tags 431939 - patch
thank you
Martín Ferrari wrote:
> llgal runs identify in each image in each run, just to verify that these
> are, in fact, images. This wastes a lot of resources, as identify is
> very slow, and can take ages on big directories. Most surely this can be
> better done with file, but for the time being, I'm just skipping the
> check (as is currently done with movies). Patch below.
>
> --- llgal.orig 2007-07-06 02:17:41.000000000 -0300
> +++ /usr/bin/llgal 2007-07-06 02:04:43.000000000 -0300
> @@ -242,6 +242,7 @@
> }
>
> sub check_image {
> + return 0;
> my $filename = shift ;
> my ($status, @output) = Llgal::Utils::system_with_output
> ( "check \"$filename\" validity",
>
Well, this patch completely breaks the check, making llgal fail to skip
corrupted images and files that are wrongly named with an image-like
extension, so the patch is not valid at all :)
I never actually looked at how identify decreases performance since I
don't often use llgal on large directory. I don't remember exactly why I
switched to identify first, but I think it was required by some perl lib
which was not dealing with non-images nicely (return an error to llgal
when retrieving image dimensions, for instance). Since I changed these
libs since then (to use exiftool), I need to test again. If exiftool can
be run on non-image without breaking anything, I'll just use it and drop
the check entirely.
Thanks,
Brice
