You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@openoffice.apache.org by Zhe Liu <al...@gmail.com> on 2011/12/16 03:43:47 UTC

Help! JunitTest_framework_unoapi.mk:28: *** Malformed target-specific variable definition. Stop.

Hi All,
I always break because of the error when build on Windows XP. I
mentioned before, nobody responsed on it.  I did a little search and
found someone also encountered the problem.  I still have no clue how
to resolve it.

JunitTest_framework_unoapi.mk:28: *** Malformed target-specific
variable definition.  Stop.

To continue my build, I have to remove the lines related to Junitest.
There are several module with the same error. It's annoying to
workaround them all.   Could anybody help me?

Thanks.

-- 
Best Regards
>From aliuzhe@gmail.com

Re: Help! JunitTest_framework_unoapi.mk:28: *** Malformed target-specific variable definition. Stop.

Posted by Zhe Liu <al...@gmail.com>.
Opened a bug on Bugzilla.
https://issues.apache.org/ooo/show_bug.cgi?id=118682



2011/12/16 Zhe Liu <al...@gmail.com>:
> Thanks for your patch. It works.
> Is there a bug in our Bugzilla to indicate the problem?
>
>
> 2011/12/16 Ariel Constenla-Haile <ar...@apache.org>:
>> Hi Zhe Liu,
>>
>> On Fri, Dec 16, 2011 at 10:43:47AM +0800, Zhe Liu wrote:
>>> Hi All,
>>> I always break because of the error when build on Windows XP. I
>>> mentioned before, nobody responsed on it.  I did a little search and
>>> found someone also encountered the problem.  I still have no clue how
>>> to resolve it.
>>>
>>> JunitTest_framework_unoapi.mk:28: *** Malformed target-specific
>>> variable definition.  Stop.
>>>
>>> To continue my build, I have to remove the lines related to Junitest.
>>> There are several module with the same error. It's annoying to
>>> workaround them all.   Could anybody help me?
>>
>> there are different ways to work around this:
>>
>> 1. comment the part in Module_XXX.mk where the tests are included
>> 2. surround this part with a condition, checking for OOO_JUNIT_JAR
>> 3. disable the code in solenv/gbuild/JunitTest.mk with the check for
>>   OOO_JUNIT_JAR
>> 4. disable gb_Module_add_subsequentcheck_targets with a similar check
>>
>>
>> I first tried (3), then (4). I'm not sure yet, I'll need to test this
>> further, but you can give it a try. The patch is a dirty mix of (3) and
>> (4).
>>
>> The problem is not solved disabling junit in configure, because the issue
>> is not with compiling the unit tests, but with parsing the Makefiles
>> JunitTest_xxx_xxx.mk , so we have to find a way not to parse those files
>> (until we find the real bug here - my guess, a gmake one).
>>
>> Testing the patch with
>> ]$ make -dsr --print-data-base --warn-undefined-variables
>> shows that at least the Makefiles are not even parsed.
>>
>> You should first unset OOO_JUNIT_JAR before compiling:
>>
>> ]$ unset OOO_JUNIT_JAR
>>
>>
>> Regards
>> --
>> Ariel Constenla-Haile
>> La Plata, Argentina
>
>
>
> --
> Best Regards
> From aliuzhe@gmail.com



-- 
Best Regards
>From aliuzhe@gmail.com

Re: Help! JunitTest_framework_unoapi.mk:28: *** Malformed target-specific variable definition. Stop.

Posted by Zhe Liu <al...@gmail.com>.
Thanks for your patch. It works.
Is there a bug in our Bugzilla to indicate the problem?


2011/12/16 Ariel Constenla-Haile <ar...@apache.org>:
> Hi Zhe Liu,
>
> On Fri, Dec 16, 2011 at 10:43:47AM +0800, Zhe Liu wrote:
>> Hi All,
>> I always break because of the error when build on Windows XP. I
>> mentioned before, nobody responsed on it.  I did a little search and
>> found someone also encountered the problem.  I still have no clue how
>> to resolve it.
>>
>> JunitTest_framework_unoapi.mk:28: *** Malformed target-specific
>> variable definition.  Stop.
>>
>> To continue my build, I have to remove the lines related to Junitest.
>> There are several module with the same error. It's annoying to
>> workaround them all.   Could anybody help me?
>
> there are different ways to work around this:
>
> 1. comment the part in Module_XXX.mk where the tests are included
> 2. surround this part with a condition, checking for OOO_JUNIT_JAR
> 3. disable the code in solenv/gbuild/JunitTest.mk with the check for
>   OOO_JUNIT_JAR
> 4. disable gb_Module_add_subsequentcheck_targets with a similar check
>
>
> I first tried (3), then (4). I'm not sure yet, I'll need to test this
> further, but you can give it a try. The patch is a dirty mix of (3) and
> (4).
>
> The problem is not solved disabling junit in configure, because the issue
> is not with compiling the unit tests, but with parsing the Makefiles
> JunitTest_xxx_xxx.mk , so we have to find a way not to parse those files
> (until we find the real bug here - my guess, a gmake one).
>
> Testing the patch with
> ]$ make -dsr --print-data-base --warn-undefined-variables
> shows that at least the Makefiles are not even parsed.
>
> You should first unset OOO_JUNIT_JAR before compiling:
>
> ]$ unset OOO_JUNIT_JAR
>
>
> Regards
> --
> Ariel Constenla-Haile
> La Plata, Argentina



-- 
Best Regards
>From aliuzhe@gmail.com

Re: Help! JunitTest_framework_unoapi.mk:28: *** Malformed target-specific variable definition. Stop.

Posted by Ariel Constenla-Haile <ar...@apache.org>.
Hi Zhe Liu,

On Fri, Dec 16, 2011 at 10:43:47AM +0800, Zhe Liu wrote:
> Hi All,
> I always break because of the error when build on Windows XP. I
> mentioned before, nobody responsed on it.  I did a little search and
> found someone also encountered the problem.  I still have no clue how
> to resolve it.
> 
> JunitTest_framework_unoapi.mk:28: *** Malformed target-specific
> variable definition.  Stop.
> 
> To continue my build, I have to remove the lines related to Junitest.
> There are several module with the same error. It's annoying to
> workaround them all.   Could anybody help me?

there are different ways to work around this:

1. comment the part in Module_XXX.mk where the tests are included
2. surround this part with a condition, checking for OOO_JUNIT_JAR
3. disable the code in solenv/gbuild/JunitTest.mk with the check for
   OOO_JUNIT_JAR
4. disable gb_Module_add_subsequentcheck_targets with a similar check


I first tried (3), then (4). I'm not sure yet, I'll need to test this
further, but you can give it a try. The patch is a dirty mix of (3) and
(4).

The problem is not solved disabling junit in configure, because the issue
is not with compiling the unit tests, but with parsing the Makefiles
JunitTest_xxx_xxx.mk , so we have to find a way not to parse those files
(until we find the real bug here - my guess, a gmake one).

Testing the patch with 
]$ make -dsr --print-data-base --warn-undefined-variables 
shows that at least the Makefiles are not even parsed.

You should first unset OOO_JUNIT_JAR before compiling: 

]$ unset OOO_JUNIT_JAR


Regards
-- 
Ariel Constenla-Haile
La Plata, Argentina

Re: Help! JunitTest_framework_unoapi.mk:28: *** Malformed target-specific variable definition. Stop.

Posted by Mathias Bauer <ma...@numberfour.eu>.
Hi,

sorry, but I don't have the Windows machine anymore where I did my OOo 
builds. Nowadays I'm only building on Linux. But you can follow my build 
instructions for make in the OOo wiki: just download the 3.81 source 
tarball, unpack it, configure, apply the patch and build.

Regards,
Mathias

On 20.12.2011 02:38, Zhe Liu wrote:
> Thank you for the comment.
> Could you share your make?  Juergen said that he can provide space to store it.
> I prefer to directly use it. -:)
>
> 2011/12/20 Mathias Bauer<Ma...@gmx.net>:
>> On 19.12.2011 23:23, Andrew Rist wrote:
>>>
>>>
>>>
>>> On 12/16/2011 9:24 AM, Mathias Bauer wrote:
>>>>
>>>> On 16.12.2011 03:43, Zhe Liu wrote:
>>>>>
>>>>> Hi All,
>>>>> I always break because of the error when build on Windows XP. I
>>>>> mentioned before, nobody responsed on it. I did a little search and
>>>>> found someone also encountered the problem. I still have no clue how
>>>>> to resolve it.
>>>>>
>>>>> JunitTest_framework_unoapi.mk:28: *** Malformed target-specific
>>>>> variable definition. Stop.
>>>>>
>>>>> To continue my build, I have to remove the lines related to Junitest.
>>>>> There are several module with the same error. It's annoying to
>>>>> workaround them all. Could anybody help me?
>>>>>
>>>>> Thanks.
>>>>>
>>>>
>>>> What version of GNU Make do you use? 3.82 has a bug that let GNU Make
>>>> spit out this error even if the variable definition is OK (and is
>>>> parsed without problems by 3.81).
>>>
>>> I have run into this issue also on Win7. Cygwin has a current version of
>>> 3.82.90 (thus hitting 3.82 'bug').
>>> What is the best way to deal with this? Is this something that is
>>> considered a bug by gmake, or is this a regression that will continue?
>>> (it involves 'make -r' not working correctly and I guess does look like
>>> a regression - not a new feature)
>>> If we can move to a new syntax, without breaking any other platforms,
>>> but also using current Cygwin packages - would that would be the best
>>> solution?
>>>
>>> For now I will switch the instructions for the buildbot, call out the
>>> use of make 3.81-2 (which is also available in current Cygwin).
>>> At this point, this seems to be a solution for the build.
>>
>>
>> I reported the problem to cygwin several months ago:
>>
>> http://cygwin.com/ml/cygwin/2011-02/msg00398.html
>>
>> See also
>>
>> http://wiki.services.openoffice.org/wiki/Build_Environment_Effort/Status_And_Next_Steps
>>
>> At the end of the page you can see my instructions to build a custom version
>> of make. This gave me the best make on cygwin so far.
>>
>> This page mentions a performance bug in make 3.82 that we also found on
>> cygwin. It's possible that this was a problem of the HEAD version at that
>> time.
>>
>> Regards,
>> Mathias
>
>
>


Re: Help! JunitTest_framework_unoapi.mk:28: *** Malformed target-specific variable definition. Stop.

Posted by Zhe Liu <al...@gmail.com>.
Thank you for the comment.
Could you share your make?  Juergen said that he can provide space to store it.
I prefer to directly use it. -:)

2011/12/20 Mathias Bauer <Ma...@gmx.net>:
> On 19.12.2011 23:23, Andrew Rist wrote:
>>
>>
>>
>> On 12/16/2011 9:24 AM, Mathias Bauer wrote:
>>>
>>> On 16.12.2011 03:43, Zhe Liu wrote:
>>>>
>>>> Hi All,
>>>> I always break because of the error when build on Windows XP. I
>>>> mentioned before, nobody responsed on it. I did a little search and
>>>> found someone also encountered the problem. I still have no clue how
>>>> to resolve it.
>>>>
>>>> JunitTest_framework_unoapi.mk:28: *** Malformed target-specific
>>>> variable definition. Stop.
>>>>
>>>> To continue my build, I have to remove the lines related to Junitest.
>>>> There are several module with the same error. It's annoying to
>>>> workaround them all. Could anybody help me?
>>>>
>>>> Thanks.
>>>>
>>>
>>> What version of GNU Make do you use? 3.82 has a bug that let GNU Make
>>> spit out this error even if the variable definition is OK (and is
>>> parsed without problems by 3.81).
>>
>> I have run into this issue also on Win7. Cygwin has a current version of
>> 3.82.90 (thus hitting 3.82 'bug').
>> What is the best way to deal with this? Is this something that is
>> considered a bug by gmake, or is this a regression that will continue?
>> (it involves 'make -r' not working correctly and I guess does look like
>> a regression - not a new feature)
>> If we can move to a new syntax, without breaking any other platforms,
>> but also using current Cygwin packages - would that would be the best
>> solution?
>>
>> For now I will switch the instructions for the buildbot, call out the
>> use of make 3.81-2 (which is also available in current Cygwin).
>> At this point, this seems to be a solution for the build.
>
>
> I reported the problem to cygwin several months ago:
>
> http://cygwin.com/ml/cygwin/2011-02/msg00398.html
>
> See also
>
> http://wiki.services.openoffice.org/wiki/Build_Environment_Effort/Status_And_Next_Steps
>
> At the end of the page you can see my instructions to build a custom version
> of make. This gave me the best make on cygwin so far.
>
> This page mentions a performance bug in make 3.82 that we also found on
> cygwin. It's possible that this was a problem of the HEAD version at that
> time.
>
> Regards,
> Mathias



