pmxbot IRC Log Viewer

[14:31:16] <dstufft> List of projects that are hosted off PyPI by the # of hits their /simple/ page got on 2014-5-14, not shown, the 1400 or so that didn't get any hits: https://gist.github.com/dstufft/5ebfb0d7e53194e5f89e

[15:30:06] <pmxbot> jaraco pushed 5 commits to setuptools (https://bitbucket.org/pypa/setuptools/) :

[15:30:06] <pmxbot> ensure write_manifest passes contents to write_file as unicode strings.

[15:30:06] <pmxbot> ignore .git* files

[15:30:06] <pmxbot> test_sdist: change the latin1 test to match the behavior of write_manifest

[15:30:06] <pmxbot> Improved comment in the latin1 sdist test.

[15:30:06] <pmxbot> Closing this head from Pull Request 50

[15:34:42] <dstufft> Projects which are not hosted on PyPI represented a total of 0.56% of the traffic to /simple/ on the 14th

[15:35:02] <dstufft> 0.44% of that was PIL itself

[15:35:29] <dstufft> meaning that if you exclude PIL, then projects which are not hosted on PyPI represented 0.12% of the traffic to /simple/ on the 14th.

[15:55:26] <jaraco> Hi ptt

[15:55:45] <ptt> hi :)

[15:56:02] <pthiem> their that nick might be better

[15:56:12] <jaraco> I ultimately ended up merging the changes from PR 50

[15:56:48] <jaraco> Now I'm working to reconcile those with the latest changes.

[15:57:27] <jaraco> But I'm slightly inclined to bypass 1758946183e6 as it is already captured in the other work.

[15:57:32] <pthiem> yeah sorry about that

[15:57:51] <pthiem> ok let me look

[16:00:21] <pthiem> I'm sorry, what are referencing?

[16:00:33] <pthiem> Having trouble finding it

[16:00:45] <jaraco> oh...

[16:00:52] <jaraco> I don't find it either. I thought it was in unicode_try_2

[16:01:05] <jaraco> But... I moved that to default so the hashes changed.

[16:01:08] <jaraco> I mean 25fbfd4

[16:02:09] <pmxbot> jaraco pushed 5 commits to setuptools (https://bitbucket.org/pypa/setuptools/) :

[16:02:09] <pmxbot> Merge in test changes from previous try

[16:02:09] <pmxbot> with_statements and output utf-8 output

[16:02:09] <pmxbot> Merge changes from Pull Request 50

[16:02:09] <pmxbot> Starting a unicode_utils module.

[16:02:09] <pmxbot> In corporate the old unicode with the spirit of what the FileList updating.

[16:02:25] <pthiem> Ahh right, look at the next commit, this branches was a sort of "stream of conscienceness"

[16:03:08] <pthiem> So this commit ends up getting "replaced"

[16:03:10] <jaraco> Right... so I just removed that commit from the history.

[16:03:18] <pthiem> Yeah no problem

[16:03:19] <jaraco> Take a look at the new repo.

[16:05:20] <jaraco> I'm seeing two test failures on Python 3 and one on Python 2 https://travis-ci.org/jaraco/setuptools/builds/25406138

[16:06:27] <pthiem> Yeah, travis was failing for me too. I had to put it up at 4am

[16:08:07] <jaraco> BTW, When it comes to branching strategy, I basically follow the git workflow, using bookmarks in favor of branches. I know at one time Bitbucket didn't handle that particularly elegantly, but I think they're getting good at it.

[16:10:22] <pthiem> Ok, I'll give that a try. I suppose that only critical thing is they need to get the conflict resolver updated. So I can "sync" in some default manner, instead of getting a conflict when I do a pull.

[16:11:36] <pthiem> Ok so in PY3: test_write_manifest_allows_utf8_filenames and test_manifest_is_written_with_utf8_encoding

[16:11:40] <jaraco> And I don't mean to suggest that you must do things differently. Do whatever works best for you. Since I don't often make pull requests in Bitbucket, I don't even know what the experience is.

[16:12:38] <pthiem> And in py2 jsut test_manifest_is_written_with_utf8_encoding

[16:13:00] <pthiem> np :) thx for the advice

[16:15:27] <pthiem> Ahh the test is using filelist.files.append instead of filelist.append

[16:16:02] <jaraco> I was worried about that.

[16:16:18] <jaraco> If that's all it takes to fix it, that's encouraging.

