>     Tier 2 — make the dual-mapping the default (weak for SELinux)
>
> The code already has a W^X path — the |MAP_SHARED| fd double-mapping at
> line 287
> <vscode-webview://0and1hd7healr0tlpsb02psleiicthqlvde48hdaraebfrgihgpa/main/bridges/source/cpp_uno/shared/vtablefactory.cxx#L287>:
>
> write through an |RW| view, execute through a separate |RX| view, never
> W+X simultaneously. Promoting it to the default fixes plain-kernel W^X
> and PaX/grsec. *But it only partially helps SELinux* — mapping a
> writable-backed object |PROT_EXEC| still draws |execmod|/|execmem|
> checks depending on labeling and |memfd| exec seals. So this satisfies
> hardened /kernels/ but not necessarily an /enforcing SELinux policy/.
> Lower priority than Tier 1.
>
>
>     Tier 3 — remove the C++ bridge entirely (the "real" redesign)
>
> The bridge exists only because |cppu|/|cppuhelper| expose UNO objects as
> native C++ objects with real vtables, so C++ code can call them with
> ordinary virtual-call syntax across an ABI boundary. If, at *IDL-compile
> time*, you generate a concrete C++ proxy /implementation/ per interface
> (extend |cppumaker| to emit method bodies that forward to the C
> |uno_Interface::pDispatcher|), there are no synthetic vtables, no
> thunks, no executable memory — the marshalling boundary moves from
> runtime machine code to compile-time-generated C++. All UNO interfaces
> are known via IDL, so nothing is lost by requiring compile-time knowledge.
>
>   * *Result:*the cleanest possible outcome —|execmem|gone by
>     construction, and the same approach generalizes the Java/.NET
>     bridges' rationale.
>   * *Cost:*large. Touches|cppumaker|, the cppu calling convention, every
>     component's build, binary-compat of the C++ UNO ABI. This is a
>     multi-quarter effort and a genuine ABI event.
>
>
>     My recommendation
>
> Do *Tier 1* for the SELinux win and *the asm-hygiene fixes*
> (|.note.GNU-stack| everywhere + |-fPIC|/no-TEXTREL + extend |-z
> noexecstack| to executables) for |execstack|/|execmod|. Together those
> let AOO run under enforcing SELinux with /no/ custom policy module or
> boolean — which is the actual goal. Tier 3 is the "if you're rewriting
> UNO anyway" answer; don't gate the SELinux fix on it.
>
>
> HTH
> Peter
>
> >
> >
> >> Best regards,
> >> --
> >> Arrigo
> >>
> >>
> >>
> > Regards
> > Damjan
> >



As usual, that AI code analysis does get a few things right and comes up
with an interesting idea, but also misses things and gets them wrong.

There are 2 related ideas:
* noexecstack, don't allow executable code on the stack.
* W^X (W XOR X), don't allow any memory page anywhere to be both writable
and executable. This is a superset of noexecstack, and offers greater
protection, as even heap or static memory cannot be used to inject
malicious code.

AFAICT, the problem with rwx permissions for those JIT memory pools is W^X,
not noexecstack.

Let's begin with "Tier 2". The way out of rwx permissions isn't necessarily
a double memory-map, one writable, one executable. As per this article (
https://freebsdfoundation.org/wp-content/uploads/2019/11/Improving-Memory-Permissions-in-FreeBSD.pdf),
"it is usually a simple matter to map pages writable, write generated code
to them, and then make them executable." In other words, a single mapping,
that is first only writable, then only readable and executable, would be
enough - and is simpler and safer than the double memory mapping the AI
proposed, where the same memory is still both writable and executable, but
in different mapped regions.

Tier 3 is a very interesting idea, that could vastly improve portability of
OpenOffice to other architectures, without needing to write any assembly.
Since we have full knowledge of the UNO interface at IDL compile time, yes,
we genuinely could generate the C++ proxy then. So why don't we? Maybe the
answer is performance. With compile-time proxies, we would have to call C++
-> UNO -> C++. With run-time proxies, and both components built with
compatible compilers and ABIs (something we don't know ahead of time), it
can short-circuit to C++ -> C++. But surely the choice of whether to use
the proxy or link objects directly also exists at runtime, so maybe
performance could be equally good. Memory usage could be higher, as each
interface would need its own marshalling code, instead of the
reusable generic marshalling we use now.

I don't really understand tier 1.

Regards
Damjan

Re: SELinux [was: WG: There is only one problem in OpenOffice....

Posted by Pe...@apache.org.
Am 08.06.2026 um 04:54 schrieb Damjan Jovanovic:
> On Sun, Jun 7, 2026 at 10:43 PM Arrigo Marchiori<ar...@apache.org> wrote:
>
>> Hello Damjan, All,
>>
>> On Sun, Jun 07, 2026 at 10:00:37PM +0200, Damjan Jovanovic wrote:
>>
>>> Could this be the problem? I think it must be, OpenGrok shows there's
>>> nowhere else in OpenOffice we use PROT_EXEC.
>>>
>>> At least on lines 105 and 111 it's granting read, write and execute
>>> permissions together:
>>>
>>> File main/bridges/source/cpp_uno/shared/vtablefactory.cxx:
>>>
>>>       77 extern "C" void * SAL_CALL allocExec(rtl_arena_type *, sal_Size *
>>> size) {
>>> ...
>>>       96     sal_Size n = (*size + (pagesize - 1)) & ~(pagesize - 1);
>>>       97     void * p;
>>>       98 #if defined SAL_UNX
>>>       99     p = mmap(
>>>      100         0, n, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANON, -1,
>>>      101         0);
>>>      102     if (p == MAP_FAILED) {
>>>      103         p = 0;
>>>      104     }
>>>      105     else if (mprotect (static_cast<char*>(p), n, PROT_READ |
>>> PROT_WRITE | PROT_EXEC) == -1)
>>>      106     {
>>>      107         munmap (static_cast<char*>(p), n);
>>>      108         p = 0;
>>>      109     }
>>>      110 #elif defined SAL_W32
>>>      111     p = VirtualAlloc(0, n, MEM_COMMIT, PAGE_EXECUTE_READWRITE);
>>>      112 #elif defined(SAL_OS2)
>>>      113     p = 0;
>>>      114     DosAllocMem( &p, n, PAG_COMMIT | PAG_READ | PAG_WRITE |
>>> OBJ_ANY);
>>>      115 #endif
>>>      116     if (p != 0) {
>>>      117         *size = n;
>>>      118     }
>>>      119     return p;
>>>      120 }
>>>
>>> There are already some workarounds for SELinux in this directory, but we
>>> probably need more.
>> I don't know SELinux well, nor those syscalls. So I cannot confirm,
>> unfortunately.
>>
>> Do you think we should put a ``workaround for SELinux'' in place,
>> rather than running execstack as I was suggesting in PR #483?  If so,
>> could you please point me to some documentation? I have access both to
>> an ``old'' build VM and a ``new'' Linux-based system, so I can easily
>> do tests.
>>
>>
> Let's understand this better first.
>
> This is an excellent blog post on the topic:
> https://linuxvox.com/blog/stacks-are-executable-even-with-noexecstack/
>
> Now as per that article, note how both FreeBSD and Linux already link with
> -Wl,-z,noexecstack for libraries, at least with gbuild:
>
> main/solenv]$ grep execstack * -R
> gbuild/platform/freebsd.mk:gb_Library_TARGETTYPEFLAGS := -shared
> -Wl,-z,noexecstack
> gbuild/platform/linux.mk:gb_Library_TARGETTYPEFLAGS := -shared
> -Wl,-z,noexecstack
>
> but even this seems unnecessary, as that should be the default with GCC >=
> 4.1.
>
> A quick audit on my build tree
> (main/instsetoo_native/unxfbsdx/Apache_OpenOffice/installed/install/en-US/openoffice4/program)
> shows how nothing has executable stacks:
>
> find . -type f | while read i; do readelf -l "$i" 2>/dev/null| grep
> GNU_STACK|grep E; done
> (no output. You can remove the "grep E" to see every GNU_STACK entry has
> only "RW" permissions.)
>
> It's of course possible that the stack is made executable at run-time,
> because when libraries needing execstack are loaded, they poison the stack
> of the entire application. Even this can be checked:
>
> procstat vm PID |grep rwx
> (or on Linux, use "cat /proc/PID/map")
>
>    PID              START                END PRT  RES PRES REF SHD FLAG  TP
> PATH
>   6460        0x89b9a5000        0x89bc15000 rwx  134  134   1   0 ----- sw
>   6460        0x89bf35000        0x89c1a5000 rwx  186  186   1   0 ----- sw
>   6460        0x8a346d000        0x8a36dd000 rwx   43   43   1   0 ----- sw
>
> The "sw" node type means swap.
>
> Each of those 3 memory mappings is 2.4375 MiB in size, and probably made by
> the bridges code from my earlier post, but isn't a stack.
>
> A stack would have flag "D" (grows Down), but none of the mappings with
> flag "D" have permission "x".
>
> So it looks like on FreeBSD noexecstack already works like it should, at
> build time and run time.
>
> Now why is it broken on Linux? I don't have time to check right now, please
> try the above on Linux and post what you get?

maybe this AI code analysis helps (take it with a grain of salt):

SELinux fights AOO because AOO does something SELinux is /specifically 
built to stop/ — and it's not the markings, it's the runtime behavior. 
vtablefactory.cxx:99-109 
<vscode-webview://0and1hd7healr0tlpsb02psleiicthqlvde48hdaraebfrgihgpa/main/bridges/source/cpp_uno/shared/vtablefactory.cxx#L99-L109>:

|p = mmap(0, n, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANON, -1, 0); 
... else if (mprotect(p, n, PROT_READ | PROT_WRITE | PROT_EXEC) == -1) 
// W+X anon |

That's the smoking gun. The UNO C++↔UNO bridge *generates machine code 
at runtime* — the vtable trampolines / code snippets (cpp2uno.cxx:412 
|codeSnippet| 
<vscode-webview://0and1hd7healr0tlpsb02psleiicthqlvde48hdaraebfrgihgpa/main/bridges/source/cpp_uno/gcc3_linux_x86-64/cpp2uno.cxx#L412>) 
— into anonymous memory it then marks executable. Those are exactly the 
2.44 MiB |rwx| pools you saw on FreeBSD. AOO is, in effect, a JIT.


    Mapping AOO's behavior onto SELinux's permission classes

Vanilla Linux DAC just allows W+X memory. SELinux doesn't — it gates 
executable memory through distinct permissions, and AOO trips four of them:

SELinux perm 	What it guards 	Where AOO triggers it
*|execmem|* 	anonymous W+X / making writable anon memory executable 
|allocExec|mmap+mprotect to RWX 
<vscode-webview://0and1hd7healr0tlpsb02psleiicthqlvde48hdaraebfrgihgpa/main/bridges/source/cpp_uno/shared/vtablefactory.cxx#L100-L105>— 
the bridge snippet pools
*|execstack|* 	|mprotect|-ing the process's own stack to executable 
glibc|_dl_make_stack_executable|when an exec-stack DSO is|dlopen|'d (our 
earlier discussion)
*|execmod|* 	making a/file-backed, relocation-modified/page executable 
*text relocations (TEXTREL)*in shipped|.so|s from non-PIC asm — the 
loader writes into|.text|, so the page is W+X
*|execheap|* 	executable heap (|brk|/|sbrk|region) 	if exec memory is 
satisfied from the heap rather than a fresh|mmap|

So everything we've discussed lands here: the unmarked/legacy asm → 
|execstack|/|execmod|; the runtime codegen → |execmem|.


    Why it's/structural/, not just a markings bug

This is the key distinction from the previous two turns. You can fix 
markings (|.note.GNU-stack|, |-z noexecstack|) and you can fix TEXTRELs 
(build everything |-fPIC|). But |execmem| is *unavoidable by design*: a 
UNO bridge /must/ synthesize trampolines at runtime to glue the C++ 
vtable ABI to UNO's dispatch. Even a perfectly noexecstack-clean, 
fully-PIC AOO still calls |allocExec|. So an SELinux-confined domain 
will deny it regardless — the only resolutions are a policy allow rule, 
the |allow_execmem|/|deny_execmem| boolean, or a labeled exception for 
the soffice domain. That's why SELinux singles AOO out where most 
desktop apps sail through: most apps never allocate executable memory; 
AOO does it on every component load.

(There's a partial escape hatch in the same file: the MAP_SHARED fd path 
at line 287 
<vscode-webview://0and1hd7healr0tlpsb02psleiicthqlvde48hdaraebfrgihgpa/main/bridges/source/cpp_uno/shared/vtablefactory.cxx#L287> 
uses a *double mapping* — one |RW| view and a separate |RX| view of the 
same shm object — which is W^X-clean and dodges |execmem|-on-anon. But 
it's conditional/fallback, not the default |allocExec| path, so the W+X 
|mprotect| is what actually runs.)


    And why FreeBSD didn't fight it

FreeBSD has no SELinux/Type-Enforcement layer in the default install — 
its MAC framework isn't enforcing |execmem|/|execstack|/|execmod| 
equivalents. Combined with rtld not poisoning the stack (prior turn), 
the /same/ JIT-ing bridge code that draws AVC denials on an enforcing 
SELinux box just runs silently on FreeBSD. The behavior is identical; 
only the enforcement regime differs.

So the one-line answer: *SELinux fights AOO because AOO is a JIT* — the 
UNO bridges allocate W+X memory (|execmem|) on every C++ component 
bridge, on top of any residual exec-stack/TEXTREL 
(|execstack|/|execmod|) from legacy asm — and SELinux's whole purpose is 
to deny exactly those operations unless explicitly permitted.

