dgilmore changed the topic of #fedora-riscv to: Fedora on RISC-V https://fedoraproject.org/wiki/Architectures/RISC-V || Logs: https://libera.irclog.whitequark.org/fedora-riscv || Alt Arch discussions are welcome in #fedora-alt-arches
Guoguo has joined #fedora-riscv
SongsongZhang has joined #fedora-riscv
<SongsongZhang> @rwmjones_ Sorry about the previous PRs, I have corrected, Cloud you help review it?
<SongsongZhang> in google spreadsheets
SongsongZhang has quit [Quit: Leaving]
Guoguo has quit [Remote host closed the connection]
Guoguo has joined #fedora-riscv
Guoguo has quit [Ping timeout: 264 seconds]
Guoguo has joined #fedora-riscv
Guoguo has quit [Remote host closed the connection]
Guoguo has joined #fedora-riscv
Guoguo has quit [Ping timeout: 264 seconds]
Guoguo has joined #fedora-riscv
Guoguo has quit [Ping timeout: 246 seconds]
Guoguo has joined #fedora-riscv
Guoguo has quit [Remote host closed the connection]
Guoguo has joined #fedora-riscv
Guoguo has quit [Ping timeout: 256 seconds]
davidlt has joined #fedora-riscv
davidlt has quit [Ping timeout: 264 seconds]
Guoguo has joined #fedora-riscv
Guoguo has quit [Ping timeout: 246 seconds]
Guoguo has joined #fedora-riscv
Guoguo_ has joined #fedora-riscv
Guoguo has quit [Ping timeout: 268 seconds]
Guoguo_ has quit [Ping timeout: 264 seconds]
Guoguo has joined #fedora-riscv
davidlt has joined #fedora-riscv
<davidlt> rwmjones_, how to those students choose what to upstream to dist-git?
<davidlt> and things like "ExclusiveArch:" are almost the last on my list
<davidlt> rwmjones_, do you want be ACK/LGTM on all PRs? Otherwise I just plan to comment if there is anything that I would like to see changed, etc.
<davidlt> All changes should go into Rawhide and then F40, and should have a new NVR.
<davidlt> otherwise we not gonna see benefits for months, and it will be hard to track things too.
<rwmjones_> morning
<davidlt> I almost done looking at all PRs
<rwmjones_> david: I don't know, best to ask fuwei
<rwmjones_> I added comments about things I *don't* like in the PRs, that's all
<rwmjones_> was going to leave the good ones to be reviewed by the packagers
<davidlt> Could they squash all the suffix commits, etc?
<rwmjones_> I can bump NVRs (in Fedora) for any package in Fedora, just let me know which one - I mean by this, not make any changes other than bump release and rebuild
<rwmjones_> yes the suffix Release changes are all wrong, I think I said that on the ones I noticed
<davidlt> If we commit changes without NVR + Changelog entry the maintainer might sit on the change, and not submit a build.
<davidlt> We want to see a build with riscv change as soon as appropriate.
<davidlt> Ask for F40 backport soonish too.
<davidlt> Hmm.. :/
<davidlt> We need to be really careful reviewing stuff.
<rwmjones_> yup that one was completely wrong
<davidlt> Anyways, now I am done with review :)
<davidlt> They should also provide more information.
<davidlt> For example is tests are passed, links to Fedora/RISCV and OpenKoji builds.
<davidlt> For JDK11 I couldn't find commits that point to those changes too. Maybe it's just me.
<davidlt> I looked at a couple of branches.
<davidlt> playonlinux is also NAK as there is no Wine support on riscv64.
<rwmjones_> I asked fuwei if he can teach them git rebase -i
JasenChao has joined #fedora-riscv
JasenChao44 has joined #fedora-riscv
JasenChao44 has quit [Client Quit]
JasenChao has quit [Quit: Client closed]
davidlt has quit [Ping timeout: 255 seconds]
Guoguo has quit [Remote host closed the connection]
Guoguo has joined #fedora-riscv
Guoguo has quit [Remote host closed the connection]
davidlt has joined #fedora-riscv
Guoguo has joined #fedora-riscv
Guoguo has quit [Ping timeout: 260 seconds]
zsun has joined #fedora-riscv
<davidlt> is there a way to get more information?
<davidlt> I want to see actual git error
zsun has quit [Quit: Leaving.]
<sorear> would "wine support for riscv64" mean integrating it with fex/box64, wine support for IMAGE_FILE_MACHINE_RISCV64, or winelib support for __riscv?
Guoguo has joined #fedora-riscv
Guoguo has quit [Remote host closed the connection]
<davidlt> which is probably this: https://gitlab.winehq.org/wine/wine
<davidlt> I don't really use it (or tested) thus don't have details
<rwmjones_> the way I understand how wine works, I don't understand how it can just run on riscv64
<davidlt> rwmjones_, ah, this is similar to meson issue
<sorear> I have a weak understanding of how wine works but much less of how it's packaged
<sorear> if I have x86 or x86_64 Windows binaries and I want to run them on an aarch64 Fedora host, can I install the wine package and use that to drive the whole process or I do I need to separately install an emulator?