[16:16:55] <pmxbot> jaraco pushed 5 commits to setuptools (https://bitbucket.org/pypa/setuptools/) :

[16:16:55] <pmxbot> Backed out changeset: 1ceaffff2d9b (restoring b0a2fcc5275a). Ref #193

[16:16:55] <pmxbot> Merge 4 out of 5 commits from Philip Thiem's unicode_try_2. Ref #193

[16:16:55] <pmxbot> with_statement is available naturally in Python 2.6

[16:16:55] <pmxbot> Compute PY3 directly and switch on that for clarity.

[16:16:55] <pmxbot> Add a PY2 also.

[16:20:48] <pthiem> Ah you're right. I must be thinking 2.5

[16:26:41] <pmxbot> jaraco pushed 1 commit to setuptools (https://bitbucket.org/pypa/setuptools/) :

[16:26:41] <pmxbot> Use PY3 and PY2 throughout

[16:35:29] <pthiem> ahh I see

[16:35:35] <pthiem> it doesn't create the file

[16:35:50] <pthiem> so the _repair procedure removes it

[16:39:30] <pmxbot> jaraco pushed 1 commit to setuptools (https://bitbucket.org/pypa/setuptools/) :

[16:39:30] <pmxbot> Convert quiet to a context manager, so its comment is no longer complaining and to simplify the code.

[16:40:03] <pmxbot> jaraco pushed 1 commit to setuptools (https://bitbucket.org/pypa/setuptools/) :

[16:40:03] <pmxbot> Remove unused import

[16:43:47] <jaraco> pthiem, so my feeling on that is that the test is bad.

[16:44:02] <jaraco> The manifest shouldn't include a filename that doesn't exist.

[16:44:28] <jaraco> That's an explicit behavior in the implementation.

[16:44:38] <jaraco> Would you agree?

[16:46:36] <pthiem> Yes basically.

[16:47:20] <pthiem> Touching the file causes the test to pass.

[16:49:20] <pthiem> Was pushing up to travis

[16:49:48] <pthiem> All green on linux

[16:50:03] <pthiem> I'll update my branch

[16:50:12] <jaraco> I don't yet understand why that test passed before.

[16:52:29] <pthiem> Before

[16:52:57] <pthiem> the tests was directly injecting into FileList.filelist

[16:53:19] <pthiem> and the code I did, only checked for encoding errors not existance.

[16:53:22] <pthiem> existence*

[16:53:24] <jaraco> indeed

[16:53:40] <jaraco> Okay. Then I feel good about touching the file.

[16:54:01] <pthiem> :D

[16:56:01] <pthiem> Working on the various branches I have is on my to-do list for today, but I need to see if I can catch the bank before it closes. I'll double check win32 and update my build_osx branch to tip

[17:02:13] <jaraco> pthiem, I'm about out of time to work on this today too.

[17:02:35] <jaraco> But if I can get the tests green and then make a comment to the distribute ticket for those interested to review, then I'll be happy.

[17:03:52] <pthiem> err channel

[17:05:04] <pthiem> jaraco, no problem the code I have in unicode_try_2 is passing.

[17:05:32] <pthiem> anyway, chat later ;)

[17:05:53] <jaraco> !thanks pthiem

[17:05:53] <pmxbot> you're doing good work, pthiem!

[17:06:59] <pthiem> lol thx.

[17:19:03] <pmxbot> jaraco pushed 3 commits to setuptools (https://bitbucket.org/pypa/setuptools/) :

[17:19:03] <pmxbot> est_manifest_is_written_with_utf8_encoding should use the filelist's append, NOT filelist.files.append

[17:19:03] <pmxbot> Must create files for tests, else they are remoed from manifests

[17:24:19] <pmxbot> jaraco pushed 1 commit to setuptools (https://bitbucket.org/pypa/setuptools/) :

[17:24:19] <pmxbot> Add whitespace for readability.

[17:31:05] <xafer> hello qwcode

[17:35:25] <xafer> I've remerged your branch into develop but have now 3 tests failing to download wheels, any idea ? :o

[23:28:55] <pmxbot> jaraco pushed 5 commits to setuptools (https://bitbucket.org/pypa/setuptools/) :

[23:28:55] <pmxbot> Extract variable

[23:28:55] <pmxbot> Write requirements in a deterministic order.

[23:28:55] <pmxbot> Use StringIO to write out requirements. Use more common convention of adding newline to each line of the file, not just intervening lines.

[23:28:55] <pmxbot> Update requires.txt to match new, predictable order.

[23:28:55] <pmxbot> Restore Python 2 compatibility.

Log file Viewer

Help | Karma | Search:

#pypa-dev logs for Saturday the 17th of May, 2014