Here is the recommended action by claude (dont know if this is rubbish, 
but maybe it gives more insight):

can we redisign uno in a SELinux save way?

Short answer: yes, and the cleanest fix is well-trodden ground — but 
it's a /source/ redesign of the bridges, which puts it in the same 
bucket as your VCL-lifetime track: a post-migration design effort, 
explicitly outside the "don't touch source" Bazel scope. Let me lay out 
the design space honestly, because there are three very different tiers 
with very different cost/payoff.


    What exactly has to go

Recall the SELinux triggers ranked by how fundamental they are:

  * *|execmem|*— the W+X|mprotect|in|allocExec|
    <vscode-webview://0and1hd7healr0tlpsb02psleiicthqlvde48hdaraebfrgihgpa/main/bridges/source/cpp_uno/shared/vtablefactory.cxx#L100-L105>./This
    is the architectural one./Every C++ component load synthesizes
    machine code.
  * |execstack|/|execmod|— legacy-asm side effects (unmarked|.s|,
    TEXTRELs). Fixable with notes/|-fPIC|,*independent*of the redesign.

So "SELinux-safe UNO" = "eliminate runtime-generated executable code 
from the C++ bridge." Everything else is hygiene.


    Why the codegen exists (the constraint you have to design around)

The C++↔UNO bridge fakes a C++ vtable. At a virtual call site the /only/ 
thing distinguishing slot N from slot M is *the address jumped to* — the 
callee gets no "which slot am I" parameter. So the bridge emits N tiny 
thunks, each baking its own |(vtableOffset, functionOffset)| as an 
immediate, then jumping to the shared |cpp_vtable_call| (cpp2uno.cxx:412 
<vscode-webview://0and1hd7healr0tlpsb02psleiicthqlvde48hdaraebfrgihgpa/main/bridges/source/cpp_uno/gcc3_linux_x86-64/cpp2uno.cxx#L412>). 
N distinct entry addresses ⇒ N distinct code blobs. That's the whole 
reason for the JIT.

The key insight: *those thunks are pure functions of their index.* 
Nothing in them depends on runtime data except an integer that's 
bounded. That's exactly the property that lets you precompute them.


    Tier 1 — static trampoline table (recommended pragmatic fix)

Generate the thunks at *compile time*: a fixed array of trampolines 
|thunk[0..N_MAX]|, each hardcoded with its own index, living in the 
bridge |.so|'s normal read-only |.text|. |vtablefactory| stops calling 
|mmap|/|mprotect| and just hands out |&thunk[i]|. No anonymous memory, 
no |mprotect|, *no |execmem|* — the code executes from a file-backed, 
never-writable segment, which is precisely what SELinux is happy with.

  * *Precedent:*this is exactly what libffi did in 2021
    (|FFI_EXEC_STATIC_TRAMP|) to become W^X / SELinux / PaX-clean — same
    problem, same solution. You're not inventing anything.
  * *Cost:*pick an|N_MAX|upper bound on slots-per-proxy (UNO interfaces
    are large but bounded; even 4096 thunks × ~24 B ≈ 100 KB, trivial).
    The per-slot variable data (|this|-relative routing) is already
    recoverable from the proxy object, so the thunk only needs to inject
    the index — which is what makes static precompute work.
  * *Catch:*the thunks are hand asm/per architecture/(x86, x86-64, arm,
    …). Each|gcc3_*|bridge needs its static table. That's mechanical but
    it's N ports.

This is the highest payoff-per-effort: it kills the structural |execmem| 
denial without changing the UNO API at all.


    Tier 2 — make the dual-mapping the default (weak for SELinux)

The code already has a W^X path — the |MAP_SHARED| fd double-mapping at 
line 287 
<vscode-webview://0and1hd7healr0tlpsb02psleiicthqlvde48hdaraebfrgihgpa/main/bridges/source/cpp_uno/shared/vtablefactory.cxx#L287>: 
write through an |RW| view, execute through a separate |RX| view, never 
W+X simultaneously. Promoting it to the default fixes plain-kernel W^X 
and PaX/grsec. *But it only partially helps SELinux* — mapping a 
writable-backed object |PROT_EXEC| still draws |execmod|/|execmem| 
checks depending on labeling and |memfd| exec seals. So this satisfies 
hardened /kernels/ but not necessarily an /enforcing SELinux policy/. 
Lower priority than Tier 1.


    Tier 3 — remove the C++ bridge entirely (the "real" redesign)

The bridge exists only because |cppu|/|cppuhelper| expose UNO objects as 
native C++ objects with real vtables, so C++ code can call them with 
ordinary virtual-call syntax across an ABI boundary. If, at *IDL-compile 
time*, you generate a concrete C++ proxy /implementation/ per interface 
(extend |cppumaker| to emit method bodies that forward to the C 
|uno_Interface::pDispatcher|), there are no synthetic vtables, no 
thunks, no executable memory — the marshalling boundary moves from 
runtime machine code to compile-time-generated C++. All UNO interfaces 
are known via IDL, so nothing is lost by requiring compile-time knowledge.

  * *Result:*the cleanest possible outcome —|execmem|gone by
    construction, and the same approach generalizes the Java/.NET
    bridges' rationale.
  * *Cost:*large. Touches|cppumaker|, the cppu calling convention, every
    component's build, binary-compat of the C++ UNO ABI. This is a
    multi-quarter effort and a genuine ABI event.


    My recommendation

Do *Tier 1* for the SELinux win and *the asm-hygiene fixes* 
(|.note.GNU-stack| everywhere + |-fPIC|/no-TEXTREL + extend |-z 
noexecstack| to executables) for |execstack|/|execmod|. Together those 
let AOO run under enforcing SELinux with /no/ custom policy module or 
boolean — which is the actual goal. Tier 3 is the "if you're rewriting 
UNO anyway" answer; don't gate the SELinux fix on it.


HTH
Peter

>
>
>> Best regards,
>> --
>> Arrigo
>>
>>
>>
> Regards
> Damjan
>

Re: SELinux [was: WG: There is only one problem in OpenOffice....

Posted by Damjan Jovanovic <da...@apache.org>.
On Sun, Jun 7, 2026 at 10:43 PM Arrigo Marchiori <ar...@apache.org> wrote:

> Hello Damjan, All,
>
> On Sun, Jun 07, 2026 at 10:00:37PM +0200, Damjan Jovanovic wrote:
>
> > Could this be the problem? I think it must be, OpenGrok shows there's
> > nowhere else in OpenOffice we use PROT_EXEC.
> >
> > At least on lines 105 and 111 it's granting read, write and execute
> > permissions together:
> >
> > File main/bridges/source/cpp_uno/shared/vtablefactory.cxx:
> >
> >      77 extern "C" void * SAL_CALL allocExec(rtl_arena_type *, sal_Size *
> > size) {
> > ...
> >      96     sal_Size n = (*size + (pagesize - 1)) & ~(pagesize - 1);
> >      97     void * p;
> >      98 #if defined SAL_UNX
> >      99     p = mmap(
> >     100         0, n, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANON, -1,
> >     101         0);
> >     102     if (p == MAP_FAILED) {
> >     103         p = 0;
> >     104     }
> >     105     else if (mprotect (static_cast<char*>(p), n, PROT_READ |
> > PROT_WRITE | PROT_EXEC) == -1)
> >     106     {
> >     107         munmap (static_cast<char*>(p), n);
> >     108         p = 0;
> >     109     }
> >     110 #elif defined SAL_W32
> >     111     p = VirtualAlloc(0, n, MEM_COMMIT, PAGE_EXECUTE_READWRITE);
> >     112 #elif defined(SAL_OS2)
> >     113     p = 0;
> >     114     DosAllocMem( &p, n, PAG_COMMIT | PAG_READ | PAG_WRITE |
> > OBJ_ANY);
> >     115 #endif
> >     116     if (p != 0) {
> >     117         *size = n;
> >     118     }
> >     119     return p;
> >     120 }
> >
> > There are already some workarounds for SELinux in this directory, but we
> > probably need more.
>
> I don't know SELinux well, nor those syscalls. So I cannot confirm,
> unfortunately.
>
> Do you think we should put a ``workaround for SELinux'' in place,
> rather than running execstack as I was suggesting in PR #483?  If so,
> could you please point me to some documentation? I have access both to
> an ``old'' build VM and a ``new'' Linux-based system, so I can easily
> do tests.
>
>
Let's understand this better first.

This is an excellent blog post on the topic:
https://linuxvox.com/blog/stacks-are-executable-even-with-noexecstack/

Now as per that article, note how both FreeBSD and Linux already link with
-Wl,-z,noexecstack for libraries, at least with gbuild:

main/solenv]$ grep execstack * -R
gbuild/platform/freebsd.mk:gb_Library_TARGETTYPEFLAGS := -shared
-Wl,-z,noexecstack
gbuild/platform/linux.mk:gb_Library_TARGETTYPEFLAGS := -shared
-Wl,-z,noexecstack

but even this seems unnecessary, as that should be the default with GCC >=
4.1.

A quick audit on my build tree
(main/instsetoo_native/unxfbsdx/Apache_OpenOffice/installed/install/en-US/openoffice4/program)
shows how nothing has executable stacks:

find . -type f | while read i; do readelf -l "$i" 2>/dev/null| grep
GNU_STACK|grep E; done
(no output. You can remove the "grep E" to see every GNU_STACK entry has
only "RW" permissions.)

It's of course possible that the stack is made executable at run-time,
because when libraries needing execstack are loaded, they poison the stack
of the entire application. Even this can be checked:

procstat vm PID |grep rwx
(or on Linux, use "cat /proc/PID/map")

  PID              START                END PRT  RES PRES REF SHD FLAG  TP
PATH
 6460        0x89b9a5000        0x89bc15000 rwx  134  134   1   0 ----- sw
 6460        0x89bf35000        0x89c1a5000 rwx  186  186   1   0 ----- sw
 6460        0x8a346d000        0x8a36dd000 rwx   43   43   1   0 ----- sw

The "sw" node type means swap.

Each of those 3 memory mappings is 2.4375 MiB in size, and probably made by
the bridges code from my earlier post, but isn't a stack.

A stack would have flag "D" (grows Down), but none of the mappings with
flag "D" have permission "x".

So it looks like on FreeBSD noexecstack already works like it should, at
build time and run time.

Now why is it broken on Linux? I don't have time to check right now, please
try the above on Linux and post what you get?


> Best regards,
> --
> Arrigo
>
>
>
Regards
Damjan

Re: SELinux [was: WG: There is only one problem in OpenOffice....

Posted by Arrigo Marchiori <ar...@apache.org>.
Hello Damjan, All,

On Sun, Jun 07, 2026 at 10:00:37PM +0200, Damjan Jovanovic wrote:

> Could this be the problem? I think it must be, OpenGrok shows there's
> nowhere else in OpenOffice we use PROT_EXEC.
> 
> At least on lines 105 and 111 it's granting read, write and execute
> permissions together:
> 
> File main/bridges/source/cpp_uno/shared/vtablefactory.cxx:
> 
>      77 extern "C" void * SAL_CALL allocExec(rtl_arena_type *, sal_Size *
> size) {
> ...
>      96     sal_Size n = (*size + (pagesize - 1)) & ~(pagesize - 1);
>      97     void * p;
>      98 #if defined SAL_UNX
>      99     p = mmap(
>     100         0, n, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANON, -1,
>     101         0);
>     102     if (p == MAP_FAILED) {
>     103         p = 0;
>     104     }
>     105     else if (mprotect (static_cast<char*>(p), n, PROT_READ |
> PROT_WRITE | PROT_EXEC) == -1)
>     106     {
>     107         munmap (static_cast<char*>(p), n);
>     108         p = 0;
>     109     }
>     110 #elif defined SAL_W32
>     111     p = VirtualAlloc(0, n, MEM_COMMIT, PAGE_EXECUTE_READWRITE);
>     112 #elif defined(SAL_OS2)
>     113     p = 0;
>     114     DosAllocMem( &p, n, PAG_COMMIT | PAG_READ | PAG_WRITE |
> OBJ_ANY);
>     115 #endif
>     116     if (p != 0) {
>     117         *size = n;
>     118     }
>     119     return p;
>     120 }
> 
> There are already some workarounds for SELinux in this directory, but we
> probably need more.

I don't know SELinux well, nor those syscalls. So I cannot confirm,
unfortunately.

Do you think we should put a ``workaround for SELinux'' in place,
rather than running execstack as I was suggesting in PR #483?  If so,
could you please point me to some documentation? I have access both to
an ``old'' build VM and a ``new'' Linux-based system, so I can easily
do tests.

Best regards,
-- 
Arrigo

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org


Re: WG: There is only one problem in OpenOffice....

Posted by Peter Kovacs <le...@posteo.de>.
Hello Wolfgang,

The 4.2. has some disadvantages. There are still various bugs, why we have not released it, but back porting to the 4.1. line. So caution is advised!

Please see:
<https://cwiki.apache.org/confluence/display/OOOUSERS/Blocker+issues+4.2.0>

We hope that they get fixed soon.

Am 1. Januar 2026 18:35:00 MEZ schrieb Wolfgang Hinzmann <Wo...@protonmail.ch>:
>Dear Dean Webber,
>
>many many thanks for this great hint!
>The new year could not better start, by installing this version
>after I downloaded the compressed files.
>
>Additionally I renamed the system folder
>"$home/wolfganghinzmann/.openoffice",
>before I start the install procedure, using
>
>"sudo rpm -Uvih -nodigest -nosignature *rpm".
>
>In this way I did it, because the original problem was,
>I always got the stupid message
>"no checksum"..
>And I thought, the compressed installation files could be damaged.
>And because of this, I send a message to the OpenOffice-Team.
>
>Using the options "-nosignature -nodigest"
>the checksum examination is suppressed,
>and the installation works very fine!
>
>I am very very glad and happy now.
>OpenOffice is very good, and I am using it now for years!
>
>I wish You too a very very good year 2026!
>
>Kind regards
>Wolfgang Hinzmann, Germany
>
>🇩🇪
>Lieber Dean,
>vielen herzlichen Dank für diesen sehr guten Hinweis!
>Das neue Jahr 2026 konnte nicht besser beginnen,
>nach dem ich OpenOffice erneut installiert habe,
>und diese Files dazu benutzt habe.
>
>Zusätzlich habe ich den Systemordner
>"$home/wolfganghinzmann/.openoffice",
>umbenannt, bevor ich die Installation startete.
>
>Die Installation selbst initialisierte ich mit dem Befehl
>
>"sudo rpm -Uvih -nodigest -nosignature *rpm".
>
>Ich benutzte den Befehl in dieser Form mit den Optionen
>"-nodigest -nosignature",
>weil ich am Anfang das Problem hatte, dass beim Versuch
>der Installatin immer der Fehler kam:
>"no checksum"..
>Und so dachte ich erst irrtümlicherweise,
>dass die Installationsdateien vielleicht beschädigt sein könnten.
>
>Mit diesen Optionen "-nodigest -nosignature"
>wird diese Prüfung schlicht unterdrückt,
>und die Installation startet endlich!
>
>Dass diese Prüfung durch den RPM-Befehl
>überhaupt ausgelöst wird,
>muss mit einem Versionswechsel von Fedora-Linux zusammenhängen.
>
>Aber nun läuft OpenOffice, Version 2, einfach sehr gut!
>Bin einfach nur sehr glücklich.
>Wünsche Dir auch ein sehr gutes neues Jahr 2026.
>
>Beste GrüßeWolfgang Hinzmann
>Idstein in Hessen
>
>Gesendet mit [Proton Mail](https://proton.me/mail/home): Ein sicherer E-Mail-Dienst.
>
>Dean Webber <de...@student.uq.edu.au> schrieb am Donnerstag, 1. Januar 2026 um 02:07:
>
>> Dear Wolfgang,
>>
>> I am not a developer, but had the same file issue when I upgraded to Debian 13.
>>
>> One of the developers suggested I use the releases here, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/ , which is an upstream build which I suspect has a fix for this issue.
>>
>> Specifically for you, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_install-rpm_de.tar.gz and https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_langpack-rpm_de.tar.gz
>>
>> It worked for me, perhaps it will work for you also.
>>
>> Happy New Year.
>>
>> Faithfully,
>> Dean
>>
>> Translation en -> german
>>
>> Lieber Wolfgang,
>>
>> Ich bin kein Entwickler, hatte aber beim Upgrade auf Debian 13 das gleiche Dateiproblem.
>>
>> Einer der Entwickler schlug vor, die Releases hier zu verwenden, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de / , Das ist ein Upstream-Build, von dem ich vermute, dass es eine Lösung für dieses Problem gibt.
>>
>> Speziell für Sie, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_install-rpm_de.tar.gz und https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_langpack-rpm_de.tar.gz
>>
>> Es hat bei mir funktioniert, vielleicht funktioniert es auch bei dir.
>>
>> Frohes neues Jahr.
>>
>> Treu,
>> Dean
>>
>> ---------------------------------------------------------------
>>
>> From: Wolfgang Hinzmann <Wo...@protonmail.ch>
>> Sent: 31 December 2025 02:12
>> To: dev@ <de...@openoffice.apache.org>
>> Subject: Re: WG: There is only one problem in OpenOffice....
>>
>> Hello Marcus,
>>
>> many thanks for these hints again.
>>
>> I tried both:
>>
>> As first step I renamed the folder
>>
>> $HOME/.openoffice.org/4/user/basic"
>>
>> After that, I started my computer new.
>> When I started now OpenOffice, I got several messages like
>> "Cannot load [file...].....".
>> Despite this fact, I was able to create a new textfile and could save it.
>> As usual with the "Error message".
>>
>> Because of this errors I tried the second step:
>> I renamed the whole folder - as You described:
>> $HOME/.openoffice.org
>>
>> Then I started my computer new,
>> and after that I started OpenOffice.
>> A window opened, and I was asked to put my whole Name, First name and family name,
>> and an abbreviation for it. By clicking then the "Coninue-Button"
>> OpenOffice opened, and I had the choice between the tools like Writer, Calculation...
>> I opened writer, opened a new document, and then saving it:
>> "The error message appears again" - as I described it.
>>
>> In effect, both steps - renaming the basic-folder, and after that,
>> renaming the "$HOME/.openoffice.org"- folder,
>> did not change the error away...
>>
>> So I decided, to change these system-folders in its originial state
>> by renaming it, and deleting the new created ones.
>>
>> If You are interested in the error-messages, I have got,
>> I would repeat the steps.
>>
>> I had really very great hope. Especially with the second step,
>> because it started with the "Welcome-Window", and asking me
>> for my name and its abbreviation.
>> Strange. Very strange.
>>
>> Despite of this:
>> I wish You a very good new year 2026!
>>
>> Best wishes Wolfgang
>>
>> Gesendet mit [Proton Mail](https://proton.me/mail/home): Ein sicherer E-Mail-Dienst.
>>
>> Marcus marcus.mail@wtnet.de schrieb am Montag, 29. Dezember 2025 um 16:15:
>>
>>> Hi Wolfgang,
>>>
>>> first of all you can try to reset your Basic settings. The following
>>> will reset only the Basic things (e.g., macros) but nothing more:
>>>
>>> - close all AOO documents and windows
>>> - open a terminal
>>> - rename the Basic directory "$HOME/.openoffice.org/4/user/basic"
>>> - cd ~
>>> - mv .openoffice.org/4/user/basic
>>> .openoffice.org/4/user/basic_maybe_broken
>>> - start AOO new
>>> - try if the error message is still showing
>>>
>>> If this will not help, try to reset the complete AOO profile:
>>>
>>> - close all AOO documents and windows
>>> - open a terminal
>>> - rename the complete profile directory "$HOME/.openoffice.org"
>>> - cd ~
>>> - mv .openoffice.org .openoffice.org_maybe_broken
>>> - start AOO new
>>> - try if the error message is still showing
>>>
>>> Please note:
>>> You should know that all your customized settings will be gone. They all
>>> need to be setup again.
>>>
>>> Please be carefully:
>>> Rename the profile directory. You can delete it anytime later when it is
>>> sure that the profile was somehow broken.
>>> Otherwise if the problem persists you can decide to go back to the old
>>> profile directory by renaming it back with the following commands in the
>>> terminal:
>>>
>>> cd ~
>>> rm -rf .openoffice.org
>>> mv .openoffice.org_maybe_broken .openoffice.org
>>>
>>> HTH
>>>
>>> Marcus
>>>
>>> Am 26.12.25 um 10:40 schrieb Wolfgang Hinzmann:
>>>
>>>> as You suggested, I will try OpenOffice, especially Writer,
>>>> with a new profile.
>>>> Because I never needed special settings and other useful extensions so far.
>>>> The only question for me in that wayis:
>>>> How can I realize that?
>>>> If You get other hints too, I would be very glad.
>>>>
>>>> Marcus marcus.mail@wtnet.de schrieb am Dienstag, 23. Dezember 2025 um 13:45:
>>>>
>>>>> Am 21.12.25 um 18:09 schrieb Wolfgang Hinzmann:
>>>>>
>>>>>> /as You suggested, I opened the terminal, and /
>>>>>> /typed then the command "soffice"./
>>>>>> /After that, the main-menu of OpenOffice started./
>>>>>> /There, I go to "Open last files", and opened the "Testfile01.odt"./
>>>>>> /After that I got the message:/
>>>>>>
>>>>>> /image.png/
>>>>>> /
>>>>>> /In the_terminal-window_, there are the following messages:/
>>>>>> /wolfganghinzmann@fedora:~$ soffice/
>>>>>> /javaldx: Could not find a Java Runtime Environment!/
>>>>>> /
>>>>>> /That's all./
>>>>>
>>>>> than I (you) need help from others as I've no idea at the moment what
>>>>> could be checked next.
>>>>>
>>>>> @All:
>>>>> Has anyone more hints? Thanks.
>>>>>
>>>>> Last resort:
>>>>> There is always the possibility to start from scratch with a new AOO
>>>>> profile. But that means you would loose all your settings that you have
>>>>> done up to know (customized toolbars, menus, keyboard shortcuts, see
>>>>> menu "Tools - Customize" or in German: "Extras - Anpassen" plus some
>>>>> more settings in other locations, e.g., "Format - AutoCorrext -
>>>>> AutoCorrect Optionen").
>>>>>
>>>>> So, it depends on you if you want to try it.
>>>>>
>>>>>> Marcus marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de schrieb am
>>>>>> Sonntag, 21. Dezember 2025 um 14:44:
>>>>>>
>>>>>> Am 21.12.25 um 14:12 schrieb Wolfgang Hinzmann:
>>>>>>
>>>>>> /I don't use macros in my documents./
>>>>>>
>>>>>> OK, a possible root cause less.
>>>>>>
>>>>>> /Question:/
>>>>>> /You suggested, I should try to start Writer from Terminal./
>>>>>> /How can I do this? I am really not so familiar with all these
>>>>>> commands./
>>>>>>
>>>>>> that should be simple. You just open the terminal and enter the
>>>>>> following command:
>>>>>>
>>>>>> soffice
>>>>>>
>>>>>> If this results in an unknown command, then please use it with the full
>>>>>> path. It depends where you have installed it in your system. E.g.,
>>>>>>
>>>>>> /opt/openoffice4/program/soffice
>>>>>>
>>>>>> The output until the error message appears could be helpful. Fore this
>>>>>> purpose start OpenOffice and then open a document that shows that error
>>>>>> message. After that please post the complete text here.
>>>>>>
>>>>>> Thanks
>>>>>>
>>>>>> Marcus
>>>>>>
>>>>>> Gesendet mit Proton Mail https://proton.me/mail/home: <https://
>>>>>> proton.me/mail/home:> Ein sicherer E-
>>>>>> Mail-Dienst.
>>>>>>
>>>>>> Marcus marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de
>>>>>> mailto:marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de schrieb am
>>>>>> Sonntag, 21. Dezember 2025 um 10:05:
>>>>>>
>>>>>> Am 20.12.25 um 16:32 schrieb Wolfgang Hinzmann:
>>>>>>
>>>>>> great that it is now installing and working.
>>>>>>
>>>>>> /There is only one problem with OpenOffice, I had already in the/
>>>>>> /earlier version too:/
>>>>>> /By opening an odt-file with Writer I receive always the error
>>>>>> message/
>>>>>> / "Allgemeiner Fehler"./
>>>>>> /image.png
>>>>>> /
>>>>>> /I had the hope, that this message will disappear by the
>>>>>> installation of
>>>>>> the new version./
>>>>>> /But it doesn't./
>>>>>> /
>>>>>> /But what is the solution for it?/
>>>>>> /After pressing the button "OK" I am able to work normally with
>>>>>> the file
>>>>>> I opened:/
>>>>>> /Changing it and saving it./
>>>>>>
>>>>>> do you have macros inside the document? Please have a look via
>>>>>> the menu
>>>>>> "Tools - Macros - Organize Macros - Basic" (German: "Extras -
>>>>>> Makros -
>>>>>> Makros verwalten - Basic")?
>>>>>> If so, it's up to you to decide if you still need them - then
>>>>>> you have
>>>>>> to adjust parts to run correctly. Then there is good chance that the
>>>>>> document will open without error messages. Or to delete them.
>>>>>>
>>>>>> Please make sure you have made a backup before deleting things
>>>>>> that you
>>>>>> still need.

Re: WG: There is only one problem in OpenOffice....

Posted by Wolfgang Hinzmann <Wo...@protonmail.ch>.
Dear Dean Webber,

many many thanks for this great hint!
The new year could not better start, by installing this version
after I downloaded the compressed files.

Additionally I renamed the system folder
"$home/wolfganghinzmann/.openoffice",
before I start the install procedure, using

"sudo rpm -Uvih -nodigest -nosignature *rpm".

In this way I did it, because the original problem was,
I always got the stupid message
"no checksum"..
And I thought, the compressed installation files could be damaged.
And because of this, I send a message to the OpenOffice-Team.

Using the options "-nosignature -nodigest"
the checksum examination is suppressed,
and the installation works very fine!

I am very very glad and happy now.
OpenOffice is very good, and I am using it now for years!

I wish You too a very very good year 2026!

Kind regards
Wolfgang Hinzmann, Germany

🇩🇪
Lieber Dean,
vielen herzlichen Dank für diesen sehr guten Hinweis!
Das neue Jahr 2026 konnte nicht besser beginnen,
nach dem ich OpenOffice erneut installiert habe,
und diese Files dazu benutzt habe.

Zusätzlich habe ich den Systemordner
"$home/wolfganghinzmann/.openoffice",
umbenannt, bevor ich die Installation startete.

Die Installation selbst initialisierte ich mit dem Befehl

"sudo rpm -Uvih -nodigest -nosignature *rpm".

Ich benutzte den Befehl in dieser Form mit den Optionen
"-nodigest -nosignature",
weil ich am Anfang das Problem hatte, dass beim Versuch
der Installatin immer der Fehler kam:
"no checksum"..
Und so dachte ich erst irrtümlicherweise,
dass die Installationsdateien vielleicht beschädigt sein könnten.

