Bug#805228: Bug#806630: libnative-platform-java: FTBFS when built with dpkg-buildpackage -A (mh_installjar fails)

[Emmanuel Bourg]

Le 20/07/2016 à 21:47, Lucas Nussbaum a écrit :

> So, it would make sense to fix that bug on the maven-repo-helper side, rather
> than require changes to each of those packages individually.

Thank you for investigating this Lucas. Please do not start a mass bug
filing for these packages though, they will just end up merged in the
maven-repo-helper bug since we aren't going to close them one by one anyway.

I'm not really happy to see this class of issue marked as serious for
stretch, it's not reasonable to raise the severity and risk an auto
removal if "dpkg-buildpackage -A" fails but not "dpkg-buildpackage".
This is just a minor convenience to enable source-only uploads (which
aren't even mandatory!).

Emmanuel Bourg

__
This is the maintainer address of Debian's Java team
. 
Please use
debian-j...@lists.debian.org for discussions and questions.

Bug#805228: Bug#806630: libnative-platform-java: FTBFS when built with dpkg-buildpackage -A (mh_installjar fails)

[tony mancill]

On 07/16/2016 03:35 PM, Santiago Vila wrote:
> On Sat, 16 Jul 2016, tony mancill wrote:
> 
>> The first two packages I tried, src:pegdown and src:libjbcrypt-java,
>> both appear in the "successfully built reproducibly" set here [1], and
>> neither has any bugs assigned to it.  Maybe I'm looking in the wrong place?
> 
> If those packages generate only "Arch: all" packages and they build ok
> in the reproducible builds setup, then those packages are ok.
> 
> The problem with the helper program you are using seems to be that it
> makes "dpkg-buildpackage -A" to fail on packages generating both
> "Arch: all" and "Arch: any" binary packages.
> 
> The same packages might appear as "ok" in reproducible builds because
> they just test for "dpkg-buildpackage".
> 
> So, it is possible that there are no other affected packages than the
> ones in my list.

Thank you the clarification.  If that's the case, we can probably
address the bug with src:libnative-platform-java by updating
debian/rules to avoid using the DH sequencer and making the required
calls to mh_install explicitly.

That doesn't detract from the severity of this bug, but does allow more
time for discussion.

Cheers,
tony



signature.asc
Description: OpenPGP digital signature
__
This is the maintainer address of Debian's Java team
. 
Please use
debian-j...@lists.debian.org for discussions and questions.

Bug#805228: Bug#806630: libnative-platform-java: FTBFS when built with dpkg-buildpackage -A (mh_installjar fails)

[Santiago Vila]

On Sat, 16 Jul 2016, tony mancill wrote:

> The first two packages I tried, src:pegdown and src:libjbcrypt-java,
> both appear in the "successfully built reproducibly" set here [1], and
> neither has any bugs assigned to it.  Maybe I'm looking in the wrong place?

If those packages generate only "Arch: all" packages and they build ok
in the reproducible builds setup, then those packages are ok.

The problem with the helper program you are using seems to be that it
makes "dpkg-buildpackage -A" to fail on packages generating both
"Arch: all" and "Arch: any" binary packages.

The same packages might appear as "ok" in reproducible builds because
they just test for "dpkg-buildpackage".

So, it is possible that there are no other affected packages than the
ones in my list.

Thanks.

__
This is the maintainer address of Debian's Java team
. 
Please use
debian-j...@lists.debian.org for discussions and questions.