<sorear> there's no NT RV64 ABI as of yet but there's a use case for running x86 binaries, however that's done
<rwmjones_> sorear: I've not tried that, but typically it wouldn't work unless there's some emulation going on (like qemu-aarch64)
<rwmjones_> typically (on x86-64) the binaries run directly and link to the wine library and talk to the separate wine-server
<rwmjones_> with the wine-server making system calls on behalf of the windows binary
<rwmjones_> I think they talk over a unix domain socket
<davidlt> I think you can use win on aarch64 and run x86_64 games
<davidlt> and, IIRC, Game Porting Toolkit (from Apple) is using Wine too
Guoguo has joined #fedora-riscv
<rwmjones_> in summary, no fscking clue about mold, so I'm going to look at test timeouts next
<rwmjones_> it has a build system which makes autoconf look sane and reasonable
<rwmjones_> davidlt: I'm going to ask about pytest_addopts on the mailing list, let's get a definitive answer there
<davidlt> rwmjones_, ok
<davidlt> I didn't find documentation, but I have built way too many packages to notice this :)
<davidlt> (and I have full dist-git checkout to look for answers)
<davidlt> GCC 14 is writing out RPMs
<rwmjones_> oh I noticed that redhat-rpm-config adds an error about <something>-rwx-segments which isn't supported by clang
<rwmjones_> (sorry, don't have the exact details now as I downgraded redhat-rpm-config to get around it)
<rwmjones_> --error-rwx-segments
<davidlt> that was a month ago
<davidlt> it's only for GCC
<davidlt> i.e. this will not get passed on Clang
<davidlt> "%{toolchain}" == "gcc" ? "-specs=/usr/lib/rpm/redhat/redhat-hardened-ld-errors" : ""
<rwmjones_> ok maybe it was gcc not clang, sorry I forgot the exact details now
<rwmjones_> davidlt: any idea about https://src.fedoraproject.org/rpms/rocm-smi/pull-request/1 ? you may actually have an AMD GPU plugged into a hifive unmatched iirc?
<davidlt> rwmjones_, replied
<davidlt> This one should also be on the end of the list
<davidlt> Not the best time to test this kind of thing
zsun has joined #fedora-riscv
<rwmjones_> so apparently the environment variable PYTEST_ADDOPTS is preferred
zsun has quit [Quit: Leaving.]
fuwei has quit [Remote host closed the connection]
fuwei has joined #fedora-riscv
Guoguo has quit [Remote host closed the connection]
<rwmjones_> compiling cmake is painful ...
<davidlt> yes it is, it's for that macro?
<rwmjones_> yup
<rwmjones_> also the rpm macro syntax is nuts
<rwmjones_> the final version (I think) is going to be:
<rwmjones_> %ctest(:-:h:j:u:v:A:C:D:E:F:H:I:L:M:N:O:Q:R:S:T:U:V:) \
<rwmjones_> %__ctest --test-dir "%{__cmake_builddir}" \\\
<rwmjones_> --output-on-failure \\\
<rwmjones_> --force-new-ctest-process \\\
<rwmjones_> %ifarch riscv64 \
<rwmjones_> --timeout 6000 \\\
<rwmjones_> %endif \
<rwmjones_> %{?_smp_mflags} %{**}
<rwmjones_> notice the lack of consistent backslashes
<davidlt> yeah, looks good
<rwmjones_> still testing it ...
<rwmjones_> so the bad news with meson is that the RPM macro is actually stored upstream for some reason
<rwmjones_> (also what on earth does %shrink do ...)
<rwmjones_> "Add macro %{shrink:…} to trim out excess white-space"
<davidlt> hm...
Guoguo has joined #fedora-riscv
<rwmjones_> davidlt: so you can't just put %ifarch riscv64 into that meson_test macro because ...
<rwmjones_> $ rpm --eval '%{meson_test}'
<rwmjones_> /usr/bin/meson test -C redhat-linux-build --num-processes 32 --print-errorlogs %ifarch riscv64 --timeout-multiplier 10 %endif
<rwmjones_> duplicating the entire macro and using %ifarch .. %else .. %endif is possible but unappealing
<rwmjones_> actually I have an idea, define a local __meson_test_timeout variable
Guoguo has quit [Ping timeout: 272 seconds]
<davidlt> it now sounds that per package is easier to handle
<rwmjones_> well I don't know, I tried several variations and couldn't get rpm to work
<rwmjones_> I'm sure it's possible but no idea how
cyberpear has joined #fedora-riscv
Guoguo has joined #fedora-riscv
Guoguo has quit [Ping timeout: 264 seconds]
davidlt has quit [Ping timeout: 246 seconds]
Guoguo has joined #fedora-riscv
Guoguo has quit [Ping timeout: 256 seconds]
cyberpear has quit [Quit: Connection closed for inactivity]
Guoguo has joined #fedora-riscv
Guoguo has quit [Ping timeout: 246 seconds]