cutter: update to 2.0.5. #35901
cutter: update to 2.0.5. #35901
Conversation
faulesocke
force-pushed
the
cutter
branch
2 times, most recently
from
6246679
to
c109e2e
Compare
| qt5-location-devel qt5-svg-devel radare2" | ||
| version=2.0.5 | ||
| revision=1 | ||
| archs="i686 x86_64" |
There was a problem hiding this comment.
Does not build for arm.
There was a problem hiding this comment.
Is this documented upstream? What's the error? Does it build for -musl?
There was a problem hiding this comment.
I haven't checked the error and haven't checked musl, this is just what I've seen from a previous pipeline run. I don't have time for fiddling around with these architectures. When someone needs it on musl or arm he can still make a PR to fix it.
There was a problem hiding this comment.
It's not acceptable to just disable a package for some architectures without a good reason. Remove these arch restrictions from both packages so we can see where it fails.
There was a problem hiding this comment.
tree-sitter has nocross=yes. So it won't compile to arm.
There was a problem hiding this comment.
But tree-sitter can still be compiled natively on arm. In that case we should set nocross=yes instead of restricting archs, but #35922 should fix cross compilation of tree-sitter, so this line can be dropped.
srcpkgs/cutter/template
Outdated
| do_fetch() { | ||
| if [ ! -e "$wrksrc" ]; then | ||
| git clone --recurse-submodules "https://github.com/rizinorg/${pkgname}" $wrksrc | ||
| fi | ||
| cd $wrksrc | ||
| git checkout "v${version}" | ||
| git submodule update --init --recursive | ||
| } |
There was a problem hiding this comment.
This is very ugly, can we just use release tarballs?
There was a problem hiding this comment.
They don't offer release source tarballs. I've several other packages in my pipeline with the same issue. Looks like package maintainers don't know (or don't care) that github does not include submodules into automatically generated tarballs. Probably because of things like flatpack. But other distributions don't seem to care as well: They just fetch the source from github directly, like I did.
There was a problem hiding this comment.
There is an issue for this which is unanswered since ~2 months: rizinorg/cutter#2878
I guess they have more important stuff to do and we have to go with the "ugly" solution.
There was a problem hiding this comment.
That's unfortunate, but could you do something similar to #32199?
|
What can I do to skip the |
srcpkgs/rizin/template
Outdated
| distfiles="https://github.com/rizinorg/${pkgname}/releases/download/v${version}/${pkgname}-src-v${version}.tar.xz" | ||
| checksum=eea49b396387c09d19705aab02a617cdb15682fca67f101ff2b27eef94a710e9 | ||
|
|
||
| subpackages="rizin-devel" |
srcpkgs/rizin/template
Outdated
|
|
||
|
|
||
| rizin-devel_package() { | ||
| depends="${makedepends} rizin>=${version}_${revision}" |
There was a problem hiding this comment.
| depends="${makedepends} rizin>=${version}_${revision}" | |
| depends="${makedepends} ${sourcepkg}>=${version}_${revision}" |
srcpkgs/rizin/template
Outdated
| pkg_install() { | ||
| vmove usr/include/librz | ||
| vmove usr/lib/pkgconfig | ||
| } |
There was a problem hiding this comment.
| pkg_install() { | |
| vmove usr/include/librz | |
| vmove usr/lib/pkgconfig | |
| } | |
| pkg_install() { | |
| vmove "usr/lib/*.so" | |
| vmove usr/include/librz | |
| vmove usr/lib/pkgconfig | |
| } |
There was a problem hiding this comment.
The .so files definitely DO NOT go into the -devel package.
There was a problem hiding this comment.
They are just symlinks to .so.$version, we don't need them in the main package, we only need the real versioned libraries.
faulesocke
force-pushed
the
cutter
branch
2 times, most recently
from
2d8ae56
to
d27c1eb
Compare
| distfiles="https://github.com/radareorg/${pkgname}/archive/v${version}.tar.gz" | ||
| checksum=868213d2ea0b4a29b9b03c9b605c2b2155c4c03b62735a9bd376a0dadcb4fe1b | ||
| homepage="https://cutter.re" | ||
| _translations_commit="974298653ba71b958e1b6c83f6011f5fefff6236" |
| qt5-location-devel qt5-svg-devel radare2" | ||
| version=2.0.5 | ||
| revision=1 | ||
| archs="i686 x86_64" |
There was a problem hiding this comment.
But tree-sitter can still be compiled natively on arm. In that case we should set nocross=yes instead of restricting archs, but #35922 should fix cross compilation of tree-sitter, so this line can be dropped.
Testing the changes
Local build testing