CI integration is broken

[rutsky]

Travis and Appveyor builds fail with different errors, e.g. in this PR #53 .

[aknuds1]

I noticed that too. I tried restarting the broken builds on Travis, but no luck. It looks like a transient error though (trouble connecting to APT repositories)?

On Mon, Nov 16, 2015 at 4:06 PM Vladimir Rutsky notifications@github.com wrote:

Travis and Appveyor builds fail with different errors, e.g. in this PR #53 https://github.com/harvimt/quamash/pull/53 .

— Reply to this email directly or view it on GitHub https://github.com/harvimt/quamash/issues/54.

[aknuds1]

As for AppVeyor I'm not familiar with that environment, but it breaks due to missing PyQt AFAICT.

[rutsky]

Travis is broken at least because Qt PPA was disappeared: looks like there is ppa:beineri/opt-qt542 instead of ppa:beineri/opt-qt541. I tried to change PPA here https://github.com/rutsky/quamash/pull/1/files, but this leads to other issues, you can see here: https://travis-ci.org/rutsky/quamash/jobs/91392601 --- I think the issue is that PySide doesn't support Qt5 (and AFAIK never did).

[aknuds1]

Yeah I think it's correct that PySide is 4 only. Thanks for the fix.

On Mon, Nov 16, 2015, 18:00 Vladimir Rutsky notifications@github.com wrote:

Travis is broken at least because Qt PPA was disappeared: looks like there is ppa:beineri/opt-qt542 instead of ppa:beineri/opt-qt541. I tried to change PPA here https://github.com/rutsky/quamash/pull/1/files, but this leads to other issues, you can see here: https://travis-ci.org/rutsky/quamash/jobs/91392601 --- I think the issue is that PySide doesn't support Qt5 (and AFAIK never was).

— Reply to this email directly or view it on GitHub https://github.com/harvimt/quamash/issues/54#issuecomment-157097659.

[harvimt]

ppa:beineri/opt-qt542 instead of ppa:beineri/opt-qt541

that's correct, I noticed that while I was working on the Dockerfiles

On Mon, Nov 16, 2015 at 9:29 AM, Arve Knudsen notifications@github.com wrote:

Yeah I think it's correct that PySide is 4 only. Thanks for the fix.

On Mon, Nov 16, 2015, 18:00 Vladimir Rutsky notifications@github.com wrote:

Travis is broken at least because Qt PPA was disappeared: looks like there is ppa:beineri/opt-qt542 instead of ppa:beineri/opt-qt541. I tried to change PPA here https://github.com/rutsky/quamash/pull/1/files, but this leads to other issues, you can see here: https://travis-ci.org/rutsky/quamash/jobs/91392601 --- I think the issue is that PySide doesn't support Qt5 (and AFAIK never was).

— Reply to this email directly or view it on GitHub https://github.com/harvimt/quamash/issues/54#issuecomment-157097659.

— Reply to this email directly or view it on GitHub https://github.com/harvimt/quamash/issues/54#issuecomment-157109736.

[harvimt]

A fork of PySide (annoyingly) called PySide2 supports Qt5, but you need to build it from source and it isn't considered stable yet.

[harvimt]

Alright Travis builds are fixed.

[harvimt]

Appveyor builds are fixed (though the code is kind of ugly) The problem was you need to set the Registry key so that the PyQt installers know where to install PyQt.

PySide tests show as passing but return a random number as a status code. (same as before) PyQt4 tests pass, but give warnings. PyQt5 tests don't pass, but not because PyQt5 is failing to install.

In other words #38 is not fixed yet.

[rutsky]

PySide tests show as passing but return a random number as a status code. (same as before)

Looking at last build it's not a random number, but 0xc0000005 error (-1073741819 is signed 32-bit representation for 0xc0000005) — Access Violation.

PyQt4 tests pass, but give warnings.

Are you sure? I see here that they still fails: https://ci.appveyor.com/project/harvimt/quamash/branch/master/job/mlr059kahlpose55

[harvimt]

you need to look at the fix-appveyor branch: https://ci.appveyor.com/project/harvimt/quamash/build/1.0.157

On Wed, Nov 18, 2015 at 5:17 PM, Vladimir Rutsky notifications@github.com wrote:

PySide tests show as passing but return a random number as a status code. (same as before)

Looking at last build it's not a random number, but 0xc0000005 error (-1073741819 is signed 32-bit representation for 0xc0000005) — Access Violation https://msdn.microsoft.com/en-us/library/cc704588%28d=lightweight,l=en-us,v=PROT.10%29.aspx .

PyQt4 tests pass, but give warnings.

Are you sure? I see here that they still fails: https://ci.appveyor.com/project/harvimt/quamash/branch/master/job/mlr059kahlpose55

— Reply to this email directly or view it on GitHub https://github.com/harvimt/quamash/issues/54#issuecomment-157915532.

[harvimt]

(maybe I should merge that into master before I close the issue)

On Wed, Nov 18, 2015 at 5:19 PM, Mark Harviston infinull@gmail.com wrote:

you need to look at the fix-appveyor branch: https://ci.appveyor.com/project/harvimt/quamash/build/1.0.157

On Wed, Nov 18, 2015 at 5:17 PM, Vladimir Rutsky <notifications@github.com

wrote:

PySide tests show as passing but return a random number as a status code. (same as before)

Looking at last build it's not a random number, but 0xc0000005 error (-1073741819 is signed 32-bit representation for 0xc0000005) — Access Violation https://msdn.microsoft.com/en-us/library/cc704588%28d=lightweight,l=en-us,v=PROT.10%29.aspx .

PyQt4 tests pass, but give warnings.

Are you sure? I see here that they still fails: https://ci.appveyor.com/project/harvimt/quamash/branch/master/job/mlr059kahlpose55

— Reply to this email directly or view it on GitHub https://github.com/harvimt/quamash/issues/54#issuecomment-157915532.

[harvimt]

Looking at last build it's not a random number, but 0xc0000005 error (-1073741819 is signed 32-bit representation for 0xc0000005) — Access Violation https://msdn.microsoft.com/en-us/library/cc704588%28d=lightweight,l=en-us,v=PROT.10%29.aspx .

So a segfault. Hrm that's not good.

On Wed, Nov 18, 2015 at 5:19 PM, Mark Harviston infinull@gmail.com wrote:

(maybe I should merge that into master before I close the issue)

On Wed, Nov 18, 2015 at 5:19 PM, Mark Harviston infinull@gmail.com wrote:

you need to look at the fix-appveyor branch: https://ci.appveyor.com/project/harvimt/quamash/build/1.0.157

On Wed, Nov 18, 2015 at 5:17 PM, Vladimir Rutsky < notifications@github.com> wrote:

PySide tests show as passing but return a random number as a status code. (same as before)

Looking at last build it's not a random number, but 0xc0000005 error (-1073741819 is signed 32-bit representation for 0xc0000005) — Access Violation https://msdn.microsoft.com/en-us/library/cc704588%28d=lightweight,l=en-us,v=PROT.10%29.aspx .

PyQt4 tests pass, but give warnings.

Are you sure? I see here that they still fails: https://ci.appveyor.com/project/harvimt/quamash/branch/master/job/mlr059kahlpose55

— Reply to this email directly or view it on GitHub https://github.com/harvimt/quamash/issues/54#issuecomment-157915532.

[harvimt]

this is now in master

[rutsky]

Thanks!