-- 
Best Regards
>From aliuzhe@gmail.com

Re: Help! JunitTest_framework_unoapi.mk:28: *** Malformed target-specific variable definition. Stop.

Posted by Mathias Bauer <Ma...@gmx.net>.
On 19.12.2011 23:23, Andrew Rist wrote:
>
>
> On 12/16/2011 9:24 AM, Mathias Bauer wrote:
>> On 16.12.2011 03:43, Zhe Liu wrote:
>>> Hi All,
>>> I always break because of the error when build on Windows XP. I
>>> mentioned before, nobody responsed on it. I did a little search and
>>> found someone also encountered the problem. I still have no clue how
>>> to resolve it.
>>>
>>> JunitTest_framework_unoapi.mk:28: *** Malformed target-specific
>>> variable definition. Stop.
>>>
>>> To continue my build, I have to remove the lines related to Junitest.
>>> There are several module with the same error. It's annoying to
>>> workaround them all. Could anybody help me?
>>>
>>> Thanks.
>>>
>>
>> What version of GNU Make do you use? 3.82 has a bug that let GNU Make
>> spit out this error even if the variable definition is OK (and is
>> parsed without problems by 3.81).
> I have run into this issue also on Win7. Cygwin has a current version of
> 3.82.90 (thus hitting 3.82 'bug').
> What is the best way to deal with this? Is this something that is
> considered a bug by gmake, or is this a regression that will continue?
> (it involves 'make -r' not working correctly and I guess does look like
> a regression - not a new feature)
> If we can move to a new syntax, without breaking any other platforms,
> but also using current Cygwin packages - would that would be the best
> solution?
>
> For now I will switch the instructions for the buildbot, call out the
> use of make 3.81-2 (which is also available in current Cygwin).
> At this point, this seems to be a solution for the build.

I reported the problem to cygwin several months ago:

http://cygwin.com/ml/cygwin/2011-02/msg00398.html

See also

http://wiki.services.openoffice.org/wiki/Build_Environment_Effort/Status_And_Next_Steps 


At the end of the page you can see my instructions to build a custom 
version of make. This gave me the best make on cygwin so far.

This page mentions a performance bug in make 3.82 that we also found on 
cygwin. It's possible that this was a problem of the HEAD version at 
that time.

Regards,
Mathias

Re: Help! JunitTest_framework_unoapi.mk:28: *** Malformed target-specific variable definition. Stop.

Posted by Andrew Rist <an...@oracle.com>.

On 12/16/2011 9:24 AM, Mathias Bauer wrote:
> On 16.12.2011 03:43, Zhe Liu wrote:
>> Hi All,
>> I always break because of the error when build on Windows XP. I
>> mentioned before, nobody responsed on it.  I did a little search and
>> found someone also encountered the problem.  I still have no clue how
>> to resolve it.
>>
>> JunitTest_framework_unoapi.mk:28: *** Malformed target-specific
>> variable definition.  Stop.
>>
>> To continue my build, I have to remove the lines related to Junitest.
>> There are several module with the same error. It's annoying to
>> workaround them all.   Could anybody help me?
>>
>> Thanks.
>>
>
> What version of GNU Make do you use? 3.82 has a bug that let GNU Make 
> spit out this error even if the variable definition is OK (and is 
> parsed without problems by 3.81).
I have run into this issue also on Win7.  Cygwin has a current version 
of 3.82.90  (thus hitting 3.82 'bug').
What is the best way to deal with this?  Is this something that is 
considered a bug by gmake, or is this a regression that will continue?
(it involves 'make -r' not working correctly and I guess does look like 
a regression - not a new feature)
If we can move to a new syntax, without breaking any other platforms, 
but also using current Cygwin packages - would that would be the best 
solution?

For now I will switch the instructions for the buildbot, call out the 
use of make 3.81-2 (which is also available in current Cygwin).
At this point, this seems to be a solution for the build.

Andrew



>
> Regards,
> Mathias


Re: Help! JunitTest_framework_unoapi.mk:28: *** Malformed target-specific variable definition. Stop.

Posted by Mathias Bauer <Ma...@gmx.net>.
On 19.12.2011 18:02, Jürgen Schmidt wrote:

> On 12/19/11 12:27 AM, Mathias Bauer wrote:
>> Am 19.12.2011 00:19, schrieb Michael Stahl:
>>
>>> On 16/12/11 18:24, Mathias Bauer wrote:
>>>> On 16.12.2011 03:43, Zhe Liu wrote:
>>>>> Hi All,
>>>>> I always break because of the error when build on Windows XP. I
>>>>> mentioned before, nobody responsed on it. I did a little search and
>>>>> found someone also encountered the problem. I still have no clue how
>>>>> to resolve it.
>>>>>
>>>>> JunitTest_framework_unoapi.mk:28: *** Malformed target-specific
>>>>> variable definition. Stop.
>>>>>
>>>>
>>>> What version of GNU Make do you use? 3.82 has a bug that let GNU Make
>>>> spit out this error even if the variable definition is OK (and is
>>>> parsed
>>>> without problems by 3.81).
>>>
>>> AFAIR this is a problem that only happens on Cygwin (i speculate the
>>> problem is triggered by the CLASSPATH that contains jar files with mixed
>>> path notation, containing a colon).
>>
>> Yes, and I wrote it because Zhe Liu had this problem on Windows. ;-)
>>
>>> getting a make that works well on windows is surprisingly difficult :-/
>>
>> Indeed. The best I could get is the one I built myself in the way I have
>> documented in the OOo Building Guide in the "old" Wiki.
> maybe you can provide your build somewhere ;-) how about
> http://people.apache.org/~mba/d´gnumake/...
>
> Or somewhere else where people can easy download it. Or send a version
> to me and i will provide it somewhere.

Please see my reply to Andrew.

Regards,
Mathias

Re: Help! JunitTest_framework_unoapi.mk:28: *** Malformed target-specific variable definition. Stop.

Posted by Jürgen Schmidt <jo...@googlemail.com>.
On 12/19/11 12:27 AM, Mathias Bauer wrote:
> Am 19.12.2011 00:19, schrieb Michael Stahl:
>
>> On 16/12/11 18:24, Mathias Bauer wrote:
>>> On 16.12.2011 03:43, Zhe Liu wrote:
>>>> Hi All,
>>>> I always break because of the error when build on Windows XP. I
>>>> mentioned before, nobody responsed on it.  I did a little search and
>>>> found someone also encountered the problem.  I still have no clue how
>>>> to resolve it.
>>>>
>>>> JunitTest_framework_unoapi.mk:28: *** Malformed target-specific
>>>> variable definition.  Stop.
>>>>
>>>
>>> What version of GNU Make do you use? 3.82 has a bug that let GNU Make
>>> spit out this error even if the variable definition is OK (and is parsed
>>> without problems by 3.81).
>>
>> AFAIR this is a problem that only happens on Cygwin (i speculate the
>> problem is triggered by the CLASSPATH that contains jar files with mixed
>> path notation, containing a colon).
>
> Yes, and I wrote it because Zhe Liu had this problem on Windows. ;-)
>
>> getting a make that works well on windows is surprisingly difficult :-/
>
> Indeed. The best I could get is the one I built myself in the way I have
> documented in the OOo Building Guide in the "old" Wiki.
maybe you can provide your build somewhere ;-)  how about 
http://people.apache.org/~mba/d´gnumake/...

