Closed GoogleCodeExporter closed 8 years ago
Sounds fair.
Are we sure that $(OS) == Windows*
for all versions of Windows ?
Original comment by yann.col...@gmail.com
on 23 Sep 2013 at 11:13
mmmh, just made a test, and I got :
$(OS) == MINGW32_NT-6.1
So Windows* doesn't work "as is".
Original comment by yann.col...@gmail.com
on 23 Sep 2013 at 11:24
OK, checked some docs, which recommended to use $(OS) for Windows, and uname
for *nix.
And the r104 Makefile, both are blended together, due to this line :
OS := $(shell uname)
Removing/commenting this line, now we have
$(OS) == Windows_NT
which is, I guess, the expected behavior.
(Note : the test is done on Windows 7 64-bits)
So, back to initial question :
what are the expected values of $(OS) for Windows ?
It seems Windows_NT is going to be the most common one, what about the others ?
Original comment by yann.col...@gmail.com
on 23 Sep 2013 at 11:31
The proposed test :
ifeq ($(OS),Windows*)
doesn't seem to work. It seems '*' wildcard support is not part of ifeq syntax.
I had to change the test to the following syntax (to keep the same test) :
ifneq (,$(filter Windows%,$(OS)))
This one works as intended.
Alternatively, it could have been possible to restrict the number of Windows OS
detected :
ifeq ($(OS),Windows_NT)
Original comment by yann.col...@gmail.com
on 23 Sep 2013 at 11:46
Original comment by yann.col...@gmail.com
on 23 Sep 2013 at 11:55
The following is a potential release candidate which includes your requirement
Original comment by yann.col...@gmail.com
on 24 Sep 2013 at 8:38
Attachments:
Will check out this new code. I didn't have a proper windows machine to test
on. I think that the only people who are going to be running gcc on windows
are likely to have the nt-based kernel? Thanks!
Original comment by bl...@wanelo.com
on 24 Sep 2013 at 8:35
integrated into r105
Original comment by yann.col...@gmail.com
on 25 Sep 2013 at 9:05
Original issue reported on code.google.com by
bl...@wanelo.com
on 23 Sep 2013 at 9:50