Issue 4288: parsermodule and grammar variable (original) (raw)

This issue has been migrated to GitHub: https://github.com/python/cpython/issues/48538

classification

process

Created on 2008-11-09 12:34 by rpetrov, last changed 2022-04-11 14:56 by admin. This issue is now closed.

Messages (8)
msg75652 - (view)	Author: Roumen Petrov (rpetrov) *	Date: 2008-11-09 12:34
After synchronization of my mingw32 cross-compilation environment with trunk some of tests fail. The reason is that parsermodule fail to link. Please check build on officially supported platform: MSVC and cygwin. The attached patch (parser-grammar.patch) solve issue in my environment.
msg75657 - (view)	Author: Christian Heimes (christian.heimes) *	Date: 2008-11-09 18:19
As far as I know neither cygwin nor MinGW32 are falling under the categoy of first class citizens. MinGW32 is only officially supported to build extension modules. The compilers aren't used by our build bots, too. In order to make both cygwin builds and MinGW32 compiler major platforms somebody has to step up and constantly provides testing and patches.
msg75658 - (view)	Author: Andy (kirkshorts)	Date: 2008-11-09 18:51
looks like it might be a similar root issue to the one I raised in #4279. Looks like this patch breaks the data hiding that I think has been attempted :-( though it doesn't mess with setup.py in the way mine does :-) Don't know how Christian's comment affects either of our patches though (still an uber newbie to the whole contributing to Python thing ;-) )
msg75659 - (view)	Author: Christian Heimes (christian.heimes) *	Date: 2008-11-09 19:00
Andy: Of course we like to support Cygwin and MinGW32. I wanted to make clear that the platforms aren't top priority. They aren't regularly tested by any of the core developers and build bots. Regarding data hiding, you are correct. I'm more fond of your solution #4279
msg75660 - (view)	Author: Christian Heimes (christian.heimes) *	Date: 2008-11-09 19:04
Oh, I retract everything I said about your patch. I didn't read your patch carefully enough. With your patch the parser module wouldn't use the same grammer as the rest of Python. That might be an issue.
msg75661 - (view)	Author: Andy (kirkshorts)	Date: 2008-11-09 19:31
Christian: Cool, thanks for the feedback d00d - it took longer than i though to get what I predicted :-) No worries on the whole "core target" platform thing - I understand it perfectly, had the same issue for work related things: too many platform and too few test resources :-( And sadly I am in not a position to gift any to the cause - sorry. Can I ask how it would mean that parser would get a different grammar? In the interests of a newbie learning if you don't mind (maybe post the reply to #4279 as I guess it is more relevant there than here)
msg75662 - (view)	Author: Andy (kirkshorts)	Date: 2008-11-09 19:42
Christian: sorry my 'find' kung fu is weak :-( :-$ I see why. Will work on a better patch.
msg79593 - (view)	Author: Martin v. Löwis (loewis) *	Date: 2009-01-11 09:35
This is now fixed with the resolution to issue 4279.