pmxbot IRC Log Viewer

[02:25:09] <pmxbot> jaraco pushed 3 commits to setuptools (https://bitbucket.org/pypa/setuptools/) :

[02:25:09] <pmxbot> Bumped to 17.0 in preparation for next release.

[02:25:10] <pmxbot> Added tag 17.0 for changeset 8e5624096101

[02:25:10] <pmxbot> Bumped to 17.1 in preparation for next release.

[18:46:01] <lifeless> dstufft: hmm

[18:46:09] <lifeless> dstufft: I'm seeing a regression in marker support in 7.0.1

[18:46:49] <dstufft> lifeless: a regression from 7.0?

[18:47:21] <lifeless> from 6

[18:47:25] <lifeless> just filing now

[18:47:58] <lifeless> there

[18:48:43] <lifeless> same in 7.0.0

[18:48:45] <sigmavirus24> lifeless: fenced code blocks make everything easier to read

[18:48:47] <sigmavirus24> ;)

[18:49:00] <lifeless> sigmavirus24: whats a fenced code block ?

[18:49:45] <sigmavirus24> lifeless: https://help.github.com/articles/github-flavored-markdown/#fenced-code-blocks

[18:50:01] <sigmavirus24> put ``` before and after text your'e pasting from the terminal (or code-blocks)

[18:50:14] <dstufft> lifeless: does https://github.com/pypa/pip/pull/2832 fix it?

[18:50:38] <lifeless> sigmavirus24: thanks

[18:50:44] <lifeless> sigmavirus24: I fixed with lots of spaces :/

[18:50:54] <lifeless> ish, no. ``` time

[18:51:05] <sigmavirus24> heh

[18:51:14] <lifeless> nope

[18:51:16] <sigmavirus24> fenced code blocks are nicer than indenting

[18:51:18] <lifeless> didn't fix. FML

[18:51:30] <sigmavirus24> ``` needs to be on its own line

[18:51:32] <lifeless> finally

[18:51:35] <sigmavirus24> Might need a blank line before hand

[18:51:45] <sigmavirus24> !m lifeless

[18:51:45] <pmxbot> you're doing good work, lifeless!

[18:51:54] <dstufft> lifeless: "didn't fix" meaning the fenced code blocks, or the PR ;P

[18:52:05] <sigmavirus24> dstufft: yes

[18:52:06] <sigmavirus24> =P

[18:52:33] <lifeless> dstufft: was looking at the editing

[18:52:40] <lifeless> dstufft: the PR changes freeze

[18:52:58] <lifeless> dstufft: this is in the guts; its throwing from add_requirement

[18:54:02] <lifeless> what sort of quoting are we needing? That seems like it will make it hard to write one requirements file for 6 and 7

[18:54:21] <lifeless> dstufft: or put another way, is the need for quoted env markers deliberate or an accident?

[18:54:24] <dstufft> umm, I'm not sure, qwcode changed that and I didn't pay much attention to it

[18:54:30] <lifeless> qwcode: ^

[18:58:53] <qwcode> lifeless, it was a byproduct of replacing the manual parsing with optparse

[18:59:31] <lifeless> ok, so happy to fix it ?

[18:59:40] <lifeless> I can see if the patch is low hanging

[18:59:48] <qwcode> fix what?

[18:59:58] <lifeless> rephrasing

[19:00:03] <qwcode> make it work w/o quoting you mean

[19:00:09] <lifeless> do you have any objection to a 7.0.2 which accepts the same requirements files 6.x did

[19:00:16] <qwcode> no

[19:00:19] <lifeless> cool

[19:00:36] <qwcode> well, accepts both formats

[19:00:51] <lifeless> mmm

[19:00:58] <lifeless> lets see what the difficulty is first

[19:01:05] <lifeless> there will be many more files in 6 format than 7

[19:01:12] <lifeless> of the small # that use markers at all that is

[19:03:19] <lifeless> qwcode: given that optparse is mangling the line, from a quick look it looks like unwinding the optparse patch is probably the cleanest thing

[19:03:23] <lifeless> as in

[19:03:26] <lifeless> leave optparse there

[19:03:39] <lifeless> but take the original line

[19:04:13] <lifeless> hmm, I'm wrong. its easy

[19:04:17] <lifeless> shlex.split is whats mangling it.

