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
<tekkamanninja> rwmjones: Hi , that work , we rebuild krb5 with %undefine _package_note_flags, then built libtirpc successfully
<tekkamanninja> Great thanks
somlo has quit [Remote host closed the connection]
somlo has joined #fedora-riscv
MichaelZhu has joined #fedora-riscv
jcajka has joined #fedora-riscv
<tekkamanninja> rwmjones: we also need to rebuild perl for this issue
<tekkamanninja> this issue blocks a lot of package
<tekkamanninja> s/package/packages
<rwmjones> tekkamanninja: hi
<rwmjones> tekkamanninja: it's a huge pain
<davidlt[m]> Is this a generic issue? or riscv64 specific one?
<rwmjones> not sure I can do anything except commisserate, but it affects regular Fedora too
<rwmjones> generic
<davidlt[m]> So how is that being solved?
<rwmjones> by ignoring it afaict
<davidlt[m]> What?
<rwmjones> you can add your voice to mine on the mailing list
<rwmjones> I've added the %undefine to just about every package I own
<rwmjones> which is the best I can do
<davidlt[m]> Is there are wiki Change describing this change?
ahs3 has quit [Ping timeout: 250 seconds]
<rwmjones> the only practical thing to do is to raise the issue at FESCo, present evidence that it's a clusterf-, and get them to revert it
<rwmjones> we'd need to know things like how many packages in Fedora have wrong build paths embedded etc
<davidlt[m]> Yeah, I was thinking that this should be FESCo thing now
<rwmjones> I've suggested a better implementation of it too, but no one has had time to implement anything
<davidlt[m]> Well typically such large changes are being tested on the side tags, wasn't it?
<davidlt[m]> I mean I see statistics after mass rebuild, but not before it.
<rwmjones> not afaik, first we saw was it was added just before the mass rebuild
<rwmjones> basically, complaining about it to me isn't going to help, because I already agree with you :-)
<rwmjones> complain on the mailing list or to fesco
<davidlt[m]> The Change itself seems simple, just a note section with RPM NVR.
<rwmjones> it's the way it was implemented which is the problem
<rwmjones> it should have been done the same place we do debuginfo, stripping
<rwmjones> instead it added this huge hack that changes LDFLAGS
<rwmjones> loads of packages embed compile-time LDFLAGS in themselves
<rwmjones> so it broke everything
<davidlt[m]> I assume (a lot to read) it's done via LDFLAGS, and not in a post linking stage?
<rwmjones> currently yes
<davidlt[m]> Interesting. They seem to use linker scripts for this
<davidlt[m]> I don't recall details, but I recall doing something similar with just objcopy in a post stage many years ago.
<davidlt[m]> Is linker script used for buildid reasons or something?
<davidlt[m]> I don't have right now too much time to dive into this :/
ahs3 has joined #fedora-riscv
zsun has joined #fedora-riscv
masami has joined #fedora-riscv
MichaelZhu has quit [Remote host closed the connection]
MichaelZhu has joined #fedora-riscv
MichaelZhu has quit [Remote host closed the connection]
masami has quit [Quit: Leaving]
jcajka has quit [Quit: Leaving]
zsun has quit [Quit: Leaving.]
chuangzhu[m] has joined #fedora-riscv
chuangzh1 has joined #fedora-riscv
rwmjones is now known as rwmjones|HOLS
defolos has joined #fedora-riscv
<defolos> What is the current state of the vanilla Fedora image on the starfive VisionFive? Has anyone tested it?
MichaelZhu has joined #fedora-riscv
MichaelZhu has quit [Remote host closed the connection]