[FFmpeg-devel] [PATCH] rename time.h to avtime.h

[Michael Niedermayer]

On Sun, Oct 07, 2012 at 12:10:58PM +0200, Nicolas George wrote:
> Le sextidi 16 vendémiaire, an CCXXI, Michael Niedermayer a écrit :
> > do you have another / better suggestion on how to fix this ?
> 
> I am not convinced there is anything to fix on ffmpeg's part at all, and if
> there is, it may be completely elsewhere. As far as I understand the
> situation, the problem is twofold:
> 
> 1. The reporter is using Android's build system instead of ffmpeg's;
> 
> 2. Android's build system adds bogus -I options.
> 
> Fixing any of these issues would fix the problem. Ideally, both issues
> should be fixed.
> 
> The first issue is the reporter's responsibility. It has been claimed that
> ffmpeg's build system does not work for Android, but only by second- or
> third-hand hearsay. Actually, about last year I was able to build the ffmpeg
> libraries with configure and make, although with an awful lot of
> --extra-cflags and --extra-ldflags. Documenting that a little better may be
> a good start, but that requires understanding the intricacies of ARM build
> flags.

theres also:
http://fate.ffmpeg.org/report.cgi?time=20121007080758&slot=armv7a-android-gcc-4.4

but still, being able to be build from other environments is a good
thing and at the cost of renaming a header it seems the right long
term change to me ...


> 
> The second issue should be reported to the Android project developers. I am
> not opposed to adding workarounds for foreign bugs, but before that, I would
> like to see genuine efforts to fix the bug in the correct place, as it would
> require less maintenance for ffmpeg and fix things for other projects too.
> When the Android developers have stated that, against all evidence, they are
> doing The Right Thing and What Users Expect and closed the bug as "wontfix",
> that is another story.

I agree it should be reportet, but we also should avoid having this
header that way because it simply makes our code fragile against
half broken build environments and i suspect android is not the only
one that puts an extraneous -I path in there

[...]

-- 
Michael     GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB

The worst form of inequality is to try to make unequal things equal.
-- Aristotle
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 198 bytes
Desc: Digital signature
URL: <http://ffmpeg.org/pipermail/ffmpeg-devel/attachments/20121007/b2134423/attachment.asc>