You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@harmony.apache.org by Stepan Mishura <st...@gmail.com> on 2007/10/04 07:48:46 UTC

[general][M3] release r580985 as M3 stable build

Hi folks,

Let's make a decision about the last milestone candidate r580985.
Shortly, as I can see there is only one issue that blocks its release
- HARMONY-4873

I'd suggest the next - we wait for objections to postpone the issue
resolution to M4 for 72 hours (starting from now) to let everyone to
speak up.
If there are no objections - we make the release. Also I think we
shouldn't keep all this time the trunk frozen so I'd suggest if there
are no objections during 24 hours we unfroze the trunk. If  after that
we get a vote for fixing the issue then a branch is created to make
the fix and testing.

Objections?

Thanks,
Stepan Mishura
Intel Enterprise Solutions Software Division

Re: [general][M3] release r580985 as M3 stable build

Posted by Mark Hindess <ma...@googlemail.com>.
On 4 October 2007 at 12:48, "Stepan Mishura" <st...@gmail.com> wrote:
> Hi folks,
> 
> Let's make a decision about the last milestone candidate r580985.
> Shortly, as I can see there is only one issue that blocks its release
> - HARMONY-4873
> 
> I'd suggest the next - we wait for objections to postpone the issue
> resolution to M4 for 72 hours (starting from now) to let everyone to
> speak up.
> If there are no objections - we make the release. Also I think we
> shouldn't keep all this time the trunk frozen so I'd suggest if there
> are no objections during 24 hours we unfroze the trunk. If  after that
> we get a vote for fixing the issue then a branch is created to make
> the fix and testing.
> 
> Objections?

None from me.  Sounds like an excellent plan.

FYI: I've placed Debian(stable) packages for r580985 at:

  http://people.apache.org/~hindessm/debian/

In case any wants to give them a try.  Not these were created using a
source tar.gz created by running 'ant bundle_src' in the federation
build.  So if there are issues/inconsistencies then they'll be issues
with the source tar.gz.

-Mark.


Re: [general][M3] release r580985 as M3 stable build

Posted by Gregory Shimansky <gs...@gmail.com>.
Stepan Mishura wrote:
> Hi folks,
> 
> Let's make a decision about the last milestone candidate r580985.
> Shortly, as I can see there is only one issue that blocks its release
> - HARMONY-4873
> 
> I'd suggest the next - we wait for objections to postpone the issue
> resolution to M4 for 72 hours (starting from now) to let everyone to
> speak up.

Fix for this bug is not straightforward. It appears that hardware 
exception handling on windows and linux became different in regard of 
exiting exception/signal context. On Linux signal context is not left 
before handling the exception as it is done on windows, and so it is 
necessary to rewrite Linux signal handler to fix this bug correctly. I 
don't think we can allow to do this in the code freeze stage.

I wrote a small and simple workaround for this bug that kind of 
"reverts" offending change for Linux. With this workaround test in the 
bug works. I'm testing the workaround patch currently on windows and 
linux to see if it doesn't break anything.

> If there are no objections - we make the release. Also I think we
> shouldn't keep all this time the trunk frozen so I'd suggest if there
> are no objections during 24 hours we unfroze the trunk. If  after that
> we get a vote for fixing the issue then a branch is created to make
> the fix and testing.
> 
> Objections?
> 
> Thanks,
> Stepan Mishura
> Intel Enterprise Solutions Software Division
> 


-- 
Gregory


Re: [general][M3] release r580985 as M3 stable build

Posted by Mikhail Loenko <ml...@gmail.com>.
The plan sounds fine for me in general and
I'm OK with postponing 4873 till M4

Thanks,
Mikhail

2007/10/4, Stepan Mishura <st...@gmail.com>:
> Hi folks,
>
> Let's make a decision about the last milestone candidate r580985.
> Shortly, as I can see there is only one issue that blocks its release
> - HARMONY-4873
>
> I'd suggest the next - we wait for objections to postpone the issue
> resolution to M4 for 72 hours (starting from now) to let everyone to
> speak up.
> If there are no objections - we make the release. Also I think we
> shouldn't keep all this time the trunk frozen so I'd suggest if there
> are no objections during 24 hours we unfroze the trunk. If  after that
> we get a vote for fixing the issue then a branch is created to make
> the fix and testing.
>
> Objections?
>
> Thanks,
> Stepan Mishura
> Intel Enterprise Solutions Software Division
>

Re: [general][M3] release r580985 as M3 stable build

Posted by Gregory Shimansky <gs...@gmail.com>.
Sian January wrote:
> Thanks Gregory.  So I guess the launcher is using JNI rather than
> java.exeto start Eclipse when the jre directory exists.
> 
> By the way, this was also failing for me with the IBM VME.  Does it make
> sense that it would have the same bug?

Hmm no. In this case it should be a different bug since IBM VME exports 
JNI_CreateJavaVM as "JNI_CreateJavaVM" (as sun does), not as 
"_JNI_CreateJavaVM@12" as DRLVM currently does. So it is something 
different. Maybe eclipse.exe actually uses java.exe when it finds 
eclipse/jre directory?...

> On 04/10/2007, Gregory Shimansky <gs...@gmail.com> wrote:
>> Sian January wrote:
>>> On 04/10/2007, Tim Ellison <t....@gmail.com> wrote:
>>>> Stepan Mishura wrote:
>>>>> Let's make a decision about the last milestone candidate r580985.
>>>>> Shortly, as I can see there is only one issue that blocks its release
>>>>> - HARMONY-4873
>>>>>
>>>>> I'd suggest the next - we wait for objections to postpone the issue
>>>>> resolution to M4 for 72 hours (starting from now) to let everyone to
>>>>> speak up.
>>>>> If there are no objections - we make the release. Also I think we
>>>>> shouldn't keep all this time the trunk frozen so I'd suggest if there
>>>>> are no objections during 24 hours we unfroze the trunk. If  after that
>>>>> we get a vote for fixing the issue then a branch is created to make
>>>>> the fix and testing.
>>>> Agreed, although it would be good to hear that the Eclipse 3.3launching
>>>> problem is fixed that Sian posted about on the thread "Re: [eclipse]
>>>> Eclipse 3.3 can not download updates".
>>>
>>> I've just done some more digging on this and it's only failing when you
>> copy
>>> the Harmony JRE to the eclipse/jre directory (which is new in Eclipse
>> 3.3).
>>> You can still add jre\bin to the path and run eclipse.exe like you did
>>> before.  So as it's not a regression and it could take a while to figure
>> out
>>> I would probably say it's not a blocking issue for M3.
>>>
>>> Next question is whether to raise a Harmony bug or an Eclipse bug...
>> There is already a bug about eclipse.exe not being able to launch
>> Harmony when DRLVM is used. It is HARMONY-4006. All patches there are
>> ready to committed after code freeze.
>>
>> BTW it is a windows only bug because of windows specific name
>> decorations for JNI_CreateJavaVM function. On Linux eclipse 3.3 should
>> work without any patches.
>>
>>>> (Sorry that I can't be of much help with fixing things at the moment,
>>>> I'm about to get on an airplane again.)
>>>>
>>>> Regards,
>>>> Tim
>>>>
>>>
>>>
>>
>> --
>> Gregory
>>
>>
> 
> 


-- 
Gregory


Re: [general][M3] release r580985 as M3 stable build

Posted by Sian January <si...@googlemail.com>.
Thanks Gregory.  So I guess the launcher is using JNI rather than
java.exeto start Eclipse when the jre directory exists.

By the way, this was also failing for me with the IBM VME.  Does it make
sense that it would have the same bug?


On 04/10/2007, Gregory Shimansky <gs...@gmail.com> wrote:
>
> Sian January wrote:
> > On 04/10/2007, Tim Ellison <t....@gmail.com> wrote:
> >> Stepan Mishura wrote:
> >>> Let's make a decision about the last milestone candidate r580985.
> >>> Shortly, as I can see there is only one issue that blocks its release
> >>> - HARMONY-4873
> >>>
> >>> I'd suggest the next - we wait for objections to postpone the issue
> >>> resolution to M4 for 72 hours (starting from now) to let everyone to
> >>> speak up.
> >>> If there are no objections - we make the release. Also I think we
> >>> shouldn't keep all this time the trunk frozen so I'd suggest if there
> >>> are no objections during 24 hours we unfroze the trunk. If  after that
> >>> we get a vote for fixing the issue then a branch is created to make
> >>> the fix and testing.
> >> Agreed, although it would be good to hear that the Eclipse 3.3launching
> >> problem is fixed that Sian posted about on the thread "Re: [eclipse]
> >> Eclipse 3.3 can not download updates".
> >
> >
> > I've just done some more digging on this and it's only failing when you
> copy
> > the Harmony JRE to the eclipse/jre directory (which is new in Eclipse
> 3.3).
> > You can still add jre\bin to the path and run eclipse.exe like you did
> > before.  So as it's not a regression and it could take a while to figure
> out
> > I would probably say it's not a blocking issue for M3.
> >
> > Next question is whether to raise a Harmony bug or an Eclipse bug...
>
> There is already a bug about eclipse.exe not being able to launch
> Harmony when DRLVM is used. It is HARMONY-4006. All patches there are
> ready to committed after code freeze.
>
> BTW it is a windows only bug because of windows specific name
> decorations for JNI_CreateJavaVM function. On Linux eclipse 3.3 should
> work without any patches.
>
> >> (Sorry that I can't be of much help with fixing things at the moment,
> >> I'm about to get on an airplane again.)
> >>
> >> Regards,
> >> Tim
> >>
> >
> >
> >
>
>
> --
> Gregory
>
>


-- 
Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number
741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU

Re: [general][M3] release r580985 as M3 stable build

Posted by Gregory Shimansky <gs...@gmail.com>.
Sian January wrote:
> On 04/10/2007, Tim Ellison <t....@gmail.com> wrote:
>> Stepan Mishura wrote:
>>> Let's make a decision about the last milestone candidate r580985.
>>> Shortly, as I can see there is only one issue that blocks its release
>>> - HARMONY-4873
>>>
>>> I'd suggest the next - we wait for objections to postpone the issue
>>> resolution to M4 for 72 hours (starting from now) to let everyone to
>>> speak up.
>>> If there are no objections - we make the release. Also I think we
>>> shouldn't keep all this time the trunk frozen so I'd suggest if there
>>> are no objections during 24 hours we unfroze the trunk. If  after that
>>> we get a vote for fixing the issue then a branch is created to make
>>> the fix and testing.
>> Agreed, although it would be good to hear that the Eclipse 3.3 launching
>> problem is fixed that Sian posted about on the thread "Re: [eclipse]
>> Eclipse 3.3 can not download updates".
> 
> 
> I've just done some more digging on this and it's only failing when you copy
> the Harmony JRE to the eclipse/jre directory (which is new in Eclipse 3.3).
> You can still add jre\bin to the path and run eclipse.exe like you did
> before.  So as it's not a regression and it could take a while to figure out
> I would probably say it's not a blocking issue for M3.
> 
> Next question is whether to raise a Harmony bug or an Eclipse bug...

There is already a bug about eclipse.exe not being able to launch 
Harmony when DRLVM is used. It is HARMONY-4006. All patches there are 
ready to committed after code freeze.

BTW it is a windows only bug because of windows specific name 
decorations for JNI_CreateJavaVM function. On Linux eclipse 3.3 should 
work without any patches.

