Overview
Request 1201876 accepted
- Update to version 1.06:
* [ARM] Added support for inferring the SoC from the uarch
* [ARM] Added support for inferring the SoC from PCI devices
* [ARM] Added support for SVE detection
* [X86] Significantly improved support for old x86 CPUs
* [X86] Added support for Hygon CPUs
* [ARM and X86] Added support for measuring max CPU frequency
* [ARM and X86] Added `--measure-max-freq` option to manually
trigger CPU max frequency measurement
* [All] Added support for NO_COLOR
Request History
RN created request
- Update to version 1.06:
* [ARM] Added support for inferring the SoC from the uarch
* [ARM] Added support for inferring the SoC from PCI devices
* [ARM] Added support for SVE detection
* [X86] Significantly improved support for old x86 CPUs
* [X86] Added support for Hygon CPUs
* [ARM and X86] Added support for measuring max CPU frequency
* [ARM and X86] Added `--measure-max-freq` option to manually
trigger CPU max frequency measurement
* [All] Added support for NO_COLOR
jubalh accepted request
Accepting due to long time no response from maintainer
Not sure what would be better to leave newer package broken or ignore this
--debugresults in%checksection for ARM architectures.sr#1201873#diff_1_n33
why do you add another whitespace at ExclusiveArch?
That's from
osc vcautomation, not me.osc vcdoesnt manipulate the spec file. I'm talking about25+ExclusiveArch: %{arm} aarch64 x86_64 ppc ppc64 ppc64leIt is.
+# Copyright (c) 2024 SUSE LLCand+License: GPL-2.0-onlyThis lines also changed from automation, not by me.This is not the line "I'm talking about". Please read the message above again.
I just showed an example that it manipulate the .spec file indeed.
Okay, my bad. It's because of
osc ci, not because ofosc vc.After "Running source_service 'format_spec_file' ..."
But change still comes from automation, not from me.
https://paste.opensuse.org/pastes/355f6e3fc6ce
ah i see now that its actually on the correct line then. all good.
Actually my question was - is it acceptable to ignore
%checksection--debugresults to for rpm build process with some architectures... so is it?depends on the maintainer and the test
Here is
%checksection with line inside"%{buildroot}%{_bindir}/%{name} --debug"
I mentioned it in sr#1201873#diff_1_n33, as you can see issue of cpufetch --debug results --
*** buffer overflow detected ***
So, I'm not sure is it acceptable or not to ignore this error for ARM devices
lets wait for @gnuykeat to review. if he doesnt have time for several days then i'll take a look. but he is official maintainer so i would like to wait for him first.
Sure, thank you.
@jubalh
Sorry, I just wanted to ask... should we fix build process for ARM architectures?
Because only issue for them it's buffer overflow detection on
%checksection.(As I asked wanted in sr#1201873)
We should fix it if we know how to. But not by just removing the check section (I think). My hope is that ARM maintainers will see the build failure and send a SR to fix it correctly. Another option might be to report it upstream so they can fix their test for ARM.
Okay, thank you.
@gnuykeat: review reminder