[19:04:36] <qwcode> I look at your PR later.. I can't really discuss now.. trying to wrap up work stuff

[19:04:45] <lifeless> kk

[19:04:55] <lifeless> simple shlex.split(line) -> line.split() seems to work

[19:05:05] <lifeless> will push up a dedicated test + that in a few minutes

[19:05:57] <qwcode> but keep this in mind. https://github.com/pypa/pip/pull/2832, which is unmerged right now

[19:08:22] <lifeless> ok

[19:08:30] <lifeless> I think the bug was introduced in 55e7bd3daa54b51392c805165dc7207f96f1f944

[19:08:57] <lifeless> which introduced the indirect through args

[19:12:02] <lifeless> all the shlex stuff was added in mar 2015

[19:12:19] <lifeless> so I'm going to call that the root cause and feel ok winding that back.

[19:13:37] <lifeless> (and in fact, this will fix a bug there, because a line

[19:13:44] <lifeless> --install-options --prefix=/opt

[19:13:46] <lifeless> sorry

[19:13:49] <lifeless> --install-options "--prefix=/opt"

[19:14:40] <lifeless> is equivalent to --install-options --prefix=/opt

[19:15:08] <lifeless> so the possible reason for shlex splitting, won't actually fix such things to be more or less embeddable

[19:24:51] <lifeless> dstufft: do you want things for stable pushed to develop first, or master, or both ?

[19:25:07] <dstufft> lifeless: develop

[19:25:27] <lifeless> oops

[19:25:43] <dstufft> I'll manually pull it into a release branch and propose a rolled up PR of all the fixes against master

[19:26:40] <lifeless> dstufft: qwcode: https://github.com/pypa/pip/pull/2841

[19:26:52] <lifeless> passes all the unit tests, dunno about functional yet

[19:28:34] <qwcode> lifeless, what about cases like --install-option="--install-purelib=/directory"

[19:28:54] <qwcode> I think that was the original motivation

[19:29:12] <lifeless> qwcode: so thats broken either way

[19:29:21] <lifeless> qwcode: let me whip up a quick demo

[19:29:58] <lifeless> >>> split('--install-option="--install-purelib=/path/to directory with spaces"')

[19:30:01] <lifeless> ['--install-option=--install-purelib=/path/to directory with spaces']

[19:30:10] <lifeless> (thats shlex.split)

[19:30:51] <lifeless> if there are no spaces in the path, then we don't need the quotes at all - --install-option=--install-purelib=/path/to

[19:31:21] <lifeless> however - I can see

[19:31:32] <lifeless> split('--install-option "--install-purelib=/path/to directory with spaces"')

[19:31:32] <lifeless> ['--install-option', '--install-purelib=/path/to directory with spaces']

[19:31:57] <lifeless> that would work - but note that the quotes around the --install-purelib etc are preserved in the first case and gone in the second

[19:32:22] <lifeless> erm

[19:32:33] <lifeless> or I'm on crack. sorry, 3 1/2 year old sitting on my head right now

[19:36:10] <lifeless> ok, so yes, we should support that

[19:37:21] <qwcode> sorry, again, I can't focus right now... too many balls in there. I'll comment later, when I'm not distracted

[19:37:27] <lifeless> thats fine

[19:37:36] <lifeless> you've been helpful enough :)

[20:26:23] <lifeless> qwcode: dstufft: pushed up a fixed version that should work sanely.

Log file Viewer

Help | Karma | Search:

#pypa-dev logs for Friday the 29th of May, 2015