>> (Sorry that I can't be of much help with fixing things at the moment,
>> I'm about to get on an airplane again.)
>>
>> Regards,
>> Tim
>>
> 
> 
> 


-- 
Gregory


Re: [general][M3] release r580985 as M3 stable build

Posted by Sian January <si...@googlemail.com>.
On 04/10/2007, Tim Ellison <t....@gmail.com> wrote:
>
> Stepan Mishura wrote:
> > Let's make a decision about the last milestone candidate r580985.
> > Shortly, as I can see there is only one issue that blocks its release
> > - HARMONY-4873
> >
> > I'd suggest the next - we wait for objections to postpone the issue
> > resolution to M4 for 72 hours (starting from now) to let everyone to
> > speak up.
> > If there are no objections - we make the release. Also I think we
> > shouldn't keep all this time the trunk frozen so I'd suggest if there
> > are no objections during 24 hours we unfroze the trunk. If  after that
> > we get a vote for fixing the issue then a branch is created to make
> > the fix and testing.
>
> Agreed, although it would be good to hear that the Eclipse 3.3 launching
> problem is fixed that Sian posted about on the thread "Re: [eclipse]
> Eclipse 3.3 can not download updates".


I've just done some more digging on this and it's only failing when you copy
the Harmony JRE to the eclipse/jre directory (which is new in Eclipse 3.3).
You can still add jre\bin to the path and run eclipse.exe like you did
before.  So as it's not a regression and it could take a while to figure out
I would probably say it's not a blocking issue for M3.

Next question is whether to raise a Harmony bug or an Eclipse bug...



> (Sorry that I can't be of much help with fixing things at the moment,
> I'm about to get on an airplane again.)
>
> Regards,
> Tim
>



-- 
Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number
741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU

Re: [general][M3] release r580985 as M3 stable build

Posted by Tim Ellison <t....@gmail.com>.
Stepan Mishura wrote:
> Let's make a decision about the last milestone candidate r580985.
> Shortly, as I can see there is only one issue that blocks its release
> - HARMONY-4873
> 
> I'd suggest the next - we wait for objections to postpone the issue
> resolution to M4 for 72 hours (starting from now) to let everyone to
> speak up.
> If there are no objections - we make the release. Also I think we
> shouldn't keep all this time the trunk frozen so I'd suggest if there
> are no objections during 24 hours we unfroze the trunk. If  after that
> we get a vote for fixing the issue then a branch is created to make
> the fix and testing.

Agreed, although it would be good to hear that the Eclipse 3.3 launching
problem is fixed that Sian posted about on the thread "Re: [eclipse]
Eclipse 3.3 can not download updates".

(Sorry that I can't be of much help with fixing things at the moment,
I'm about to get on an airplane again.)

Regards,
Tim

Re: [general][M3] release r580985 as M3 stable build

Posted by Stepan Mishura <st...@gmail.com>.
On 10/4/07, Stepan Mishura wrote:
> Hi folks,
>
> Let's make a decision about the last milestone candidate r580985.
> Shortly, as I can see there is only one issue that blocks its release
> - HARMONY-4873
>
> I'd suggest the next - we wait for objections to postpone the issue
> resolution to M4 for 72 hours (starting from now) to let everyone to
> speak up.
> If there are no objections - we make the release. Also I think we
> shouldn't keep all this time the trunk frozen so I'd suggest if there
> are no objections during 24 hours we unfroze the trunk. If  after that
> we get a vote for fixing the issue then a branch is created to make
> the fix and testing.

No votes for fixing the issue in M3 - so I'm going to promote r580985
snapshot as M3 stable build. I've already copied it to the milestone
dir and updated download page. Also I think the announcement should be
made and the news page should be updated.

Currently I'm gathering info about what was fixed/improved/added in
M3. Please let me know what worth to be mentioned.

Thanks,
Stepan.

>
> Objections?
>
> Thanks,
> Stepan Mishura
> Intel Enterprise Solutions Software Division
>