Bug#805228: Bug#806630: libnative-platform-java: FTBFS when built with dpkg-buildpackage -A (mh_installjar fails)

[tony mancill]

On 07/16/2016 01:58 PM, Santiago Vila wrote:
> On Sat, 16 Jul 2016, tony mancill wrote:
> 
>>> https://bugs.debian.org/cgi-bin/pkgreport.cgi?tag=binary-indep;users=sanv...@debian.org
>>
>> Hi Santiago,
>>
>> Thanks for the list.  For 806640, the only bug that appears in this list
>> related to mh_install 805228, I will mark it as blocked by 805228, but I
>> won't reassign.
>>
>> It might be worth noting that there are other arch:all packages that
>> FTBFS with the same mh_install issue, but don't appear in your list of
>> binary-indep packages.  Did your test build perhaps exclude packages
>> that are strictly arch:all - as compared to src:libnative-platform-java,
>> which builds both arch:any and arch:all binary packages? [...]
> 
> Yes, I have only tested packages which generate both "Arch: any" and
> "Arch: all" packages, so source packages which only generate "Arch: all"
> packages are excluded from my tests.
> 
> The reason is that those packages are already tested (and reported)
> in the reproducible builds project, so I decided to concentrate
> on packages that nobody tested.

The first two packages I tried, src:pegdown and src:libjbcrypt-java,
both appear in the "successfully built reproducibly" set here [1], and
neither has any bugs assigned to it.  Maybe I'm looking in the wrong place?

In any event, that entire class of bug (I am sure that there are other
instances) will be addressed by the resolution of 805228.

Thank you,
tony

[1]
https://tests.reproducible-builds.org/debian/unstable/amd64/pkg_set_maint_pkg-java-maintainers.html





signature.asc
Description: OpenPGP digital signature
__
This is the maintainer address of Debian's Java team
. 
Please use
debian-j...@lists.debian.org for discussions and questions.

Bug#805228: Bug#806630: libnative-platform-java: FTBFS when built with dpkg-buildpackage -A (mh_installjar fails)

[Santiago Vila]

On Sat, 16 Jul 2016, tony mancill wrote:

> > https://bugs.debian.org/cgi-bin/pkgreport.cgi?tag=binary-indep;users=sanv...@debian.org
> 
> Hi Santiago,
> 
> Thanks for the list.  For 806640, the only bug that appears in this list
> related to mh_install 805228, I will mark it as blocked by 805228, but I
> won't reassign.
> 
> It might be worth noting that there are other arch:all packages that
> FTBFS with the same mh_install issue, but don't appear in your list of
> binary-indep packages.  Did your test build perhaps exclude packages
> that are strictly arch:all - as compared to src:libnative-platform-java,
> which builds both arch:any and arch:all binary packages? [...]

Yes, I have only tested packages which generate both "Arch: any" and
"Arch: all" packages, so source packages which only generate "Arch: all"
packages are excluded from my tests.

The reason is that those packages are already tested (and reported)
in the reproducible builds project, so I decided to concentrate
on packages that nobody tested.

Thanks.

__
This is the maintainer address of Debian's Java team
. 
Please use
debian-j...@lists.debian.org for discussions and questions.

Bug#805228: Bug#806630: libnative-platform-java: FTBFS when built with dpkg-buildpackage -A (mh_installjar fails)

[tony mancill]

On 07/16/2016 01:47 PM, tony mancill wrote:

> Thanks for the list.  For 806640, the only bug that appears in this list
> related to mh_install 805228, I will mark it as blocked by 805228, but I
> won't reassign.

Grrr sorry for the typos.  I meant 806630, and to say that I will
mark it blocked by 805228.  That's all.




signature.asc
Description: OpenPGP digital signature
__
This is the maintainer address of Debian's Java team
. 
Please use
debian-j...@lists.debian.org for discussions and questions.

Bug#805228: Bug#806630: libnative-platform-java: FTBFS when built with dpkg-buildpackage -A (mh_installjar fails)

[tony mancill]