Mit diesen Optionen "-nodigest -nosignature"
wird diese Prüfung schlicht unterdrückt,
und die Installation startet endlich!

Dass diese Prüfung durch den RPM-Befehl
überhaupt ausgelöst wird,
muss mit einem Versionswechsel von Fedora-Linux zusammenhängen.

Aber nun läuft OpenOffice, Version 2, einfach sehr gut!
Bin einfach nur sehr glücklich.
Wünsche Dir auch ein sehr gutes neues Jahr 2026.

Beste GrüßeWolfgang Hinzmann
Idstein in Hessen

Gesendet mit [Proton Mail](https://proton.me/mail/home): Ein sicherer E-Mail-Dienst.

Dean Webber <de...@student.uq.edu.au> schrieb am Donnerstag, 1. Januar 2026 um 02:07:

> Dear Wolfgang,
>
> I am not a developer, but had the same file issue when I upgraded to Debian 13.
>
> One of the developers suggested I use the releases here, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/ , which is an upstream build which I suspect has a fix for this issue.
>
> Specifically for you, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_install-rpm_de.tar.gz and https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_langpack-rpm_de.tar.gz
>
> It worked for me, perhaps it will work for you also.
>
> Happy New Year.
>
> Faithfully,
> Dean
>
> Translation en -> german
>
> Lieber Wolfgang,
>
> Ich bin kein Entwickler, hatte aber beim Upgrade auf Debian 13 das gleiche Dateiproblem.
>
> Einer der Entwickler schlug vor, die Releases hier zu verwenden, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de / , Das ist ein Upstream-Build, von dem ich vermute, dass es eine Lösung für dieses Problem gibt.
>
> Speziell für Sie, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_install-rpm_de.tar.gz und https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_langpack-rpm_de.tar.gz
>
> Es hat bei mir funktioniert, vielleicht funktioniert es auch bei dir.
>
> Frohes neues Jahr.
>
> Treu,
> Dean
>
> ---------------------------------------------------------------
>
> From: Wolfgang Hinzmann <Wo...@protonmail.ch>
> Sent: 31 December 2025 02:12
> To: dev@ <de...@openoffice.apache.org>
> Subject: Re: WG: There is only one problem in OpenOffice....
>
> Hello Marcus,
>
> many thanks for these hints again.
>
> I tried both:
>
> As first step I renamed the folder
>
> $HOME/.openoffice.org/4/user/basic"
>
> After that, I started my computer new.
> When I started now OpenOffice, I got several messages like
> "Cannot load [file...].....".
> Despite this fact, I was able to create a new textfile and could save it.
> As usual with the "Error message".
>
> Because of this errors I tried the second step:
> I renamed the whole folder - as You described:
> $HOME/.openoffice.org
>
> Then I started my computer new,
> and after that I started OpenOffice.
> A window opened, and I was asked to put my whole Name, First name and family name,
> and an abbreviation for it. By clicking then the "Coninue-Button"
> OpenOffice opened, and I had the choice between the tools like Writer, Calculation...
> I opened writer, opened a new document, and then saving it:
> "The error message appears again" - as I described it.
>
> In effect, both steps - renaming the basic-folder, and after that,
> renaming the "$HOME/.openoffice.org"- folder,
> did not change the error away...
>
> So I decided, to change these system-folders in its originial state
> by renaming it, and deleting the new created ones.
>
> If You are interested in the error-messages, I have got,
> I would repeat the steps.
>
> I had really very great hope. Especially with the second step,
> because it started with the "Welcome-Window", and asking me
> for my name and its abbreviation.
> Strange. Very strange.
>
> Despite of this:
> I wish You a very good new year 2026!
>
> Best wishes Wolfgang
>
> Gesendet mit [Proton Mail](https://proton.me/mail/home): Ein sicherer E-Mail-Dienst.
>
> Marcus marcus.mail@wtnet.de schrieb am Montag, 29. Dezember 2025 um 16:15:
>
>> Hi Wolfgang,
>>
>> first of all you can try to reset your Basic settings. The following
>> will reset only the Basic things (e.g., macros) but nothing more:
>>
>> - close all AOO documents and windows
>> - open a terminal
>> - rename the Basic directory "$HOME/.openoffice.org/4/user/basic"
>> - cd ~
>> - mv .openoffice.org/4/user/basic
>> .openoffice.org/4/user/basic_maybe_broken
>> - start AOO new
>> - try if the error message is still showing
>>
>> If this will not help, try to reset the complete AOO profile:
>>
>> - close all AOO documents and windows
>> - open a terminal
>> - rename the complete profile directory "$HOME/.openoffice.org"
>> - cd ~
>> - mv .openoffice.org .openoffice.org_maybe_broken
>> - start AOO new
>> - try if the error message is still showing
>>
>> Please note:
>> You should know that all your customized settings will be gone. They all
>> need to be setup again.
>>
>> Please be carefully:
>> Rename the profile directory. You can delete it anytime later when it is
>> sure that the profile was somehow broken.
>> Otherwise if the problem persists you can decide to go back to the old
>> profile directory by renaming it back with the following commands in the
>> terminal:
>>
>> cd ~
>> rm -rf .openoffice.org
>> mv .openoffice.org_maybe_broken .openoffice.org
>>
>> HTH
>>
>> Marcus
>>
>> Am 26.12.25 um 10:40 schrieb Wolfgang Hinzmann:
>>
>>> as You suggested, I will try OpenOffice, especially Writer,
>>> with a new profile.
>>> Because I never needed special settings and other useful extensions so far.
>>> The only question for me in that wayis:
>>> How can I realize that?
>>> If You get other hints too, I would be very glad.
>>>
>>> Marcus marcus.mail@wtnet.de schrieb am Dienstag, 23. Dezember 2025 um 13:45:
>>>
>>>> Am 21.12.25 um 18:09 schrieb Wolfgang Hinzmann:
>>>>
>>>>> /as You suggested, I opened the terminal, and /
>>>>> /typed then the command "soffice"./
>>>>> /After that, the main-menu of OpenOffice started./
>>>>> /There, I go to "Open last files", and opened the "Testfile01.odt"./
>>>>> /After that I got the message:/
>>>>>
>>>>> /image.png/
>>>>> /
>>>>> /In the_terminal-window_, there are the following messages:/
>>>>> /wolfganghinzmann@fedora:~$ soffice/
>>>>> /javaldx: Could not find a Java Runtime Environment!/
>>>>> /
>>>>> /That's all./
>>>>
>>>> than I (you) need help from others as I've no idea at the moment what
>>>> could be checked next.
>>>>
>>>> @All:
>>>> Has anyone more hints? Thanks.
>>>>
>>>> Last resort:
>>>> There is always the possibility to start from scratch with a new AOO
>>>> profile. But that means you would loose all your settings that you have
>>>> done up to know (customized toolbars, menus, keyboard shortcuts, see
>>>> menu "Tools - Customize" or in German: "Extras - Anpassen" plus some
>>>> more settings in other locations, e.g., "Format - AutoCorrext -
>>>> AutoCorrect Optionen").
>>>>
>>>> So, it depends on you if you want to try it.
>>>>
>>>>> Marcus marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de schrieb am
>>>>> Sonntag, 21. Dezember 2025 um 14:44:
>>>>>
>>>>> Am 21.12.25 um 14:12 schrieb Wolfgang Hinzmann:
>>>>>
>>>>> /I don't use macros in my documents./
>>>>>
>>>>> OK, a possible root cause less.
>>>>>
>>>>> /Question:/
>>>>> /You suggested, I should try to start Writer from Terminal./
>>>>> /How can I do this? I am really not so familiar with all these
>>>>> commands./
>>>>>
>>>>> that should be simple. You just open the terminal and enter the
>>>>> following command:
>>>>>
>>>>> soffice
>>>>>
>>>>> If this results in an unknown command, then please use it with the full
>>>>> path. It depends where you have installed it in your system. E.g.,
>>>>>
>>>>> /opt/openoffice4/program/soffice
>>>>>
>>>>> The output until the error message appears could be helpful. Fore this
>>>>> purpose start OpenOffice and then open a document that shows that error
>>>>> message. After that please post the complete text here.
>>>>>
>>>>> Thanks
>>>>>
>>>>> Marcus
>>>>>
>>>>> Gesendet mit Proton Mail https://proton.me/mail/home: <https://
>>>>> proton.me/mail/home:> Ein sicherer E-
>>>>> Mail-Dienst.
>>>>>
>>>>> Marcus marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de
>>>>> mailto:marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de schrieb am
>>>>> Sonntag, 21. Dezember 2025 um 10:05:
>>>>>
>>>>> Am 20.12.25 um 16:32 schrieb Wolfgang Hinzmann:
>>>>>
>>>>> great that it is now installing and working.
>>>>>
>>>>> /There is only one problem with OpenOffice, I had already in the/
>>>>> /earlier version too:/
>>>>> /By opening an odt-file with Writer I receive always the error
>>>>> message/
>>>>> / "Allgemeiner Fehler"./
>>>>> /image.png
>>>>> /
>>>>> /I had the hope, that this message will disappear by the
>>>>> installation of
>>>>> the new version./
>>>>> /But it doesn't./
>>>>> /
>>>>> /But what is the solution for it?/
>>>>> /After pressing the button "OK" I am able to work normally with
>>>>> the file
>>>>> I opened:/
>>>>> /Changing it and saving it./
>>>>>
>>>>> do you have macros inside the document? Please have a look via
>>>>> the menu
>>>>> "Tools - Macros - Organize Macros - Basic" (German: "Extras -
>>>>> Makros -
>>>>> Makros verwalten - Basic")?
>>>>> If so, it's up to you to decide if you still need them - then
>>>>> you have
>>>>> to adjust parts to run correctly. Then there is good chance that the
>>>>> document will open without error messages. Or to delete them.
>>>>>
>>>>> Please make sure you have made a backup before deleting things
>>>>> that you
>>>>> still need.

Re: SELinux [was: WG: There is only one problem in OpenOffice....

Posted by Damjan Jovanovic <da...@apache.org>.
Could this be the problem? I think it must be, OpenGrok shows there's
nowhere else in OpenOffice we use PROT_EXEC.

At least on lines 105 and 111 it's granting read, write and execute
permissions together:

File main/bridges/source/cpp_uno/shared/vtablefactory.cxx:

     77 extern "C" void * SAL_CALL allocExec(rtl_arena_type *, sal_Size *
size) {
...
     96     sal_Size n = (*size + (pagesize - 1)) & ~(pagesize - 1);
     97     void * p;
     98 #if defined SAL_UNX
     99     p = mmap(
    100         0, n, PROT_READ | PROT_WRITE, MAP_PRIVATE | MAP_ANON, -1,
    101         0);
    102     if (p == MAP_FAILED) {
    103         p = 0;
    104     }
    105     else if (mprotect (static_cast<char*>(p), n, PROT_READ |
PROT_WRITE | PROT_EXEC) == -1)
    106     {
    107         munmap (static_cast<char*>(p), n);
    108         p = 0;
    109     }
    110 #elif defined SAL_W32
    111     p = VirtualAlloc(0, n, MEM_COMMIT, PAGE_EXECUTE_READWRITE);
    112 #elif defined(SAL_OS2)
    113     p = 0;
    114     DosAllocMem( &p, n, PAG_COMMIT | PAG_READ | PAG_WRITE |
OBJ_ANY);
    115 #endif
    116     if (p != 0) {
    117         *size = n;
    118     }
    119     return p;
    120 }

There are already some workarounds for SELinux in this directory, but we
probably need more.

Regards
Damjan

On Sun, Jan 11, 2026 at 4:56 PM Arrigo Marchiori <ar...@apache.org> wrote:

> Hello Dean,
>
> On Sun, Jan 11, 2026 at 12:04:38PM +0000, Dean Webber wrote:
>
> > Out of curiosity I made an apparmor profile for soffice, I wanted to
> > see what it did. I ended up logging information using journalctl to
> > compare the general error version (4.1.16) with my known working
> > version (some old release from forever ago).
>
> Thank you for your efforts!
>
> I believe that your "known working version" is in fact a 4.2.xxx, not
> 4.1.yyy, right?
>
> Would it be possible for you to grab the execstack program from
> another source, and try that, please?
>
> Instructions are here:
>
> https://blog.nashcom.de/nashcomblog.nsf/dx/domino-14.5-is-not-supported-on-brand-new-debian-13-trixie.htm
>
> I tried on a Docker instance of Debian 13 and it worked.
> The command is, for your reference:
>
> $ sudo execstack -s /opt/openoffice4/program/soffice.bin
>
> Thank you in advance and best regards,
> --
> Arrigo
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
> For additional commands, e-mail: dev-help@openoffice.apache.org
>
>

Re: SELinux [was: WG: There is only one problem in OpenOffice....

Posted by Dean Webber <de...@student.uq.edu.au>.
Hi Arrigo,

>I believe that your "known working version" is in fact a 4.2.xxx, not
> 4.1.yyy, right?

Known working refers to Apache_OpenOffice_4.2.0_Linux_x86-64_install-deb_en-GB.tar.gz ( AOO420m5(Build:9824)  -  Rev. c51e9806e6 )

>Would it be possible for you to grab the execstack program from
>another source, and try that, please?

