HAVE_FDATASYNC incorrectly set for 2.3.4 under cygwin

[gregsmith]

BPO	985154
Nosy	@mwhudson

^{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 = None
closed_at = <Date 2004-08-17.03:39:27.000>
created_at = <Date 2004-07-05.02:38:00.000>
labels = ['build', 'invalid']
title = 'HAVE_FDATASYNC incorrectly set for 2.3.4 under cygwin'
updated_at = <Date 2004-08-17.03:39:27.000>
user = 'https://bugs.python.org/gregsmith'

bugs.python.org fields:

activity = <Date 2004-08-17.03:39:27.000>
actor = 'gregsmith'
assignee = 'jlt63'
closed = True
closed_date = None
closer = None
components = ['Build']
creation = <Date 2004-07-05.02:38:00.000>
creator = 'gregsmith'
dependencies = []
files = []
hgrepos = []
issue_num = 985154
keywords = []
message_count = 9.0
messages = ['21405', '21406', '21407', '21408', '21409', '21410', '21411', '21412', '21413']
nosy_count = 4.0
nosy_names = ['mwh', 'nnorwitz', 'jlt63', 'gregsmith']
pr_nums = []
priority = 'normal'
resolution = 'not a bug'
stage = None
status = 'closed'
superseder = None
type = None
url = 'https://bugs.python.org/issue985154'
versions = ['Python 2.3']

[gregsmith]

Building 2.3.4 from the tarball, 
#define HAVE_FDATASYNC 1
appears in the pyconfig.h file, and
the build of 'posixmodule' fails with
undefined symbol _fdatasync

Workaround is to delete the line
and proceed, of course, but I've
never seen this before when building
under cygwin.

There is no 'fdatasync' mentioned
in any /usr/include/.h or /usr/local/include/.h

The cygwin installation is kind of a mongrel
(4 or 5 upgrades since install, some selective)
This is the ID string from the built python:

Python 2.3.4 (#2, Jul 4 2004, 22:10:22)
[GCC 3.3.1 (cygming special)] on cygwin

Let me know of any details of the cygwin install
which might be useful.

[mwhudson]

Logged In: YES
user_id=6656

That's odd. Why does configure think you have fdatasync? Maybe
config.log will help...

[gregsmith]

Logged In: YES
user_id=292741

looked into config.log and 'configure' itself.. the compile
for fdatasync succeeded, but the script does not attempt
to link it. gcc -E led me to the following
chunk of /usr/include/sys/unistd.h:

(not inside any #ifdefs, as far as I can tell)

int     _EXFUN(fdatasync, (int __fd));

So it's declared in the header, but it's not in the
library. This symbol is defined in libc.a on a RH8.0
system, I can't find it anywhere on the cygwin system.

[mwhudson]

Logged In: YES
user_id=6656

Well, I guess that's an arguable bug in cygwin, but if you want to
supply a patch that modifies configure to do the linking test too,
I'll at least think about applying it :-)

[nnorwitz]

Logged In: YES
user_id=33168

Jason, do you know what the deal is with cygwin?

[jlt63]

Logged In: YES
user_id=86216

nnorwitz> Jason, do you know what the deal is with cygwin?

AFAICT, Greg's Cygwin setup seems suspect. Actually, he
alluded to this fact in the following:

gregsmith> The cygwin installation is kind of a mongrel...

On my Cygwin setup (i.e., 1.5.10-3), I have the following:

/usr/include/sys/unistd.h:
    #ifndef __CYGWIN__
    int     _EXFUN(fdatasync, (int __fd));
    #endif

pyconfig.h:

/* Define if you have the 'fdatasync' function. */
/* #undef HAVE_FDATASYNC */

Additionally, the official Cygwin Python package built
cleanly.

Greg, what version of Cygwin are you using? If not 1.5.10-3,
then please upgrade and try again. If so, then your setup
seems to be corrupted. Try reinstalling. Does it work any
better now?

[jlt63]

Logged In: YES
user_id=86216

There has been no follow-up from Greg in a month.
OK to close?