Or somewhere else where people can easy download it. Or send a version 
to me and i will provide it somewhere.

Juergen


>
> Regards,
> Mathias


Re: Help! JunitTest_framework_unoapi.mk:28: *** Malformed target-specific variable definition. Stop.

Posted by Mathias Bauer <Ma...@gmx.net>.
Am 19.12.2011 00:19, schrieb Michael Stahl:

> On 16/12/11 18:24, Mathias Bauer wrote:
>> On 16.12.2011 03:43, Zhe Liu wrote:
>>> Hi All,
>>> I always break because of the error when build on Windows XP. I
>>> mentioned before, nobody responsed on it.  I did a little search and
>>> found someone also encountered the problem.  I still have no clue how
>>> to resolve it.
>>>
>>> JunitTest_framework_unoapi.mk:28: *** Malformed target-specific
>>> variable definition.  Stop.
>>>
>> 
>> What version of GNU Make do you use? 3.82 has a bug that let GNU Make 
>> spit out this error even if the variable definition is OK (and is parsed 
>> without problems by 3.81).
> 
> AFAIR this is a problem that only happens on Cygwin (i speculate the
> problem is triggered by the CLASSPATH that contains jar files with mixed
> path notation, containing a colon).

Yes, and I wrote it because Zhe Liu had this problem on Windows. ;-)

> getting a make that works well on windows is surprisingly difficult :-/

Indeed. The best I could get is the one I built myself in the way I have
documented in the OOo Building Guide in the "old" Wiki.

Regards,
Mathias

Re: Help! JunitTest_framework_unoapi.mk:28: *** Malformed target-specific variable definition. Stop.

Posted by Michael Stahl <ms...@openoffice.org>.
On 16/12/11 18:24, Mathias Bauer wrote:
> On 16.12.2011 03:43, Zhe Liu wrote:
>> Hi All,
>> I always break because of the error when build on Windows XP. I
>> mentioned before, nobody responsed on it.  I did a little search and
>> found someone also encountered the problem.  I still have no clue how
>> to resolve it.
>>
>> JunitTest_framework_unoapi.mk:28: *** Malformed target-specific
>> variable definition.  Stop.
>>
> 
> What version of GNU Make do you use? 3.82 has a bug that let GNU Make 
> spit out this error even if the variable definition is OK (and is parsed 
> without problems by 3.81).

AFAIR this is a problem that only happens on Cygwin (i speculate the
problem is triggered by the CLASSPATH that contains jar files with mixed
path notation, containing a colon).

getting a make that works well on windows is surprisingly difficult :-/


Re: Help! JunitTest_framework_unoapi.mk:28: *** Malformed target-specific variable definition. Stop.

Posted by Mathias Bauer <Ma...@gmx.net>.
On 16.12.2011 03:43, Zhe Liu wrote:
> Hi All,
> I always break because of the error when build on Windows XP. I
> mentioned before, nobody responsed on it.  I did a little search and
> found someone also encountered the problem.  I still have no clue how
> to resolve it.
>
> JunitTest_framework_unoapi.mk:28: *** Malformed target-specific
> variable definition.  Stop.
>
> To continue my build, I have to remove the lines related to Junitest.
> There are several module with the same error. It's annoying to
> workaround them all.   Could anybody help me?
>
> Thanks.
>

What version of GNU Make do you use? 3.82 has a bug that let GNU Make 
spit out this error even if the variable definition is OK (and is parsed 
without problems by 3.81).

Regards,
Mathias