>Instructions are here:
>https://blog.nashcom.de/nashcomblog.nsf/dx/domino-14.5-is-not-supported-on-brand-new-debian-13-trixie.htm

Okay, I used https://archive.ubuntu.com/ubuntu/pool/universe/p/prelink/execstack_0.0.20131005-1.1_amd64.deb

>I tried on a Docker instance of Debian 13 and it worked.
>The command is, for your reference:

>$ sudo execstack -s /opt/openoffice4/program/soffice.bin

Using -s to "mark binary or shared library as requiring executable stack" results in Apache_OpenOffice_4.1.16_Linux_x86_install-deb_en-GB.tar.gz ( AOO4116m3(Build:9816)  -  Rev. 277251aa7e ) being able to save (and open documents) with no general error :-)

It also results in Apache_OpenOffice_4.1.15_Linux_x86-64_install-deb_en-GB.tar.gz ( AOO4115m2(Build:9813)  -  Rev. 5f13fa0070 ) being able to save (and open documents) with no general error :-))

I will add, these saved documents do actually get saved as files and contain the changes, but just result in "general error". They are also able to be opened, but result in general error.

I reverted to the original conditions using $ sudo execstack -c /opt/openoffice4/program/soffice.bin

Again, reverting system to normal. No such issue with the 4.2.0 build, works fine without having stack marked executable.

So, looks like marking the binary or shared library as requiring executable stack using https://archive.ubuntu.com/ubuntu/pool/universe/p/prelink/execstack_0.0.20131005-1.1_amd64.deb and $ sudo execstack -s /opt/openoffice4/program/soffice.bin is a known fix or workaround for this issue.

Well done,
Dean

________________________________
From: Arrigo Marchiori <ar...@apache.org>
Sent: 12 January 2026 00:56
To: dev@openoffice.apache.org <de...@openoffice.apache.org>
Subject: Re: SELinux [was: WG: There is only one problem in OpenOffice....

Hello Dean,

On Sun, Jan 11, 2026 at 12:04:38PM +0000, Dean Webber wrote:

> Out of curiosity I made an apparmor profile for soffice, I wanted to
> see what it did. I ended up logging information using journalctl to
> compare the general error version (4.1.16) with my known working
> version (some old release from forever ago).

Thank you for your efforts!

I believe that your "known working version" is in fact a 4.2.xxx, not
4.1.yyy, right?

Would it be possible for you to grab the execstack program from
another source, and try that, please?

Instructions are here:
https://blog.nashcom.de/nashcomblog.nsf/dx/domino-14.5-is-not-supported-on-brand-new-debian-13-trixie.htm

I tried on a Docker instance of Debian 13 and it worked.
The command is, for your reference:

$ sudo execstack -s /opt/openoffice4/program/soffice.bin

Thank you in advance and best regards,
--
Arrigo

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org


Re: SELinux [was: WG: There is only one problem in OpenOffice....

Posted by Arrigo Marchiori <ar...@apache.org>.
Hello Dean,

On Sun, Jan 11, 2026 at 12:04:38PM +0000, Dean Webber wrote:

> Out of curiosity I made an apparmor profile for soffice, I wanted to
> see what it did. I ended up logging information using journalctl to
> compare the general error version (4.1.16) with my known working
> version (some old release from forever ago).

Thank you for your efforts!

I believe that your "known working version" is in fact a 4.2.xxx, not
4.1.yyy, right?

Would it be possible for you to grab the execstack program from
another source, and try that, please?

Instructions are here:
https://blog.nashcom.de/nashcomblog.nsf/dx/domino-14.5-is-not-supported-on-brand-new-debian-13-trixie.htm

I tried on a Docker instance of Debian 13 and it worked.
The command is, for your reference:

$ sudo execstack -s /opt/openoffice4/program/soffice.bin

Thank you in advance and best regards,
-- 
Arrigo

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org


Re: SELinux [was: WG: There is only one problem in OpenOffice....

Posted by Dean Webber <de...@student.uq.edu.au>.
Out of curiosity I made an apparmor profile for soffice, I wanted to see what it did. I ended up logging information using journalctl to compare the general error version (4.1.16) with my known working version (some old release from forever ago).

Please note, I have no idea if this is actually helpful, but just the information I was able to capture, which might help to diagnose where the general error occurs.

I had a look to compare them, but not sure what soffice does to save files. The general error log was shorter, I presume because it fails to save the file.

The apparmor profile;
located in /etc/apparmor.d/
The file is,
___
# This profile allows everything and only exists to give the
# application a name instead of having the label "unconfined"

abi <abi/4.0>,
include <tunables/global>

profile soffice /opt/openoffice4/program/soffice flags=(unconfined) {
  userns,

  # Site-specific additions and overrides. See local/README for details.
  include if exists <local/soffice>
}
___

I used "sudo aa-complain /etc/apparmor.d/soffice" so policy violations are logged.
I did the logging using "sudo journalctl -fx".

Now the logs, you can get a nice version of the general error version here (https://pastebin.com/0PVZVQ4S) and a nice version of the known working version here (https://pastebin.com/8K8gdi69).

Otherwise, the saving part of the logs are below.

During saving with the general error version, apparmor+journalctl logged;
___
Jan 11 20:40:10 USERNAME kernel: kauditd_printk_skb: 620 callbacks suppressed

Jan 11 20:40:10 USERNAME kernel: audit: type=1400 audit(1768128010.915:77084): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/home/USERNAME/" pid=21516 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=1000

Jan 11 20:40:10 USERNAME kernel: audit: type=1400 audit(1768128010.915:77085): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/etc/nsswitch.conf" pid=21516 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:40:10 USERNAME kernel: audit: type=1400 audit(1768128010.915:77086): apparmor="ALLOWED" operation="open" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/etc/passwd" pid=21516 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:40:10 USERNAME kernel: audit: type=1400 audit(1768128010.915:77087): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/etc/passwd" pid=21516 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:40:10 USERNAME kernel: audit: type=1400 audit(1768128010.947:77088): apparmor="ALLOWED" operation="open" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/tmp/sv9lqrla.tmp/" pid=21516 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=1000

Jan 11 20:40:10 USERNAME kernel: audit: type=1400 audit(1768128010.947:77089): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/tmp/sv9lqrla.tmp/" pid=21516 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=1000

Jan 11 20:40:10 USERNAME kernel: audit: type=1400 audit(1768128010.947:77090): apparmor="ALLOWED" operation="mknod" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/tmp/sv9lqrla.tmp/sv9lrkqw.tmp" pid=21516 comm="soffice.bin" requested_mask="c" denied_mask="c" fsuid=1000 ouid=1000

Jan 11 20:40:10 USERNAME kernel: audit: type=1400 audit(1768128010.947:77091): apparmor="ALLOWED" operation="open" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/tmp/sv9lqrla.tmp/sv9lrkqw.tmp" pid=21516 comm="soffice.bin" requested_mask="wrc" denied_mask="wrc" fsuid=1000 ouid=1000

Jan 11 20:40:10 USERNAME kernel: audit: type=1400 audit(1768128010.947:77092): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/tmp/sv9lqrla.tmp/sv9lrkqw.tmp" pid=21516 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=1000

Jan 11 20:40:10 USERNAME kernel: audit: type=1400 audit(1768128010.947:77093): apparmor="ALLOWED" operation="open" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/tmp/sv9lqrla.tmp/sv9lrkqw.tmp" pid=21516 comm="soffice.bin" requested_mask="wr" denied_mask="wr" fsuid=1000 ouid=1000
___

During saving with the known good working version, apparmor+journalctl logged;
___
Jan 11 20:55:02 USERNAME kernel: kauditd_printk_skb: 192 callbacks suppressed

Jan 11 20:55:02 USERNAME kernel: audit: type=1400 audit(1768128902.607:95459): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.db" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:02 USERNAME kernel: audit: type=1400 audit(1768128902.607:95460): apparmor="ALLOWED" operation="open" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.db" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:02 USERNAME kernel: audit: type=1400 audit(1768128902.607:95461): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.db" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:02 USERNAME kernel: audit: type=1400 audit(1768128902.607:95462): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.ht" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:02 USERNAME kernel: audit: type=1400 audit(1768128902.607:95463): apparmor="ALLOWED" operation="open" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.ht" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:02 USERNAME kernel: audit: type=1400 audit(1768128902.607:95464): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.ht" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:02 USERNAME kernel: audit: type=1400 audit(1768128902.759:95465): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.db" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:02 USERNAME kernel: audit: type=1400 audit(1768128902.759:95466): apparmor="ALLOWED" operation="open" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.db" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:02 USERNAME kernel: audit: type=1400 audit(1768128902.759:95467): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.db" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:02 USERNAME kernel: audit: type=1400 audit(1768128902.759:95468): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.ht" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:07 USERNAME kernel: kauditd_printk_skb: 435 callbacks suppressed

Jan 11 20:55:07 USERNAME kernel: audit: type=1400 audit(1768128907.611:95904): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/home/USERNAME/FILE_IN_MY_HOME_DIRECTORY/" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=1000

Jan 11 20:55:07 USERNAME kernel: audit: type=1400 audit(1768128907.611:95905): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/home/USERNAME/FILE_IN_MY_HOME_DIRECTORY/" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=1000

Jan 11 20:55:07 USERNAME kernel: audit: type=1400 audit(1768128907.611:95906): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/home/USERNAME/FILE_IN_MY_HOME_DIRECTORY" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=1000

Jan 11 20:55:07 USERNAME kernel: audit: type=1400 audit(1768128907.611:95907): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name=2F686F6D652F73747564656E742F5073616C6D203930 pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=1000

Jan 11 20:55:07 USERNAME kernel: audit: type=1400 audit(1768128907.611:95908): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/home/USERNAME/FILE_IN_MY_HOME_DIRECTORY" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=1000

Jan 11 20:55:07 USERNAME kernel: audit: type=1400 audit(1768128907.611:95909): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/home/USERNAME/FILE_IN_MY_HOME_DIRECTORY" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=1000

Jan 11 20:55:07 USERNAME kernel: audit: type=1400 audit(1768128907.611:95910): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/home/USERNAME/FOLDER_IN_MY_HOME_DIRECTORY/" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=1000

Jan 11 20:55:07 USERNAME kernel: audit: type=1400 audit(1768128907.611:95911): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name=2F686F6D652F73747564656E742F6920616D20616E20656C65637472696369616E202D2041736B2042726176652E706466 pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=1000

Jan 11 20:55:07 USERNAME kernel: audit: type=1400 audit(1768128907.611:95912): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/home/USERNAME/FILE_IN_MY_HOME_DIRECTORY" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=1000

Jan 11 20:55:07 USERNAME kernel: audit: type=1400 audit(1768128907.611:95913): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/home/USERNAME/FILE_IN_MY_HOME_DIRECTORY" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=1000

Jan 11 20:55:32 USERNAME kernel: kauditd_printk_skb: 610 callbacks suppressed

Jan 11 20:55:32 USERNAME kernel: audit: type=1400 audit(1768128932.678:96524): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.db" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:32 USERNAME kernel: audit: type=1400 audit(1768128932.678:96525): apparmor="ALLOWED" operation="open" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.db" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:32 USERNAME kernel: audit: type=1400 audit(1768128932.678:96526): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.db" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:32 USERNAME kernel: audit: type=1400 audit(1768128932.678:96527): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.ht" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:32 USERNAME kernel: audit: type=1400 audit(1768128932.678:96528): apparmor="ALLOWED" operation="open" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.ht" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:32 USERNAME kernel: audit: type=1400 audit(1768128932.678:96529): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.ht" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:32 USERNAME kernel: audit: type=1400 audit(1768128932.694:96530): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.db" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:32 USERNAME kernel: audit: type=1400 audit(1768128932.694:96531): apparmor="ALLOWED" operation="open" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.db" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:32 USERNAME kernel: audit: type=1400 audit(1768128932.694:96532): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.db" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:32 USERNAME kernel: audit: type=1400 audit(1768128932.694:96533): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/opt/openoffice4/help/en/swriter.ht" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:39 USERNAME kernel: kauditd_printk_skb: 135 callbacks suppressed

Jan 11 20:55:39 USERNAME kernel: audit: type=1400 audit(1768128939.342:96669): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/etc/nsswitch.conf" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:39 USERNAME kernel: audit: type=1400 audit(1768128939.342:96670): apparmor="ALLOWED" operation="open" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/etc/passwd" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:39 USERNAME kernel: audit: type=1400 audit(1768128939.342:96671): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/etc/passwd" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=0

Jan 11 20:55:40 USERNAME kernel: audit: type=1400 audit(1768128940.342:96672): apparmor="ALLOWED" operation="mknod" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/home/USERNAME/.config/.openoffice/4/user/n3ZgV9" pid=22366 comm="soffice.bin" requested_mask="c" denied_mask="c" fsuid=1000 ouid=1000

Jan 11 20:55:40 USERNAME kernel: audit: type=1400 audit(1768128940.342:96673): apparmor="ALLOWED" operation="open" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/home/USERNAME/.config/.openoffice/4/user/n3ZgV9" pid=22366 comm="soffice.bin" requested_mask="wrc" denied_mask="wrc" fsuid=1000 ouid=1000

Jan 11 20:55:40 USERNAME kernel: audit: type=1400 audit(1768128940.342:96674): apparmor="ALLOWED" operation="getattr" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/home/USERNAME/.config/.openoffice/4/user/n3ZgV9" pid=22366 comm="soffice.bin" requested_mask="r" denied_mask="r" fsuid=1000 ouid=1000

Jan 11 20:55:40 USERNAME kernel: audit: type=1400 audit(1768128940.342:96675): apparmor="ALLOWED" operation="file_lock" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/home/USERNAME/.config/.openoffice/4/user/n3ZgV9" pid=22366 comm="soffice.bin" requested_mask="wk" denied_mask="wk" fsuid=1000 ouid=1000

Jan 11 20:55:40 USERNAME kernel: audit: type=1400 audit(1768128940.346:96676): apparmor="ALLOWED" operation="rename_src" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/home/USERNAME/.config/.openoffice/4/user/n3ZgV9" pid=22366 comm="soffice.bin" requested_mask="wrd" denied_mask="wrd" fsuid=1000 ouid=1000

Jan 11 20:55:40 USERNAME kernel: audit: type=1400 audit(1768128940.346:96677): apparmor="ALLOWED" operation="rename_dest" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/home/USERNAME/.config/.openoffice/4/user/registrymodifications.xcu" pid=22366 comm="soffice.bin" requested_mask="wc" denied_mask="wc" fsuid=1000 ouid=1000

Jan 11 20:55:55 USERNAME kernel: audit: type=1400 audit(1768128955.326:96678): apparmor="ALLOWED" operation="signal" class="signal" profile="soffice" pid=22350 comm="soffice" requested_mask="send" denied_mask="send" signal=int peer="soffice"

Jan 11 20:55:55 USERNAME kernel: audit: type=1400 audit(1768128955.326:96679): apparmor="ALLOWED" operation="signal" class="signal" profile="soffice" pid=22350 comm="soffice" requested_mask="receive" denied_mask="receive" signal=int peer="soffice"

Jan 11 20:55:55 USERNAME kernel: audit: type=1400 audit(1768128955.326:96680): apparmor="ALLOWED" operation="unlink" class="file" profile="soffice//null-/opt/openoffice4/program/soffice.bin" name="/tmp/OSL_PIPE_1000_SingleOfficeIPC_dc60b6c0421a28528e98ba19dfde6f" pid=22366 comm="soffice.bin" requested_mask="d" denied_mask="d" fsuid=1000 ouid=1000
___

I dont know what this known good version was doing looking at just some of my files and folders in my home directory, FILE_IN_MY_HOME_DIRECTORY and FOLDER_IN_MY_HOME_DIRECTORY, but am sure it had some good reason.