On 07/16/2016 01:31 PM, Santiago Vila wrote:
> On Sat, 16 Jul 2016, Markus Koschany wrote:
> 
>> On 16.07.2016 18:22, tony mancill wrote:
>> [...]
>>> I'd like to discuss options with the Java Team.  For example, we could
>>> change mh_install to ignore '-i' when not passed in conjunction with a
>>> maven.ignoreRules file, or change the DH sequencer, and probably some
>>> other better options that I'm not creative enough to think of right now.
>>
>> Hi,
>>
>> I think
>>
>> https://bugs.debian.org/805228
>>
>> is the best place to discuss this issue.
> 
> Indeed.
> 
> 
> One thing that you might want to do is to identify the affected
> packages. The complete list of bugs I reported regarding
> "dpkg-buildpackage -A" is available here in case it helps:
> 
> https://bugs.debian.org/cgi-bin/pkgreport.cgi?tag=binary-indep;users=sanv...@debian.org

Hi Santiago,

Thanks for the list.  For 806640, the only bug that appears in this list
related to mh_install 805228, I will mark it as blocked by 805228, but I
won't reassign.

It might be worth noting that there are other arch:all packages that
FTBFS with the same mh_install issue, but don't appear in your list of
binary-indep packages.  Did your test build perhaps exclude packages
that are strictly arch:all - as compared to src:libnative-platform-java,
which builds both arch:any and arch:all binary packages?  Put another
way, do we need to be concerned that the strictly arch:all packages for
this binary-indep tag?

Cheers,
tony



signature.asc
Description: OpenPGP digital signature
__
This is the maintainer address of Debian's Java team
. 
Please use
debian-j...@lists.debian.org for discussions and questions.

Bug#805228: Bug#806630: libnative-platform-java: FTBFS when built with dpkg-buildpackage -A (mh_installjar fails)

[Santiago Vila]

On Sat, 16 Jul 2016, Markus Koschany wrote:

> On 16.07.2016 18:22, tony mancill wrote:
> [...]
> > I'd like to discuss options with the Java Team.  For example, we could
> > change mh_install to ignore '-i' when not passed in conjunction with a
> > maven.ignoreRules file, or change the DH sequencer, and probably some
> > other better options that I'm not creative enough to think of right now.
> 
> Hi,
> 
> I think
> 
> https://bugs.debian.org/805228
> 
> is the best place to discuss this issue.

Indeed.


One thing that you might want to do is to identify the affected
packages. The complete list of bugs I reported regarding
"dpkg-buildpackage -A" is available here in case it helps:

https://bugs.debian.org/cgi-bin/pkgreport.cgi?tag=binary-indep;users=sanv...@debian.org


Thanks.

__
This is the maintainer address of Debian's Java team
. 
Please use
debian-j...@lists.debian.org for discussions and questions.

Bug#806630: libnative-platform-java: FTBFS when built with dpkg-buildpackage -A (mh_installjar fails)

[Santiago Vila]

On Sat, 16 Jul 2016, tony mancill wrote:

> There are 320-odd packages that build-depend on maven-repo-helper,
> although only a subset of them will use the DH sequencer.  Still, I
> think the bug belongs there as a problem in the tool chain and not
> against all of the packages that depend on it.

Hi.

As Markus Koschany has pointed out, there is already a bug reported
specifically for that.

So, even if this is considered as a bug in maven-repo-helper, please
don't reassign the bugs to that package, because then it would be a
lot more difficult for me to track the affected packages.

Thanks.

__
This is the maintainer address of Debian's Java team
. 
Please use
debian-j...@lists.debian.org for discussions and questions.

Bug#806630: libnative-platform-java: FTBFS when built with dpkg-buildpackage -A (mh_installjar fails)

[Markus Koschany]

On 16.07.2016 18:22, tony mancill wrote:
[...]
> I'd like to discuss options with the Java Team.  For example, we could
> change mh_install to ignore '-i' when not passed in conjunction with a
> maven.ignoreRules file, or change the DH sequencer, and probably some
> other better options that I'm not creative enough to think of right now.

Hi,

I think

https://bugs.debian.org/805228

is the best place to discuss this issue.

Regards,

Markus



signature.asc
Description: OpenPGP digital signature
__
This is the maintainer address of Debian's Java team
. 
Please use
debian-j...@lists.debian.org for discussions and questions.

