This issue tracker has been migrated to GitHub, and is currently read-only.
For more information, see the GitHub FAQs in the Python's Developer Guide.

classification
Title: Allow use of GNU arch on Darwin
Type: behavior Stage: resolved
Components: Build Versions: Python 2.7, Python 2.6
process
Status: closed Resolution: fixed
Dependencies: Superseder:
Assigned To: ronaldoussoren Nosy List: grobian, ned.deily, ronaldoussoren
Priority: normal Keywords: needs review, patch

Created on 2010-01-16 16:33 by grobian, last changed 2022-04-11 14:56 by admin. This issue is now closed.

Files
File name Uploaded Description Edit
python-2.6.4-gnu-arch-darwin.patch grobian, 2010-01-16 16:33 Recognise output from GNU arch
Messages (6)
msg97884 - (view) Author: Fabian Groffen (grobian) Date: 2010-01-16 16:33
Configure will die on Darwin with the message

Unexpected output of 'arch' on OSX

when GNU arch is in use.  The following patch simply adds the output as given by GNU arch to allow a successful configure run.
msg97930 - (view) Author: Ned Deily (ned.deily) * (Python committer) Date: 2010-01-17 01:27
Note that with current trunk and in-the-pipeline proposed changes, configure is depending on using the Apple-supplied enhanced version of arch on 10.5+ (e.g. to force execution in 32-bit mode) so, while the patch here doesn't harm anything, it points to another potential problem: configure should probably ensure it *is* using the Apple arch.
msg97936 - (view) Author: Ronald Oussoren (ronaldoussoren) * (Python committer) Date: 2010-01-17 11:42
I agree with Ned: the correct patch is to hardcode usage of /usr/bin/arch, because that's the one the build environment is expecting to use and the only one I care to support.
msg97969 - (view) Author: Ronald Oussoren (ronaldoussoren) * (Python committer) Date: 2010-01-17 19:33
I've applied a patch to use /usr/bin/arch instead of just 'arch' in r7715 (trunk). 

I will port this to the other branches soonish.
msg97971 - (view) Author: Ned Deily (ned.deily) * (Python committer) Date: 2010-01-17 19:54
BTW, I believe there is a problem with ARCH_RUN_32BIT as it stands: as far as I can tell, unlike lipo, /usr/bin/arch requires -ppc and rejects -ppc7400 regardless if the executable is -arch ppc7400 or -arch ppc.
msg99205 - (view) Author: Ronald Oussoren (ronaldoussoren) * (Python committer) Date: 2010-02-11 13:40
Ned: I fixed the ARCH_RUN_32BIT definition for 3-way universal builds in r78150. 

Also ported the patch to 2.6, 3.2 and 3.1.
History
Date User Action Args
2022-04-11 14:56:56adminsetgithub: 51964
2010-02-11 13:40:19ronaldoussorensetstatus: open -> closed

messages: + msg99205
2010-01-17 19:54:41ned.deilysetmessages: + msg97971
2010-01-17 19:33:15ronaldoussorensetresolution: fixed
messages: + msg97969
stage: patch review -> resolved
2010-01-17 11:42:37ronaldoussorensetmessages: + msg97936
2010-01-17 01:27:52ned.deilysetnosy: + ned.deily
messages: + msg97930
2010-01-16 22:05:18r.david.murraysetassignee: ronaldoussoren

nosy: + ronaldoussoren
2010-01-16 21:21:22brian.curtinsetpriority: normal
keywords: + needs review
stage: patch review
2010-01-16 16:33:47grobiancreate