Anyway, might be easier to see what is going on with this information, again, get a nice version of the general error version here (https://pastebin.com/0PVZVQ4S) and a nice version of the known working version here (https://pastebin.com/8K8gdi69).

All the best,
Dean


________________________________
From: Dean Webber <de...@student.uq.edu.au>
Sent: 11 January 2026 15:59
To: dev@openoffice.apache.org <de...@openoffice.apache.org>
Subject: Re: SELinux [was: WG: There is only one problem in OpenOffice....

Okay, I had a good look. It seems I was unable to find any workaround. It still resulted in the general error when saving.

Below is my response, please excuse the "extra" information, it was of help at the time, and search results are no longer great in recent years.

Interesting, I will have a look. Out of interest, I wanted to know what SELinux was before getting into this.

"Security-Enhanced Linux (SELinux) is a Linux kernel security module that provides a mechanism for supporting access control security policies, including mandatory access controls (MAC). Originally developed by the United States National Security Agency (NSA) and released to the open source community under the GNU GPL on 22 December 2000, SELinux was merged into the Linux kernel mainline in the 2.6 series. It operates by enforcing security policies that confine user programs and system services, limiting their privileges to the minimum required to function, thereby reducing the potential damage from flawed or compromised applications." - Brave (Kit)

Recommended reading -> https://debian-handbook.info/browse/stable/sect.selinux.html

To start off I uninstalled my current openoffice. Extracted, Apache_OpenOffice_4.1.16_Linux_x86-64_install-deb_en-GB.tar.gz (fb4d7a4fe86b15d05e051a7e1b969ee031b858e0e49e04c7af429efdf29b6c97) [sha256sum]
sudo dpkg -i *.deb; cd desktop-integration/; sudo dpkg -i *.deb
I confirmed that saving a document resulted in general error.

I tried to install execstack, but it seems that it is actually unavailable.

"The execstack utility is not available as a package in Debian 13 (Trixie), as it has been removed from the distribution. This removal is part of a broader trend where newer versions of glibc, such as version 2.41 introduced in Debian 13 on March 13, 2025, no longer allow libraries requiring executable stacks to be loaded if the stack is not already executable. As a result, attempting to use execstack to modify or query ELF binaries that require executable stacks will fail, and the tool itself is no longer provided by the system.

For users needing to manage executable stack requirements, alternatives exist. A workaround involves using patchelf to clear the executable stack flag, though this method can corrupt files and cause server crashes if not applied carefully. Another approach is to use a script based on readelf to query whether an ELF binary has an executable stack, which can serve as a replacement for the execstack -q functionality. However, these methods are not recommended for production environments without proper testing." - Brave (Kit)

So instead, I decided to just disable whatever security features were available on the system. Further investigation turned out that SELinux is permanently disabled on my system. This is confirmed by running "sestatus". I was unable to enable it.

"You may not be able to enable SELinux on Debian 13 because: 1. AppArmor is Enabled by Default. Debian uses AppArmor as its default Linux Security Module (LSM), and only one LSM can be active at a time. SELinux cannot be enabled while AppArmor is running." - Brave, (Kit)

This was confirmed by running, "cat /sys/module/apparmor/parameters/enabled", which results in "Y", meaning apparmor is running.

So I disabled apparmor, using "sudo systemctl stop apparmor; sudo systemctl disable apparmor", restarted the system.

I opened writer and this still resulted in the general error when saving.. Not too sure where to go from there.

Unless, it has something to do with the "newer versions of glibc, such as version 2.41 introduced in Debian 13 on March 13, 2025, no longer allow libraries requiring executable stacks to be loaded if the stack is not already executable ... attempting to use execstack to modify or query ELF binaries that require executable stacks will fail "?

Next weekend I could reimage this machine to Debian 12, to see if the issue persists on the older operating system.

Hopefully this is of some help,
Dean
________________________________
From: Arrigo Marchiori <ar...@apache.org>
Sent: 11 January 2026 03:43
To: dev@openoffice.apache.org <de...@openoffice.apache.org>
Subject: SELinux [was: WG: There is only one problem in OpenOffice....

Dear All,

please allow me this top-posting.

The "General error" message also appears on the latest openSUSE
distribution (codename Leap 16.0).

Apparently, it is due to SELinux.

It did not appear on previous openSUSE versions because SELinux was
not enabled by default until this release.

I suppose the same could apply to the latest Fedora and Debian
releases.

The solution consists in enabling the "executable stack flag" of the
OpenOffice binary, with the following command:

$ sudo execstack -s /opt/openoffice4/program/soffice.bin

Note that the execstack program may not be installed by default.

A proof of this cause is found using the "ausearch" command as root
(i.e. you may need sudo to launch it as normal user):

# ausearch -m avc,user_avc,selinux_err,user_selinux_err -ts boot

On my system, the output contained lines like the following:

time->Sat Jan 10 18:23:08 2026
type=AVC msg=audit(1768065788.356:276): avc:  denied  { execstack } for  pid=5517 comm="soffice.bin" scontext=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 tcontext=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 tclass=process permissive=0

Dean, Wolfgang, could you please check if this command solves the
"General error" message on your systems? Repeating it below:

$ sudo execstack -s /opt/openoffice4/program/soffice.bin

Thank you and best regards.

On Thu, Jan 01, 2026 at 01:07:13AM +0000, Dean Webber wrote:

> Dear Wolfgang,
>
> I am not a developer, but had the same file issue when I upgraded to Debian 13.
>
> One of the developers suggested I use the releases here, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/ , which is an upstream build which I suspect has a fix for this issue.
>
> Specifically for you, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_install-rpm_de.tar.gz and https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_langpack-rpm_de.tar.gz
>
> It worked for me, perhaps it will work for you also.
>
> Happy New Year.
>
> Faithfully,
> Dean
>
> Translation en -> german
>
> Lieber Wolfgang,
>
> Ich bin kein Entwickler, hatte aber beim Upgrade auf Debian 13 das gleiche Dateiproblem.
>
> Einer der Entwickler schlug vor, die Releases hier zu verwenden, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de / , Das ist ein Upstream-Build, von dem ich vermute, dass es eine Lösung für dieses Problem gibt.
>
> Speziell für Sie, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_install-rpm_de.tar.gz und https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_langpack-rpm_de.tar.gz
>
> Es hat bei mir funktioniert, vielleicht funktioniert es auch bei dir.
>
> Frohes neues Jahr.
>
> Treu,
> Dean
>
> ________________________________
> From: Wolfgang Hinzmann <Wo...@protonmail.ch>
> Sent: 31 December 2025 02:12
> To: dev@ <de...@openoffice.apache.org>
> Subject: Re: WG: There is only one problem in OpenOffice....
>
> Hello Marcus,
>
> many thanks for these hints again.
>
> I tried both:
>
> As first step I renamed the folder
>
> $HOME/.openoffice.org/4/user/basic"
>
> After that, I started my computer new.
> When I started now OpenOffice, I got several messages like
> "Cannot load [file...].....".
> Despite this fact, I was able to create a new textfile and could save it.
> As usual with the "Error message".
>
> Because of this errors I tried the second step:
> I renamed the whole folder - as You described:
> $HOME/.openoffice.org
>
> Then I started my computer new,
> and after that I started OpenOffice.
> A window opened, and I was asked to put my whole Name, First name and family name,
> and an abbreviation for it. By clicking then the "Coninue-Button"
> OpenOffice opened, and I had the choice between the tools like Writer, Calculation...
> I opened writer, opened a new document, and then saving it:
> "The error message appears again" - as I described it.
>
> In effect, both steps - renaming the basic-folder, and after that,
> renaming the "$HOME/.openoffice.org"- folder,
> did not change the error away...
>
> So I decided, to change these system-folders in its originial state
> by renaming it, and deleting the new created ones.
>
> If You are interested in the error-messages, I have got,
> I would repeat the steps.
>
> I had really very great hope. Especially with the second step,
> because it started with the "Welcome-Window", and asking me
> for my name and its abbreviation.
> Strange. Very strange.
>
> Despite of this:
> I wish You a very good new year 2026!
>
> Best wishes Wolfgang
>
> Gesendet mit [Proton Mail](https://proton.me/mail/home): Ein sicherer E-Mail-Dienst.
>
> Marcus marcus.mail@wtnet.de schrieb am Montag, 29. Dezember 2025 um 16:15:
>
> > Hi Wolfgang,
> >
> > first of all you can try to reset your Basic settings. The following
> > will reset only the Basic things (e.g., macros) but nothing more:
> >
> > - close all AOO documents and windows
> > - open a terminal
> > - rename the Basic directory "$HOME/.openoffice.org/4/user/basic"
> > - cd ~
> > - mv .openoffice.org/4/user/basic
> > .openoffice.org/4/user/basic_maybe_broken
> > - start AOO new
> > - try if the error message is still showing
> >
> > If this will not help, try to reset the complete AOO profile:
> >
> > - close all AOO documents and windows
> > - open a terminal
> > - rename the complete profile directory "$HOME/.openoffice.org"
> > - cd ~
> > - mv .openoffice.org .openoffice.org_maybe_broken
> > - start AOO new
> > - try if the error message is still showing
> >
> > Please note:
> > You should know that all your customized settings will be gone. They all
> > need to be setup again.
> >
> > Please be carefully:
> > Rename the profile directory. You can delete it anytime later when it is
> > sure that the profile was somehow broken.
> > Otherwise if the problem persists you can decide to go back to the old
> > profile directory by renaming it back with the following commands in the
> > terminal:
> >
> > cd ~
> > rm -rf .openoffice.org
> > mv .openoffice.org_maybe_broken .openoffice.org
> >
> > HTH
> >
> > Marcus
> >
> > Am 26.12.25 um 10:40 schrieb Wolfgang Hinzmann:
> >
> >> as You suggested, I will try OpenOffice, especially Writer,
> >> with a new profile.
> >> Because I never needed special settings and other useful extensions so far.
> >> The only question for me in that wayis:
> >> How can I realize that?
> >> If You get other hints too, I would be very glad.
> >>
> >> Marcus marcus.mail@wtnet.de schrieb am Dienstag, 23. Dezember 2025 um 13:45:
> >>
> >>> Am 21.12.25 um 18:09 schrieb Wolfgang Hinzmann:
> >>>
> >>>> /as You suggested, I opened the terminal, and /
> >>>> /typed then the command "soffice"./
> >>>> /After that, the main-menu of OpenOffice started./
> >>>> /There, I go to "Open last files", and opened the "Testfile01.odt"./
> >>>> /After that I got the message:/
> >>>>
> >>>> /image.png/
> >>>> /
> >>>> /In the_terminal-window_, there are the following messages:/
> >>>> /wolfganghinzmann@fedora:~$ soffice/
> >>>> /javaldx: Could not find a Java Runtime Environment!/
> >>>> /
> >>>> /That's all./
> >>>
> >>> than I (you) need help from others as I've no idea at the moment what
> >>> could be checked next.
> >>>
> >>> @All:
> >>> Has anyone more hints? Thanks.
> >>>
> >>> Last resort:
> >>> There is always the possibility to start from scratch with a new AOO
> >>> profile. But that means you would loose all your settings that you have
> >>> done up to know (customized toolbars, menus, keyboard shortcuts, see
> >>> menu "Tools - Customize" or in German: "Extras - Anpassen" plus some
> >>> more settings in other locations, e.g., "Format - AutoCorrext -
> >>> AutoCorrect Optionen").
> >>>
> >>> So, it depends on you if you want to try it.
> >>>
> >>>> Marcus marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de schrieb am
> >>>> Sonntag, 21. Dezember 2025 um 14:44:
> >>>>
> >>>> Am 21.12.25 um 14:12 schrieb Wolfgang Hinzmann:
> >>>>
> >>>> /I don't use macros in my documents./
> >>>>
> >>>> OK, a possible root cause less.
> >>>>
> >>>> /Question:/
> >>>> /You suggested, I should try to start Writer from Terminal./
> >>>> /How can I do this? I am really not so familiar with all these
> >>>> commands./
> >>>>
> >>>> that should be simple. You just open the terminal and enter the
> >>>> following command:
> >>>>
> >>>> soffice
> >>>>
> >>>> If this results in an unknown command, then please use it with the full
> >>>> path. It depends where you have installed it in your system. E.g.,
> >>>>
> >>>> /opt/openoffice4/program/soffice
> >>>>
> >>>> The output until the error message appears could be helpful. Fore this
> >>>> purpose start OpenOffice and then open a document that shows that error
> >>>> message. After that please post the complete text here.
> >>>>
> >>>> Thanks
> >>>>
> >>>> Marcus
> >>>>
> >>>> Gesendet mit Proton Mail https://proton.me/mail/home: <https://
> >>>> proton.me/mail/home:> Ein sicherer E-
> >>>> Mail-Dienst.
> >>>>
> >>>> Marcus marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de
> >>>> mailto:marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de schrieb am
> >>>> Sonntag, 21. Dezember 2025 um 10:05:
> >>>>
> >>>> Am 20.12.25 um 16:32 schrieb Wolfgang Hinzmann:
> >>>>
> >>>> great that it is now installing and working.
> >>>>
> >>>> /There is only one problem with OpenOffice, I had already in the/
> >>>> /earlier version too:/
> >>>> /By opening an odt-file with Writer I receive always the error
> >>>> message/
> >>>> / "Allgemeiner Fehler"./
> >>>> /image.png
> >>>> /
> >>>> /I had the hope, that this message will disappear by the
> >>>> installation of
> >>>> the new version./
> >>>> /But it doesn't./
> >>>> /
> >>>> /But what is the solution for it?/
> >>>> /After pressing the button "OK" I am able to work normally with
> >>>> the file
> >>>> I opened:/
> >>>> /Changing it and saving it./
> >>>>
> >>>> do you have macros inside the document? Please have a look via
> >>>> the menu
> >>>> "Tools - Macros - Organize Macros - Basic" (German: "Extras -
> >>>> Makros -
> >>>> Makros verwalten - Basic")?
> >>>> If so, it's up to you to decide if you still need them - then
> >>>> you have
> >>>> to adjust parts to run correctly. Then there is good chance that the
> >>>> document will open without error messages. Or to delete them.
> >>>>
> >>>> Please make sure you have made a backup before deleting things
> >>>> that you
> >>>> still need.