Bug#806630: libnative-platform-java: FTBFS when built with dpkg-buildpackage -A (mh_installjar fails)

[tony mancill]

On 07/14/2016 03:10 PM, Santiago Vila wrote:
> Greetings.
> 
> I have the ok from the Release Managers to consider this issue as RC
> for stretch. I'm going to wait at least one week before raising
> this to "serious".
> 
> If you need help to fix this bug, please tag it as "help".

Hi Santiago,

I took a look at it and this is a more general issue with the
maven-repo-helper DH sequencer and the fact that mh_install makes use of
'-i' with completely different semantics than those of debhelper.

Also, I'm kind of surprised that it's not affecting more packages.  It
seems like it could affect any package that uses DH construction like this:

dh $@ --with maven_repo_helper

(In fact, I spot-checked a few other packages that use maven_repo_helper
in this way and confirmed that they fail in the same way.)

There are 320-odd packages that build-depend on maven-repo-helper,
although only a subset of them will use the DH sequencer.  Still, I
think the bug belongs there as a problem in the tool chain and not
against all of the packages that depend on it.

I'd like to discuss options with the Java Team.  For example, we could
change mh_install to ignore '-i' when not passed in conjunction with a
maven.ignoreRules file, or change the DH sequencer, and probably some
other better options that I'm not creative enough to think of right now.

Thank you,
tony




signature.asc
Description: OpenPGP digital signature
__
This is the maintainer address of Debian's Java team
. 
Please use
debian-j...@lists.debian.org for discussions and questions.

Bug#806630: libnative-platform-java: FTBFS when built with dpkg-buildpackage -A (mh_installjar fails)

[Santiago Vila]

Greetings.

I have the ok from the Release Managers to consider this issue as RC
for stretch. I'm going to wait at least one week before raising
this to "serious".

If you need help to fix this bug, please tag it as "help".

Thanks.

__
This is the maintainer address of Debian's Java team
. 
Please use
debian-j...@lists.debian.org for discussions and questions.

Bug#806630: libnative-platform-java: FTBFS when built with dpkg-buildpackage -A (mh_installjar fails)

[Santiago Vila]

Package: src:libnative-platform-java
Version: 0.10+dfsg-1
User: sanv...@debian.org
Usertags: binary-indep
Severity: important

Dear maintainer:

I tried to build this package with "dpkg-buildpackage -A"
(i.e. only architecture-independent packages), and it failed:


[...]
 debian/rules build-indep
dh build-indep --with maven_repo_helper
   dh_testdir -i
   dh_auto_configure -i
   debian/rules override_dh_auto_build
make[1]: Entering directory '/<>/libnative-platform-java-0.10+dfsg'
make -f debian/Makefile
make[2]: Entering directory '/<>/libnative-platform-java-0.10+dfsg'
mkdir -p build/classes
javac -source 1.5 -target 1.5 -d build/classes -classpath 
/usr/share/java/joptsimple.jar 
src/main/java/net/rubygrapefruit/platform/Terminals.java 
src/main/java/net/rubygrapefruit/platform/FileSystems.java 
src/main/java/net/rubygrapefruit/platform/ProcessLauncher.java 
src/main/java/net/rubygrapefruit/platform/NativeIntegration.java 
src/main/java/net/rubygrapefruit/platform/Terminal.java 
src/main/java/net/rubygrapefruit/platform/SystemInfo.java 
src/main/java/net/rubygrapefruit/platform/MissingRegistryEntryException.java 
src/main/java/net/rubygrapefruit/platform/ThreadSafe.java 
src/main/java/net/rubygrapefruit/platform/Native.java 
src/main/java/net/rubygrapefruit/platform/TerminalSize.java 
src/main/java/net/rubygrapefruit/platform/PosixFile.java 
src/main/java/net/rubygrapefruit/platform/NativeException.java 
src/main/java/net/rubygrapefruit/platform/FileSystem.java 
src/main/java/net/rubygrapefruit/platform/package-info.java 
src/main/java/net/rubygrapefruit/platform/NativeIntegrationUnava
 ilableException.java 
src/main/java/net/rubygrapefruit/platform/WindowsRegistry.java 
src/main/java/net/rubygrapefruit/platform/Process.java 
src/main/java/net/rubygrapefruit/platform/PosixFiles.java 
src/main/java/net/rubygrapefruit/platform/internal/WrapperProcessLauncher.java 
src/main/java/net/rubygrapefruit/platform/internal/NativeLibraryLocator.java 
src/main/java/net/rubygrapefruit/platform/internal/WindowsTerminals.java 
src/main/java/net/rubygrapefruit/platform/internal/DefaultFileSystem.java 
src/main/java/net/rubygrapefruit/platform/internal/Platform.java 
src/main/java/net/rubygrapefruit/platform/internal/WrapperProcess.java 
src/main/java/net/rubygrapefruit/platform/internal/PosixFileSystems.java 
src/main/java/net/rubygrapefruit/platform/internal/WrapperTerminal.java 
src/main/java/net/rubygrapefruit/platform/internal/DefaultPosixFiles.java 
src/main/java/net/rubygrapefruit/platform/internal/DefaultProcessLauncher.java 
src/main/java/net/rubygrapefruit/platform/internal/AbstractTerm
 inals.java 
src/main/java/net/rubygrapefruit/platform/internal/AbstractTerminal.java 
src/main/java/net/rubygrapefruit/platform/internal/MutableTerminalSize.java 
src/main/java/net/rubygrapefruit/platform/internal/FileSystemList.java 
src/main/java/net/rubygrapefruit/platform/internal/DefaultWindowsRegistry.java 
src/main/java/net/rubygrapefruit/platform/internal/DefaultSystemInfo.java 
src/main/java/net/rubygrapefruit/platform/internal/TerminalCapabilities.java 
src/main/java/net/rubygrapefruit/platform/internal/MutableSystemInfo.java 
src/main/java/net/rubygrapefruit/platform/internal/TerminfoTerminals.java 
src/main/java/net/rubygrapefruit/platform/internal/NativeLibraryLoader.java 
src/main/java/net/rubygrapefruit/platform/internal/WindowsProcessLauncher.java 
src/main/java/net/rubygrapefruit/platform/internal/DefaultProcess.java 
src/main/java/net/rubygrapefruit/platform/internal/LibraryDef.java 
src/main/java/net/rubygrapefruit/platform/internal/TerminfoTerminal.java 
src/main/java/net/ruby
 grapefruit/platform/internal/FileStat.java 
src/main/java/net/rubygrapefruit/platform/internal/WindowsTerminal.java 
src/main/java/net/rubygrapefruit/platform/internal/jni/WindowsConsoleFunctions.java
 
src/main/java/net/rubygrapefruit/platform/internal/jni/NativeLibraryFunctions.java
 src/main/java/net/rubygrapefruit/platform/internal/jni/TerminfoFunctions.java 
src/main/java/net/rubygrapefruit/platform/internal/jni/WindowsHandleFunctions.java
 
src/main/java/net/rubygrapefruit/platform/internal/jni/PosixFileSystemFunctions.java
 src/main/java/net/rubygrapefruit/platform/internal/jni/PosixFileFunctions.java 
src/main/java/net/rubygrapefruit/platform/internal/jni/PosixProcessFunctions.java
 
src/main/java/net/rubygrapefruit/platform/internal/jni/WindowsRegistryFunctions.java
 
src/main/java/net/rubygrapefruit/platform/internal/jni/PosixTerminalFunctions.java
 src/main/java/net/rubygrapefruit/platform/internal/AnsiTerminal.java 
src/main/java/net/rubygrapefruit/platform/internal/FunctionResult.java
warning: [options] bootstrap class path not set in conjunction with -source 1.5
Note: src/main/java/net/rubygrapefruit/platform/internal/WrapperProcess.java 
uses unchecked or unsafe operations.

[... snipped ...]

-V --version: show the version