Arch Linux User Repository

Can you please fix the PKGBUILD replacing

patch -p0 -i "${srcdir}/libpng-$_apngver-apng.patch"

with:

zcat "${srcdir}/libpng-$_apngver-apng.patch" | patch -p0

as suggested before? thanks

i click download file on "Download snapshot" and run this commands:

makepkg -Acs

to fix error: extract file "src/libpng-1.6.37-apng.patch.gz" -> "libpng-1.6.37-apng.patch"

for me it's working.

zcat "${srcdir}/libpng-$_apngver-apng.patch" | patch -p0

worked for me

makepkg did not extract patch any more...

==> Extracting sources...
  -> Extracting libpng-1.6.37.tar.xz with bsdtar
==> Starting prepare()...
patch: **** Can't open patch file /tmp/makepkg/mingw-w64-libpng/src/libpng-1.6.37-apng.patch : No such file or directory
==> ERROR: A failure occurred in prepare().
    Aborting...

Hello, the build (both yay and manual) is failing for me due to the 'patch' command in prepare() not finding the uncompressed patch file.

Not sure why the patch is not actually uncompressed, because I know makepkg uncompresses sources in some cases.

Anyway, I fixed the build by replacing the 'patch' command in prepare with:

zcat "${srcdir}/libpng-$_apngver-apng.patch" | patch -p0

and with regular packages archlinux sets its own flags too, hence the 'buildflags' option, as they're not compatible with mingw

Well, I do understand the flags - basically they are all about hardening security and as such are actually reasonable - I was just wondering where they were coming from and why one would set them automatically instead of letting users set them as they like, but I suppose Fedora is a reasonable reference, yes, so I'll start using them for my packages as well, thanks.

I don't know the significance of all "-g -O2" is obvious, FORTIFY_SOURCE adds security to basic libc function against overflow etc, ...
Basically it's the ones of Fedora, and it's doesn't seem unreasonable flags.
xan.

Although you put them there yourself... I'd still like to know what's the reasoning behind using these flags and especially why it makes sense to force users to set these flags...

OK, they are suggested in the MinGW package guidelines... well I suppose it can't hurt...