--
Arrigo

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org


Re: SELinux [was: WG: There is only one problem in OpenOffice....

Posted by Dean Webber <de...@student.uq.edu.au>.
Okay, I had a good look. It seems I was unable to find any workaround. It still resulted in the general error when saving.

Below is my response, please excuse the "extra" information, it was of help at the time, and search results are no longer great in recent years.

Interesting, I will have a look. Out of interest, I wanted to know what SELinux was before getting into this.

"Security-Enhanced Linux (SELinux) is a Linux kernel security module that provides a mechanism for supporting access control security policies, including mandatory access controls (MAC). Originally developed by the United States National Security Agency (NSA) and released to the open source community under the GNU GPL on 22 December 2000, SELinux was merged into the Linux kernel mainline in the 2.6 series. It operates by enforcing security policies that confine user programs and system services, limiting their privileges to the minimum required to function, thereby reducing the potential damage from flawed or compromised applications." - Brave (Kit)

Recommended reading -> https://debian-handbook.info/browse/stable/sect.selinux.html

To start off I uninstalled my current openoffice. Extracted, Apache_OpenOffice_4.1.16_Linux_x86-64_install-deb_en-GB.tar.gz (fb4d7a4fe86b15d05e051a7e1b969ee031b858e0e49e04c7af429efdf29b6c97) [sha256sum]
sudo dpkg -i *.deb; cd desktop-integration/; sudo dpkg -i *.deb
I confirmed that saving a document resulted in general error.

I tried to install execstack, but it seems that it is actually unavailable.

"The execstack utility is not available as a package in Debian 13 (Trixie), as it has been removed from the distribution. This removal is part of a broader trend where newer versions of glibc, such as version 2.41 introduced in Debian 13 on March 13, 2025, no longer allow libraries requiring executable stacks to be loaded if the stack is not already executable. As a result, attempting to use execstack to modify or query ELF binaries that require executable stacks will fail, and the tool itself is no longer provided by the system.

For users needing to manage executable stack requirements, alternatives exist. A workaround involves using patchelf to clear the executable stack flag, though this method can corrupt files and cause server crashes if not applied carefully. Another approach is to use a script based on readelf to query whether an ELF binary has an executable stack, which can serve as a replacement for the execstack -q functionality. However, these methods are not recommended for production environments without proper testing." - Brave (Kit)

So instead, I decided to just disable whatever security features were available on the system. Further investigation turned out that SELinux is permanently disabled on my system. This is confirmed by running "sestatus". I was unable to enable it.

"You may not be able to enable SELinux on Debian 13 because: 1. AppArmor is Enabled by Default. Debian uses AppArmor as its default Linux Security Module (LSM), and only one LSM can be active at a time. SELinux cannot be enabled while AppArmor is running." - Brave, (Kit)

This was confirmed by running, "cat /sys/module/apparmor/parameters/enabled", which results in "Y", meaning apparmor is running.

So I disabled apparmor, using "sudo systemctl stop apparmor; sudo systemctl disable apparmor", restarted the system.

I opened writer and this still resulted in the general error when saving.. Not too sure where to go from there.

Unless, it has something to do with the "newer versions of glibc, such as version 2.41 introduced in Debian 13 on March 13, 2025, no longer allow libraries requiring executable stacks to be loaded if the stack is not already executable ... attempting to use execstack to modify or query ELF binaries that require executable stacks will fail "?

Next weekend I could reimage this machine to Debian 12, to see if the issue persists on the older operating system.

Hopefully this is of some help,
Dean
________________________________
From: Arrigo Marchiori <ar...@apache.org>
Sent: 11 January 2026 03:43
To: dev@openoffice.apache.org <de...@openoffice.apache.org>
Subject: SELinux [was: WG: There is only one problem in OpenOffice....

Dear All,

please allow me this top-posting.

The "General error" message also appears on the latest openSUSE
distribution (codename Leap 16.0).

Apparently, it is due to SELinux.

It did not appear on previous openSUSE versions because SELinux was
not enabled by default until this release.

I suppose the same could apply to the latest Fedora and Debian
releases.

The solution consists in enabling the "executable stack flag" of the
OpenOffice binary, with the following command:

$ sudo execstack -s /opt/openoffice4/program/soffice.bin

Note that the execstack program may not be installed by default.

A proof of this cause is found using the "ausearch" command as root
(i.e. you may need sudo to launch it as normal user):

# ausearch -m avc,user_avc,selinux_err,user_selinux_err -ts boot

On my system, the output contained lines like the following:

time->Sat Jan 10 18:23:08 2026
type=AVC msg=audit(1768065788.356:276): avc:  denied  { execstack } for  pid=5517 comm="soffice.bin" scontext=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 tcontext=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 tclass=process permissive=0

Dean, Wolfgang, could you please check if this command solves the
"General error" message on your systems? Repeating it below:

$ sudo execstack -s /opt/openoffice4/program/soffice.bin

Thank you and best regards.

On Thu, Jan 01, 2026 at 01:07:13AM +0000, Dean Webber wrote:

> Dear Wolfgang,
>
> I am not a developer, but had the same file issue when I upgraded to Debian 13.
>
> One of the developers suggested I use the releases here, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/ , which is an upstream build which I suspect has a fix for this issue.
>
> Specifically for you, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_install-rpm_de.tar.gz and https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_langpack-rpm_de.tar.gz
>
> It worked for me, perhaps it will work for you also.
>
> Happy New Year.
>
> Faithfully,
> Dean
>
> Translation en -> german
>
> Lieber Wolfgang,
>
> Ich bin kein Entwickler, hatte aber beim Upgrade auf Debian 13 das gleiche Dateiproblem.
>
> Einer der Entwickler schlug vor, die Releases hier zu verwenden, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de / , Das ist ein Upstream-Build, von dem ich vermute, dass es eine Lösung für dieses Problem gibt.
>
> Speziell für Sie, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_install-rpm_de.tar.gz und https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_langpack-rpm_de.tar.gz
>
> Es hat bei mir funktioniert, vielleicht funktioniert es auch bei dir.
>
> Frohes neues Jahr.
>
> Treu,
> Dean
>
> ________________________________
> From: Wolfgang Hinzmann <Wo...@protonmail.ch>
> Sent: 31 December 2025 02:12
> To: dev@ <de...@openoffice.apache.org>
> Subject: Re: WG: There is only one problem in OpenOffice....
>
> Hello Marcus,
>
> many thanks for these hints again.
>
> I tried both:
>
> As first step I renamed the folder
>
> $HOME/.openoffice.org/4/user/basic"
>
> After that, I started my computer new.
> When I started now OpenOffice, I got several messages like
> "Cannot load [file...].....".
> Despite this fact, I was able to create a new textfile and could save it.
> As usual with the "Error message".
>
> Because of this errors I tried the second step:
> I renamed the whole folder - as You described:
> $HOME/.openoffice.org
>
> Then I started my computer new,
> and after that I started OpenOffice.
> A window opened, and I was asked to put my whole Name, First name and family name,
> and an abbreviation for it. By clicking then the "Coninue-Button"
> OpenOffice opened, and I had the choice between the tools like Writer, Calculation...
> I opened writer, opened a new document, and then saving it:
> "The error message appears again" - as I described it.
>
> In effect, both steps - renaming the basic-folder, and after that,
> renaming the "$HOME/.openoffice.org"- folder,
> did not change the error away...
>
> So I decided, to change these system-folders in its originial state
> by renaming it, and deleting the new created ones.
>
> If You are interested in the error-messages, I have got,
> I would repeat the steps.
>
> I had really very great hope. Especially with the second step,
> because it started with the "Welcome-Window", and asking me
> for my name and its abbreviation.
> Strange. Very strange.
>
> Despite of this:
> I wish You a very good new year 2026!
>
> Best wishes Wolfgang
>
> Gesendet mit [Proton Mail](https://proton.me/mail/home): Ein sicherer E-Mail-Dienst.
>
> Marcus marcus.mail@wtnet.de schrieb am Montag, 29. Dezember 2025 um 16:15:
>
> > Hi Wolfgang,
> >
> > first of all you can try to reset your Basic settings. The following
> > will reset only the Basic things (e.g., macros) but nothing more:
> >
> > - close all AOO documents and windows
> > - open a terminal
> > - rename the Basic directory "$HOME/.openoffice.org/4/user/basic"
> > - cd ~
> > - mv .openoffice.org/4/user/basic
> > .openoffice.org/4/user/basic_maybe_broken
> > - start AOO new
> > - try if the error message is still showing
> >
> > If this will not help, try to reset the complete AOO profile:
> >
> > - close all AOO documents and windows
> > - open a terminal
> > - rename the complete profile directory "$HOME/.openoffice.org"
> > - cd ~
> > - mv .openoffice.org .openoffice.org_maybe_broken
> > - start AOO new
> > - try if the error message is still showing
> >
> > Please note:
> > You should know that all your customized settings will be gone. They all
> > need to be setup again.
> >
> > Please be carefully:
> > Rename the profile directory. You can delete it anytime later when it is
> > sure that the profile was somehow broken.
> > Otherwise if the problem persists you can decide to go back to the old
> > profile directory by renaming it back with the following commands in the
> > terminal:
> >
> > cd ~
> > rm -rf .openoffice.org
> > mv .openoffice.org_maybe_broken .openoffice.org
> >
> > HTH
> >
> > Marcus
> >
> > Am 26.12.25 um 10:40 schrieb Wolfgang Hinzmann:
> >
> >> as You suggested, I will try OpenOffice, especially Writer,
> >> with a new profile.
> >> Because I never needed special settings and other useful extensions so far.
> >> The only question for me in that wayis:
> >> How can I realize that?
> >> If You get other hints too, I would be very glad.
> >>
> >> Marcus marcus.mail@wtnet.de schrieb am Dienstag, 23. Dezember 2025 um 13:45:
> >>
> >>> Am 21.12.25 um 18:09 schrieb Wolfgang Hinzmann:
> >>>
> >>>> /as You suggested, I opened the terminal, and /
> >>>> /typed then the command "soffice"./
> >>>> /After that, the main-menu of OpenOffice started./
> >>>> /There, I go to "Open last files", and opened the "Testfile01.odt"./
> >>>> /After that I got the message:/
> >>>>
> >>>> /image.png/
> >>>> /
> >>>> /In the_terminal-window_, there are the following messages:/
> >>>> /wolfganghinzmann@fedora:~$ soffice/
> >>>> /javaldx: Could not find a Java Runtime Environment!/
> >>>> /
> >>>> /That's all./
> >>>
> >>> than I (you) need help from others as I've no idea at the moment what
> >>> could be checked next.
> >>>
> >>> @All:
> >>> Has anyone more hints? Thanks.
> >>>
> >>> Last resort:
> >>> There is always the possibility to start from scratch with a new AOO
> >>> profile. But that means you would loose all your settings that you have
> >>> done up to know (customized toolbars, menus, keyboard shortcuts, see
> >>> menu "Tools - Customize" or in German: "Extras - Anpassen" plus some
> >>> more settings in other locations, e.g., "Format - AutoCorrext -
> >>> AutoCorrect Optionen").
> >>>
> >>> So, it depends on you if you want to try it.
> >>>
> >>>> Marcus marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de schrieb am
> >>>> Sonntag, 21. Dezember 2025 um 14:44:
> >>>>
> >>>> Am 21.12.25 um 14:12 schrieb Wolfgang Hinzmann:
> >>>>
> >>>> /I don't use macros in my documents./
> >>>>
> >>>> OK, a possible root cause less.
> >>>>
> >>>> /Question:/
> >>>> /You suggested, I should try to start Writer from Terminal./
> >>>> /How can I do this? I am really not so familiar with all these
> >>>> commands./
> >>>>
> >>>> that should be simple. You just open the terminal and enter the
> >>>> following command:
> >>>>
> >>>> soffice
> >>>>
> >>>> If this results in an unknown command, then please use it with the full
> >>>> path. It depends where you have installed it in your system. E.g.,
> >>>>
> >>>> /opt/openoffice4/program/soffice
> >>>>
> >>>> The output until the error message appears could be helpful. Fore this
> >>>> purpose start OpenOffice and then open a document that shows that error
> >>>> message. After that please post the complete text here.
> >>>>
> >>>> Thanks
> >>>>
> >>>> Marcus
> >>>>
> >>>> Gesendet mit Proton Mail https://proton.me/mail/home: <https://
> >>>> proton.me/mail/home:> Ein sicherer E-
> >>>> Mail-Dienst.
> >>>>
> >>>> Marcus marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de
> >>>> mailto:marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de schrieb am
> >>>> Sonntag, 21. Dezember 2025 um 10:05:
> >>>>
> >>>> Am 20.12.25 um 16:32 schrieb Wolfgang Hinzmann:
> >>>>
> >>>> great that it is now installing and working.
> >>>>
> >>>> /There is only one problem with OpenOffice, I had already in the/
> >>>> /earlier version too:/
> >>>> /By opening an odt-file with Writer I receive always the error
> >>>> message/
> >>>> / "Allgemeiner Fehler"./
> >>>> /image.png
> >>>> /
> >>>> /I had the hope, that this message will disappear by the
> >>>> installation of
> >>>> the new version./
> >>>> /But it doesn't./
> >>>> /
> >>>> /But what is the solution for it?/
> >>>> /After pressing the button "OK" I am able to work normally with
> >>>> the file
> >>>> I opened:/
> >>>> /Changing it and saving it./
> >>>>
> >>>> do you have macros inside the document? Please have a look via
> >>>> the menu
> >>>> "Tools - Macros - Organize Macros - Basic" (German: "Extras -
> >>>> Makros -
> >>>> Makros verwalten - Basic")?
> >>>> If so, it's up to you to decide if you still need them - then
> >>>> you have
> >>>> to adjust parts to run correctly. Then there is good chance that the
> >>>> document will open without error messages. Or to delete them.
> >>>>
> >>>> Please make sure you have made a backup before deleting things
> >>>> that you
> >>>> still need.

