New package: sane-airscan-0.99.24

[pudiva]

General

• [?] This is a new package and it conforms to the quality requirements

Have the results of the proposed changes been tested?

• I use the packages affected by the proposed changes on a regular basis and confirm this PR works for me
• I generally don't use the affected packages but briefly tested this PR

[ericonr]

Duplicate of #29392, but I haven't looked at the difference between the templates yet.

[oynqr]

#29392 uses meson and this uses make.
It seems like both are supported by upstream and other distros are not largely favoring one either, with Arch and Fedora using make and Nix and ALT using meson.

[pudiva]

In case people prefer to go with my version (make) rather than #29392 (meson), I just added the line to shlibs that I had forgotten...

...my template had to define the do_install() because the gnu-makefile style defines the variable STRIP=true, which breaks this particular pkg. I usually favour make over other build systems because it's a standard tool ~ which means less dependencies ~ but since void already suports meson, then perhaps it's better to go with it, as it requires less template code to maintain.

[ericonr]

It's okay to hardcode these in the command. What you might be able to do is add make_install_args="STRIP=echo" or use some conditionals to get the correct strip for cross and native compilation there.

[pudiva]

@ericonr, thanks for the suggestion! Indeed I could get around defining a custom do_install() by setting $make_install_args="STRIP=-s"...

However, Makefile uses $(STRIP) as options to install:
https://github.com/alexpevzner/sane-airscan/blob/master/Makefile#L118-L132

and defaults to -s (which instructs install to strip):
https://github.com/alexpevzner/sane-airscan/blob/master/Makefile#L30-L35

so, for cross-compilation, I think we would need to override $(INSTALL) with target-linux-gnu-install ~ which could be a nice addition to the gnu-makefile build style, as it already does many similar overrides for CC, LD and friends.

What do you think?

[pudiva]

It's actually install --strip-program=target-linux-gnu-strip. Would be a bit strange to have platform-specific install...

[ericonr]

We can do STRIP=-s, because we have an install wrapper that removes the flag anyway :p So you don't need to care about passing the correct --strip-program :)

We do the stripping ourselves, to generate debug packages. I quite dislike build systems that try to strip binaries themselves, tbh.

[ericonr]

Maybe report it upstream, but the makefile version is missing the unversioned symlink:

/usr/lib/sane/libsane-airscan.so -> /usr/lib/sane/libsane-airscan.so.1

[pudiva]

Maybe report it upstream, but the makefile version is missing the unversioned symlink:

/usr/lib/sane/libsane-airscan.so -> /usr/lib/sane/libsane-airscan.so.1

I submitted a PR, but upstream didn't like the unversioned symlink, and also shared that they don't really maintain the meson build, but just try to not break it ~ sounds like they prefer make. Their full response:

Nobody links against these drivers directly, they are always loaded via sane-dll, which doesn't use unversioned symlink. So I believe that SANE itself creates these links only for legacy reasons and there is no need to reproduce this behaviour in the independent package.

Official packages for Fedora, Debian and Ubuntu come without this symlink and it causes no problems and no complains.

Regarding meson build, it was contributed by my users and I don't maintain it, I only keep it here for everybody's convenience and trying not to break when something changes in the project.

So really thank for your contribution, but in this particular case I don't think we really need it.

alexpevzner/sane-airscan#132 (comment)

[ericonr]

In that case, I think we should go with the Makefile based build.

[ericonr]

Could you make the description closer to what's used by upstream?

Scanner Access Now Easy - universal driver for eSCL (Apple AirScan) and WSD

[pudiva]

✨

[ericonr]

Wooh!