Zbigniew Jędrzejewski-Szmek wrote:
Right. So it sounds like the package could be made to
"build" very easily.
By shipping the upstream blob as is? That would not really be compliant to
Fedora packaging guidelines, would it? At least I would hope it would not
be! (This is not really firmware, WINE loads it into the regular userspace.)
The problem is that the code no longer compiles with the current cross-MinGW
toolchain, and upstream has not done anything to fix that because they still
ship their old binary.
The fact that maintainers could not do this since 2016 is IMO a good
reason to retire the package. We need to drop packages which FTBFS, FTI,
or have multiple security bugs open. Each package that fails at the mass
rebuild sucks up time of other maintainers who will try to figure out what
it doesn't build and fix it. Every package that FTI sucks up time of users
who try to install the package. Every package that has CVEs without any
response means that users (reasonably) have doubts about the overall
security of Fedora, and also obscures other issues which might be really
important. Over the years Fedora has collected a lot of "cruft". We need
to do cleanup sweeps because otherwise the distro becomes a garbage site
where it's impossible to do any kind of improvement at scale.
Retiring packages leaves the user with no software at all to fulfill their
needs. Users want the best software available to perform their tasks. That
software may be in a less than ideal state, it may even be insecure, but
that can still be better than not having the software available at all. If
the software does not install at all (or installs but does not run), that
obviously needs fixing, but if it just FTBFS and/or has some CVEs that are
not critical in the user's environment (e.g., mingw-wine-gecko is more
likely to be used to view either some local HTML files in Windows programs
or parts of a specific ISV's website than to browse the entire untrusted
Internet), having the software available is often better than having nothing
at all.
Kevin Kofler