--
Arrigo

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org


SELinux [was: WG: There is only one problem in OpenOffice....

Posted by Arrigo Marchiori <ar...@apache.org>.
Dear All,

please allow me this top-posting.

The "General error" message also appears on the latest openSUSE
distribution (codename Leap 16.0).

Apparently, it is due to SELinux.

It did not appear on previous openSUSE versions because SELinux was
not enabled by default until this release.

I suppose the same could apply to the latest Fedora and Debian
releases.

The solution consists in enabling the "executable stack flag" of the
OpenOffice binary, with the following command:

$ sudo execstack -s /opt/openoffice4/program/soffice.bin

Note that the execstack program may not be installed by default.

A proof of this cause is found using the "ausearch" command as root
(i.e. you may need sudo to launch it as normal user):

# ausearch -m avc,user_avc,selinux_err,user_selinux_err -ts boot

On my system, the output contained lines like the following:

time->Sat Jan 10 18:23:08 2026
type=AVC msg=audit(1768065788.356:276): avc:  denied  { execstack } for  pid=5517 comm="soffice.bin" scontext=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 tcontext=unconfined_u:unconfined_r:unconfined_t:s0-s0:c0.c1023 tclass=process permissive=0

Dean, Wolfgang, could you please check if this command solves the
"General error" message on your systems? Repeating it below:

$ sudo execstack -s /opt/openoffice4/program/soffice.bin

Thank you and best regards.

On Thu, Jan 01, 2026 at 01:07:13AM +0000, Dean Webber wrote:

> Dear Wolfgang,
> 
> I am not a developer, but had the same file issue when I upgraded to Debian 13.
> 
> One of the developers suggested I use the releases here, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/ , which is an upstream build which I suspect has a fix for this issue.
> 
> Specifically for you, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_install-rpm_de.tar.gz and https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_langpack-rpm_de.tar.gz
> 
> It worked for me, perhaps it will work for you also.
> 
> Happy New Year.
> 
> Faithfully,
> Dean
> 
> Translation en -> german
> 
> Lieber Wolfgang,
> 
> Ich bin kein Entwickler, hatte aber beim Upgrade auf Debian 13 das gleiche Dateiproblem.
> 
> Einer der Entwickler schlug vor, die Releases hier zu verwenden, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de / , Das ist ein Upstream-Build, von dem ich vermute, dass es eine Lösung für dieses Problem gibt.
> 
> Speziell für Sie, https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_install-rpm_de.tar.gz und https://dist.apache.org/repos/dist/dev/openoffice/4.2.0-Dev5/de/Apache_OpenOffice_4.2.0_Linux_x86-64_langpack-rpm_de.tar.gz
> 
> Es hat bei mir funktioniert, vielleicht funktioniert es auch bei dir.
> 
> Frohes neues Jahr.
> 
> Treu,
> Dean
> 
> ________________________________
> From: Wolfgang Hinzmann <Wo...@protonmail.ch>
> Sent: 31 December 2025 02:12
> To: dev@ <de...@openoffice.apache.org>
> Subject: Re: WG: There is only one problem in OpenOffice....
> 
> Hello Marcus,
> 
> many thanks for these hints again.
> 
> I tried both:
> 
> As first step I renamed the folder
> 
> $HOME/.openoffice.org/4/user/basic"
> 
> After that, I started my computer new.
> When I started now OpenOffice, I got several messages like
> "Cannot load [file...].....".
> Despite this fact, I was able to create a new textfile and could save it.
> As usual with the "Error message".
> 
> Because of this errors I tried the second step:
> I renamed the whole folder - as You described:
> $HOME/.openoffice.org
> 
> Then I started my computer new,
> and after that I started OpenOffice.
> A window opened, and I was asked to put my whole Name, First name and family name,
> and an abbreviation for it. By clicking then the "Coninue-Button"
> OpenOffice opened, and I had the choice between the tools like Writer, Calculation...
> I opened writer, opened a new document, and then saving it:
> "The error message appears again" - as I described it.
> 
> In effect, both steps - renaming the basic-folder, and after that,
> renaming the "$HOME/.openoffice.org"- folder,
> did not change the error away...
> 
> So I decided, to change these system-folders in its originial state
> by renaming it, and deleting the new created ones.
> 
> If You are interested in the error-messages, I have got,
> I would repeat the steps.
> 
> I had really very great hope. Especially with the second step,
> because it started with the "Welcome-Window", and asking me
> for my name and its abbreviation.
> Strange. Very strange.
> 
> Despite of this:
> I wish You a very good new year 2026!
> 
> Best wishes Wolfgang
> 
> Gesendet mit [Proton Mail](https://proton.me/mail/home): Ein sicherer E-Mail-Dienst.
> 
> Marcus marcus.mail@wtnet.de schrieb am Montag, 29. Dezember 2025 um 16:15:
> 
> > Hi Wolfgang,
> >
> > first of all you can try to reset your Basic settings. The following
> > will reset only the Basic things (e.g., macros) but nothing more:
> >
> > - close all AOO documents and windows
> > - open a terminal
> > - rename the Basic directory "$HOME/.openoffice.org/4/user/basic"
> > - cd ~
> > - mv .openoffice.org/4/user/basic
> > .openoffice.org/4/user/basic_maybe_broken
> > - start AOO new
> > - try if the error message is still showing
> >
> > If this will not help, try to reset the complete AOO profile:
> >
> > - close all AOO documents and windows
> > - open a terminal
> > - rename the complete profile directory "$HOME/.openoffice.org"
> > - cd ~
> > - mv .openoffice.org .openoffice.org_maybe_broken
> > - start AOO new
> > - try if the error message is still showing
> >
> > Please note:
> > You should know that all your customized settings will be gone. They all
> > need to be setup again.
> >
> > Please be carefully:
> > Rename the profile directory. You can delete it anytime later when it is
> > sure that the profile was somehow broken.
> > Otherwise if the problem persists you can decide to go back to the old
> > profile directory by renaming it back with the following commands in the
> > terminal:
> >
> > cd ~
> > rm -rf .openoffice.org
> > mv .openoffice.org_maybe_broken .openoffice.org
> >
> > HTH
> >
> > Marcus
> >
> > Am 26.12.25 um 10:40 schrieb Wolfgang Hinzmann:
> >
> >> as You suggested, I will try OpenOffice, especially Writer,
> >> with a new profile.
> >> Because I never needed special settings and other useful extensions so far.
> >> The only question for me in that wayis:
> >> How can I realize that?
> >> If You get other hints too, I would be very glad.
> >>
> >> Marcus marcus.mail@wtnet.de schrieb am Dienstag, 23. Dezember 2025 um 13:45:
> >>
> >>> Am 21.12.25 um 18:09 schrieb Wolfgang Hinzmann:
> >>>
> >>>> /as You suggested, I opened the terminal, and /
> >>>> /typed then the command "soffice"./
> >>>> /After that, the main-menu of OpenOffice started./
> >>>> /There, I go to "Open last files", and opened the "Testfile01.odt"./
> >>>> /After that I got the message:/
> >>>>
> >>>> /image.png/
> >>>> /
> >>>> /In the_terminal-window_, there are the following messages:/
> >>>> /wolfganghinzmann@fedora:~$ soffice/
> >>>> /javaldx: Could not find a Java Runtime Environment!/
> >>>> /
> >>>> /That's all./
> >>>
> >>> than I (you) need help from others as I've no idea at the moment what
> >>> could be checked next.
> >>>
> >>> @All:
> >>> Has anyone more hints? Thanks.
> >>>
> >>> Last resort:
> >>> There is always the possibility to start from scratch with a new AOO
> >>> profile. But that means you would loose all your settings that you have
> >>> done up to know (customized toolbars, menus, keyboard shortcuts, see
> >>> menu "Tools - Customize" or in German: "Extras - Anpassen" plus some
> >>> more settings in other locations, e.g., "Format - AutoCorrext -
> >>> AutoCorrect Optionen").
> >>>
> >>> So, it depends on you if you want to try it.
> >>>
> >>>> Marcus marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de schrieb am
> >>>> Sonntag, 21. Dezember 2025 um 14:44:
> >>>>
> >>>> Am 21.12.25 um 14:12 schrieb Wolfgang Hinzmann:
> >>>>
> >>>> /I don't use macros in my documents./
> >>>>
> >>>> OK, a possible root cause less.
> >>>>
> >>>> /Question:/
> >>>> /You suggested, I should try to start Writer from Terminal./
> >>>> /How can I do this? I am really not so familiar with all these
> >>>> commands./
> >>>>
> >>>> that should be simple. You just open the terminal and enter the
> >>>> following command:
> >>>>
> >>>> soffice
> >>>>
> >>>> If this results in an unknown command, then please use it with the full
> >>>> path. It depends where you have installed it in your system. E.g.,
> >>>>
> >>>> /opt/openoffice4/program/soffice
> >>>>
> >>>> The output until the error message appears could be helpful. Fore this
> >>>> purpose start OpenOffice and then open a document that shows that error
> >>>> message. After that please post the complete text here.
> >>>>
> >>>> Thanks
> >>>>
> >>>> Marcus
> >>>>
> >>>> Gesendet mit Proton Mail https://proton.me/mail/home: <https://
> >>>> proton.me/mail/home:> Ein sicherer E-
> >>>> Mail-Dienst.
> >>>>
> >>>> Marcus marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de
> >>>> mailto:marcus.mail@wtnet.de mailto:marcus.mail@wtnet.de schrieb am
> >>>> Sonntag, 21. Dezember 2025 um 10:05:
> >>>>
> >>>> Am 20.12.25 um 16:32 schrieb Wolfgang Hinzmann:
> >>>>
> >>>> great that it is now installing and working.
> >>>>
> >>>> /There is only one problem with OpenOffice, I had already in the/
> >>>> /earlier version too:/
> >>>> /By opening an odt-file with Writer I receive always the error
> >>>> message/
> >>>> / "Allgemeiner Fehler"./
> >>>> /image.png
> >>>> /
> >>>> /I had the hope, that this message will disappear by the
> >>>> installation of
> >>>> the new version./
> >>>> /But it doesn't./
> >>>> /
> >>>> /But what is the solution for it?/
> >>>> /After pressing the button "OK" I am able to work normally with
> >>>> the file
> >>>> I opened:/
> >>>> /Changing it and saving it./
> >>>>
> >>>> do you have macros inside the document? Please have a look via
> >>>> the menu
> >>>> "Tools - Macros - Organize Macros - Basic" (German: "Extras -
> >>>> Makros -
> >>>> Makros verwalten - Basic")?
> >>>> If so, it's up to you to decide if you still need them - then
> >>>> you have
> >>>> to adjust parts to run correctly. Then there is good chance that the
> >>>> document will open without error messages. Or to delete them.
> >>>>
> >>>> Please make sure you have made a backup before deleting things
> >>>> that you
> >>>> still need.

-- 
Arrigo

---------------------------------------------------------------------
To unsubscribe, e-mail: dev-unsubscribe@openoffice.apache.org
For additional commands, e-mail: dev-help@openoffice.apache.org