(partial?) fix for Misc/python-config.in

[levinsm]

BPO	1523356
Nosy	@loewis, @birkenfeld
Files	python-config-patch: patch

^{Note: these values reflect the state of the issue at the time it was migrated and might not reflect the current state.}

Show more details

GitHub fields:

assignee = 'https://github.com/birkenfeld'
closed_at = <Date 2006-07-24.13:29:07.000>
created_at = <Date 2006-07-16.11:41:04.000>
labels = []
title = '(partial?) fix for Misc/python-config.in'
updated_at = <Date 2006-07-24.13:29:07.000>
user = 'https://bugs.python.org/levinsm'

bugs.python.org fields:

activity = <Date 2006-07-24.13:29:07.000>
actor = 'georg.brandl'
assignee = 'georg.brandl'
closed = True
closed_date = None
closer = None
components = ['None']
creation = <Date 2006-07-16.11:41:04.000>
creator = 'levinsm'
dependencies = []
files = ['7408']
hgrepos = []
issue_num = 1523356
keywords = ['patch']
message_count = 3.0
messages = ['50704', '50705', '50706']
nosy_count = 3.0
nosy_names = ['loewis', 'georg.brandl', 'levinsm']
pr_nums = []
priority = 'normal'
resolution = 'accepted'
stage = None
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue1523356'
versions = ['Python 2.5']

[levinsm]

I've found that the current output from python-config
isn't quite right for compiling and linking an
application with an embedded interpreter on platforms
that need $(SYSLIBS). The attached patch removes
extraneous directories from the include path and adds
the missing libraries, which was enough to make it
work for me in my limited testing.

But even with these changes, I haven't quite managed
to convince myself that python-config is currently
general enough to work on all platforms. Couldn't
$(LINKFORSHARED) also be required in some cases, for
example? And perhaps there are other missing flags
I haven't thought of.

I feel it's worth double-checking because python-config
strikes me as a potentially convenient tool, but only
if its output is known to be correct: If a broken
version ever became widely installed in the wild then
even if it were to be fixed in some future release,
the authors of other applications would forever after
need to worry about distinguishing between good and
bad versions of python-config before relying on its
output, which would undermine its convenience.
(Also: if/when it is included in a release, a man page
and some mention of it in Doc/ext would be a plus. :-)

I hope this is helpful. Thanks for all your great work!

[loewis]

Logged In: YES
user_id=21627

Georg, can you please take a look?

[birkenfeld]

Logged In: YES
user_id=849994

Thanks for the patch, applied as rev. 50800.