[Ffmpeg-devel] [PATCH] mingw and temorary files

Rich Felker dalias
Wed Mar 29 01:04:55 CEST 2006


On Tue, Mar 28, 2006 at 02:28:45PM -0800, Corey Hickey wrote:
> Corey Hickey wrote:
> >>>Ah, I didn't see those were the headers for installation -- I thought 
> >>>they were just general-purpose headers. Ok, I've removed that hunk.
> >>I'm against this whole tempfile thing. Why doesn't xvid just use the
> >>user-configured name for a 2pass file if it needs one? Or is this used
> >>for something else?? There's been this assumption that a temp file is
> >>necessary with no explanation of why, and IMO it's extremely bad for a
> >>_codec_ to create files on disk, temporary or not, if not requested.
> >
> >I can't really comment on that one way or another. All I know is:
> >
> >(a) There are two places in the codebase where temporary files are 
> >created. One has a workaround for MinGW; the other doesn't.
> >
> >(b) Compilation fails on MinGW because there is no workaround.
> >
> >All I did was centralize the tempfile creation code, along with the 
> >workaround. Now, if it is possible to avoid tempfile creation 
> >altogether, then, sure, I'm all for it. I don't know how to do that, 
> >though, and somehow I wouldn't expect Michael to have written that in 
> >the first place if it weren't somehow necessary.
> >
> >At the very least, this patch just makes an ugly situation a bit cleaner 
> >and more compatible.
> 
> Rich, if you object enough to not want me to apply the patch, please say 
> so; otherwise, I'll assume you reluctantly agree that it doesn't make 
> the situation any worse. I'll give this one more day, and if there's no 
> dissent I'll apply.

I won't object to you committing the patch but I honestly hope someone
will look into the issue of WHY a dependency on temp files was added
(it does not belong!!) rather than letting this issue get burried
after you commit your workaround.

Rich





More information about the ffmpeg-devel mailing list