[issue35693] test_httpservers fails

Irit Katriel added the comment: Closing as 3.7 is no longer maintained and this problem was not reproduced in 3.8. -- nosy: +iritkatriel resolution: -> out of date stage: test needed -> resolved status: open -> closed ___ Python tracker

[issue35693] test_httpservers fails

Jorge Ramos added the comment: E:\RepoGiT\3.7>"E:\RepoGiT\3.7\PCbuild\amd64\python.exe" -u -Wd -E -bb -m test -v test_httpservers == CPython 3.7.2+ (heads/3.7:e1259886ab, Jan 13 2019, 19:16:24) [MSC v.1916 64 bit (AMD64)] == Windows-10-10.0.17763-SP0 little-endian == cwd:

[issue35693] test_httpservers fails

Jorge Ramos added the comment: the output is stored on the file run.txt above -- ___ Python tracker ___ ___ Python-bugs-list

[issue35693] test_httpservers fails

Steve Dower added the comment: > It is true that I worry about failing tests during PGO training because they > stop me from compiling Python -with- PGO. They shouldn't prevent you from compiling with PGO... even if they cause Python to crash, you should be able to compile still. How is it

[issue35693] test_httpservers fails

Zachary Ware added the comment: Test failures during the PGO training (*not* those after the final PGO compilation) should be ignored; if they're not, that's an issue that should be fixed. -- ___ Python tracker

[issue35693] test_httpservers fails

Change by Jorge Ramos : -- stage: test needed -> resolved status: open -> closed ___ Python tracker ___ ___ Python-bugs-list

[issue35693] test_httpservers fails

Jorge Ramos added the comment: It is true that I worry about failing tests during PGO training because they stop me from compiling Python -with- PGO. I also got here to report my experience and problems back to you guys but, if the problems are known, well, there's nothing much I can do

[issue35693] test_httpservers fails

Steve Dower added the comment: Defining new training sets for PGO would be a new job (that you could volunteer to do if you like). The test suite is convenient because when you clone our GitHub repository, it's already there. Another training suite would have to be installed somehow as part

[issue35693] test_httpservers fails

Jorge Ramos added the comment: Thanks! About the alternate set of training data, how (where) do I find it? and about the layouts, can you point me in the right direction as to learn about them, into what they do? or try to achieve? and how to use them? Quoting Steve Dower (

[issue35693] test_httpservers fails

Steve Dower added the comment: The tests mostly fail in the PGO run because they are highly sensitive to the source tree layout. We've never fixed them because they aren't that important to the PGO run - failure cases help with training - and the real tests pass fine under their more

[issue35693] test_httpservers fails

Jorge Ramos added the comment: I have realized that results from rt.bat differ from those when running buildrelease.bat (see example below) Why do the tests run OK in rt.bat but fail in buildrelase.bat? To my best understanding, the tests are summoned directly via rt.bat but also from

[issue35693] test_httpservers fails

Change by Jorge Ramos : -- versions: +Python 3.6 ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe:

[issue35693] test_httpservers fails

Zachary Ware added the comment: Certainly! Bugfixes, in library or test code, are still welcome on 3.7. -- stage: -> test needed versions: +Python 3.7 -Python 3.6 ___ Python tracker

[issue35693] test_httpservers fails

Jorge Ramos added the comment: Ok, thank you. Here are my findings: steps taken: 1) download today's version of 3.7.2 and 3.8.0 alpha 0 2) navigate to tools/msi 3) run build.bat -x64 (to create amd64 folder under pcbuild) 4) navigate to pcbuild 5) run rt.bat -x64 -v test_httpservers

[issue35693] test_httpservers fails

Zachary Ware added the comment: No need to paste the output; it's linked between the form at the top and the first message. It's not shown on the issue list view because it's not a patch. The next step is to see if it still fails when you build from the `master` branch since 3.6 is now

[issue35693] test_httpservers fails

Jorge Ramos added the comment: The thread doesn't display the attachment icon in bugtracker webpage, so I will copy the output here: E:\RepoGiT\3.6>PCbuild\rt.bat -x64 -v test_httpservers Deleting .pyc files ... 0 .pyc deleted Cleaning _pth files ...

[issue35693] test_httpservers fails

Jorge Ramos added the comment: The test failed, the results on attached file! -- Added file: https://bugs.python.org/file48037/failed.txt ___ Python tracker ___

[issue35693] test_httpservers fails

Jorge Ramos added the comment: Got it, i'll get back with the results -- ___ Python tracker ___ ___ Python-bugs-list mailing list

[issue35693] test_httpservers fails

Zachary Ware added the comment: Try running `rt.bat -x64 -v test_httpservers`, which will run just that test in verbose mode. If that happens to pass, try `rt.bat -x64 -wW`, which will show the verbose output of a failed test and retry just the failed test at the end of the full run.

[issue35693] test_httpservers fails

New submission from Jorge Ramos : when running test_httpservers fails: 0:04:53 [171/407] test_httpservers E:\RepoGiT\3.6\lib\socket.py:144: ResourceWarning: unclosed _socket.socket.__init__(self, family, type, proto, fileno) E:\RepoGiT\3.6\lib\test\support\__init__.py:1542: