Appveyor build failing #9176
Labels
Milestone
Comments
6 tasks
|
It should be noted, though, that PR passed AppVeyor at the time. |
|
There's #9282 that I just opened--at least as far as At any rate while I think I have the failing test identified...I have no idea why it's causing the problem. |
|
@cgohlke these exceptions are properly handled (other way python would crash), I do not why, but |
|
Adding '-v' seems to also change the behavior #9229 |
|
So to clarify, does it look like this is related to the changes I made, or unrelated? If people suspect my modifications may have caused some of the problems I can try to find some time to dig into it. |
Crash 1I have some success in reproducing the crash in And it happens only with specific matplotlib/src/ft2font_wrapper.cpp Line 349 in 062beb6 Stack trace
Crash 2This crash happened to me several times upon stopping pytest with Line 136 in 062beb6 matplotlib/src/ft2font_wrapper.cpp Line 364 in 062beb6 Stack trace
|
|
@Kojoley are your builds using Appveyor or freetype built with setup_external_compile.py? In that case, crashes are not really unexpected. See comments at #6969 (diff) |
|
@cgohlke I cannot catch what are you saying. Freetype builds successfully by |
|
When using |
|
You are right about different runtimes and it might be a root cause if |
|
Yes, the |
|
No, not VS2015 (I do not even have it), look here https://wiki.python.org/moin/WindowsCompilers Update: Actually it does not matter, VS2015 and VS2017 provides same compiler version of 14.0 |
6 tasks
|
It was really not worth it, but... Today it does not crash, but at the end I got |
Are the file pointer related crashes gone or not when using 2015 Build Tools? |
|
The python exits with error code -1073740777 but does not actually crashes. Line 144 in 062beb6 |
I do it that way: @set /a x=0
:begin
@set /a x+=1
set PYTHONHASHSEED=%x%
python tests.py -v lib/matplotlib/tests/test_agg.py lib\matplotlib\tests\test_animation.py lib\matplotlib\tests\test_artist.py
goto begin
I already do it, otherwise I could not get a crash line, only disassembled code mess. I wanted to try to build with debug python, have downloaded it but it crashes on
Stack trace
Now I suspect that something wrong with python itself or I just do not understand what happens at all. |
|
Thanks for both of your work on this! Is it time to escalate to the cpython folks? |
Using a constant
You can use remote debug or download a crash dump to host and load it in a debugger.
I think at least one thing we can do before escalating this to cpython devs is to check for different python versions. |
|
I now can reproduce a crash and stack trace on Python 3.4.4, 64-bit, Windows 7. Basically the same as |
Output log |
|
FINALLY: It is a CPython bug, and it is already fixed in upstream python/cpython#1224 |
|
How about removing Python 3.5 from Appveyor testing as this bug will never be fixed in official CPython 3.5 binaries? Python 3.6.3 contains this bug fix and passes Appveyor testing. Also, the |
|
Only v3.7.0a1 has that fix at this point. |
|
|
Maybe I was fooled by GitHub |
|
Python 3.6.3 contains this bug fix. Unfortunately it also contains a faulthandler version that prints all the caught C++ exception messages. Probably it should be disabled on Appveyor for that version? |
|
Indeed Python 3.6.3 does have it. |
As we see, on windows faulthandler is completely useless, so I would just remove it from appveyor, for a while at least. |
|
Can we replace 3.5 by 3.4 on Appveyor? |
3.4 contains that bug too. |
|
I am for removing 3.5 from appveyor, but What about faulthandler and C++ exceptions there is a thread https://bugs.python.org/issue31701 and a fix python/cpython#3929 |
|
If I understand correctly we did not see the bug before because it's due to a change in the CRT? or something else? |
|
@anntzer I had seen problems with descriptors a year ago and that's why |
|
👍 to dropping 3.5 from the test matrix and documenting this issue in the windows install docs. |
|
Thanks a lot to everyone who chased this one! |
pauldmccarthy
added a commit
to pauldmccarthy/staged-recipes
that referenced
this issue
Jun 17, 2018
pauldmccarthy
added a commit
to pauldmccarthy/staged-recipes
that referenced
this issue
Jun 18, 2018
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
The appveyor build is failing fairly consistently on the python
3.5build (eg. see https://ci.appveyor.com/project/matplotlib/matplotlib/build/1.0.2957/job/dqhano9lykge7ffb#L1375). The actual failure is somewhere intest_artist:The text was updated successfully, but these errors were encountered: