[issue828450] sdist generates bad MANIFEST on Windows

Steve Dower added the comment: Distutils is now deprecated (see PEP 632) and all tagged issues are being closed. From now until removal, only release blocking issues will be considered for distutils. If this issue does not relate to distutils, please remove the component and reopen it. If y

[issue828450] sdist generates bad MANIFEST on Windows

Change by Mark Lawrence : -- nosy: -BreamoreBoy ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: https://mai

[issue828450] sdist generates bad MANIFEST on Windows

Mark Lawrence added the comment: @Éric did you ever ask on distutils-sig if MANIFEST is meant to be cross-platform? -- components: -Distutils2, Windows nosy: +BreamoreBoy, dstufft versions: +Python 3.4, Python 3.5 -3rd party, Python 3.2, Python 3.3

[issue828450] sdist generates bad MANIFEST on Windows

Éric Araujo added the comment: > One cann't let Python generate MANIFEST files taking Unix-style LF as newline > endings On Windows, I think. Why? Python can open it fine, and it’s not meant for human edition, so the stupidity of notepad.exe is not a problem . > So, does it mean even though w

[issue828450] sdist generates bad MANIFEST on Windows

higery added the comment: >>I’ve just found another problem with MANIFEST files created in Windows: they >>use CRLF. One cann't let Python generate MANIFEST files taking Unix-style LF as newline endings On Windows, I think. So, does it mean even though we have already made much effort for th

[issue828450] sdist generates bad MANIFEST on Windows

Éric Araujo added the comment: I’ve just found another problem with MANIFEST files created in Windows: they use CRLF. -- ___ Python tracker ___ ___

[issue828450] sdist generates bad MANIFEST on Windows

Changes by Thomas Holmes : -- nosy: +thomas.holmes ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail

[issue828450] sdist generates bad MANIFEST on Windows

Éric Araujo added the comment: I have changed some things in your patch. There are still two issues: 1) setting os.sep to \ in the tests is not enough to trigger the bug. This means that the tests really test something only on Windows. I’ll edit them to mock the OS layer and return paths w

[issue828450] sdist generates bad MANIFEST on Windows

Changes by higery : Removed file: http://bugs.python.org/file22328/test_manifest_reading_sdist.diff ___ Python tracker ___ ___ Python-bugs-li

[issue828450] sdist generates bad MANIFEST on Windows

Changes by higery : Removed file: http://bugs.python.org/file21713/change_path_separator_in_MANIFEST.patch ___ Python tracker ___ ___ Python

[issue828450] sdist generates bad MANIFEST on Windows

higery added the comment: OK. I recreated a full version patch. I'll remove old patches. -- Added file: http://bugs.python.org/file22333/change_path_separator_fullversion.patch ___ Python tracker ___

[issue828450] sdist generates bad MANIFEST on Windows

Éric Araujo added the comment: > I just recreated this patch against version 2.7, so I'm not sure it > can be applied to all the listed versions. It’s okay, distutils in 2.7 and 3.x is very similar, I can port. > Note: there still are two pathes, one for sdist.py and another for > test_sdist.py

[issue828450] sdist generates bad MANIFEST on Windows

higery added the comment: I just recreated this patch against version 2.7, so I'm not sure it can be applied to all the listed versions. Note: there still are two pathes, one for sdist.py and another for test_sdist.py -- Added file: http://bugs.python.org/file22328/test_manifest_readi

[issue828450] sdist generates bad MANIFEST on Windows

Changes by higery : Removed file: http://bugs.python.org/file21764/test_manifest_reading_sdist_v2.diff ___ Python tracker ___ ___ Python-bug

[issue828450] sdist generates bad MANIFEST on Windows

Changes by higery : Removed file: http://bugs.python.org/file21725/test_manifest_reading_sdist.diff ___ Python tracker ___ ___ Python-bugs-li

[issue828450] sdist generates bad MANIFEST on Windows

Éric Araujo added the comment: I wanted to review the patch but it cannot be applied. It looks like your third patch applies on top of the second. Can you generate a full patch? -- status: pending -> open versions: -Python 3.1 ___ Python tracker

[issue828450] sdist generates bad MANIFEST on Windows

Changes by Éric Araujo : -- resolution: -> accepted stage: test needed -> patch review status: open -> pending ___ Python tracker ___ __

[issue828450] sdist generates bad MANIFEST on Windows

Changes by higery : Added file: http://bugs.python.org/file21764/test_manifest_reading_sdist_v2.diff ___ Python tracker ___ ___ Python-bugs-l

[issue828450] sdist generates bad MANIFEST on Windows

higery added the comment: >> I'm not sure it is necessary to use TempdirManager here to write tests >> for MANIFEST reading. >Well, the sdist command has to run on some directory where some files >exist, >right? So it’s better to do it in a temp dir that will automatically get >>cleaned up.

[issue828450] sdist generates bad MANIFEST on Windows

Éric Araujo added the comment: > I'm not sure it is necessary to use TempdirManager here to write tests > for MANIFEST reading. Well, the sdist command has to run on some directory where some files exist, right? So it’s better to do it in a temp dir that will automatically get cleaned up. >

[issue828450] sdist generates bad MANIFEST on Windows

higery added the comment: I'm not sure it is necessary to use TempdirManager here to write tests for MANIFEST reading. The attachment is the test diff file against my last patch with the latest version. Detail: Step 1: Write sample MANIFEST strings to the MANIFEST file with '/' as separator

[issue828450] sdist generates bad MANIFEST on Windows

Changes by higery : Added file: http://bugs.python.org/file21713/change_path_separator_in_MANIFEST.patch ___ Python tracker ___ ___ Python-b

[issue828450] sdist generates bad MANIFEST on Windows

Changes by higery : Removed file: http://bugs.python.org/file21697/change_path_separator_in_MANIFEST.patch ___ Python tracker ___ ___ Python

[issue828450] sdist generates bad MANIFEST on Windows

Éric Araujo added the comment: Thanks for the patch. Follow the “review” link above to find my review. +1 on making the smallest possible change. The paths inside the Filelist/Manifest objects can continue to use os.sep, we only care about write_file for this bug. Oh, and also read_file: c

[issue828450] sdist generates bad MANIFEST on Windows

Changes by higery : Added file: http://bugs.python.org/file21697/change_path_separator_in_MANIFEST.patch ___ Python tracker ___ ___ Python-b

[issue828450] sdist generates bad MANIFEST on Windows

Changes by higery : Removed file: http://bugs.python.org/file21691/change_path_separator_in_MANIFEST.patch ___ Python tracker ___ ___ Python

[issue828450] sdist generates bad MANIFEST on Windows

higery added the comment: It may be just necessary to hack the write_manifest funtion in sdist.py- replace all '\' in MANIFEST with '/', thus it will not have other bad side effects, for instance, it would not change the content of self.filelist and people can also use '/' in MANIFEST templat

[issue828450] sdist generates bad MANIFEST on Windows

Changes by higery : Removed file: http://bugs.python.org/file21667/test_sdist.diff ___ Python tracker ___ ___ Python-bugs-list mailing list U

[issue828450] sdist generates bad MANIFEST on Windows

Changes by Santoso Wijaya : -- nosy: +santa4nt ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe: http://mail.pyt

[issue828450] sdist generates bad MANIFEST on Windows

higery added the comment: OK. I used this method just because I thought '\' is a special character and if it's in a file path line, then it must be the separator. As you say, it may be not that clear for others to know what does this test do. -- _

[issue828450] sdist generates bad MANIFEST on Windows

Éric Araujo added the comment: Thanks! I would also like it if you could use a more specific test, comparing a line with a path instead of using the overly broad assertIn, to make the intent of the test clearer. -- assignee: tarek -> eric.araujo _

[issue828450] sdist generates bad MANIFEST on Windows

higery added the comment: Yes, the test fails and the output msg is: AssertionError: '\\' unexpectedly found in '# file GENERATED by distutils, do NOT edit\nREADME\nsetup.py\nsomecode\\__init__.py\n' It means that distutils generates MANIFEST with '\' as file path separator. OK, I'll try to

[issue828450] sdist generates bad MANIFEST on Windows

Éric Araujo added the comment: setuptools sdist uses a wholly different machinery than distutils, so it’s a red herring. Have you tested that your patch does reproduce the bug? From the diff header, I see that you’ve patched your installed Python instead of using a developpers’ environment.

[issue828450] sdist generates bad MANIFEST on Windows

higery added the comment: I will join GSOC2011 and I find this bug is a good test for me to submit my patch as a scoring judgment. I have created a diff file to confirm this bug, but setuptools may have already fix it, because when using 'python setup.py sdist' to create the source distribut

[issue828450] sdist generates bad MANIFEST on Windows

Éric Araujo added the comment: Adding the “easy” keyword to encourage someone to write a test. -- keywords: +easy versions: +3rd party ___ Python tracker ___ __

[issue828450] sdist generates bad MANIFEST on Windows

Changes by Mark Lawrence : -- type: -> behavior versions: +Python 2.7, Python 3.1, Python 3.2 -Python 2.5, Python 2.6 ___ Python tracker ___ ___

[issue828450] sdist generates bad MANIFEST on Windows

Éric Araujo added the comment: Thanks for the report, and sorry noone replied earlier. Could you perhaps make a diff against test_sdist.py to confirm this bug? -- components: +Distutils2 nosy: +merwok stage: -> unit test needed ___ Python tracker

[issue828450] sdist generates bad MANIFEST on Windows

Changes by Daniel Diniz : -- assignee: -> tarek nosy: +tarek ___ Python tracker ___ ___ Python-bugs-list mailing list Unsubscribe:

[issue828450] sdist generates bad MANIFEST on Windows

Changes by Christian Heimes: -- components: +Windows versions: +Python 2.5, Python 2.6 Tracker <[EMAIL PROTECTED]> ___ Python-bugs-l