Dne 11. 02. 23 v 18:53 Chris Adams napsal(a):
Once upon a time, Kevin Kofler via devel
<devel(a)lists.fedoraproject.org> said:
> Ben Cotton wrote:
>> 1–3. distribution — {Workstation,Everything,Server} boot x86_64 image
>> exceeds maximum size — ASSIGNED
> Let me guess, this is from the -fno-omit-frame-pointers Change? I have been
> asking for data on the size impact and I have been completely ignored both
> by the Change owners and by FESCo.
Kevin, if you are serious about this, then you probably want to analyze
the size changes from the "Fedora rawhide compose report: 20230125.n.0
changes" email, which is the compose after mass rebuild.
Looking randomly on e.g. Coin packages, their size increased considerably:
~~~
Package: Coin3-3.1.3-33.fc38
Old package: Coin3-3.1.3-32.fc37
Summary: High-level 3D visualization library
RPMs: Coin3 Coin3-devel
Size: 47.99 MiB
Size change: 4.03 MiB
Changelog:
* Wed Jan 18 2023 Fedora Release Engineering
<releng(a)fedoraproject.org> - 3.1.3-33
- Rebuilt for
https://fedoraproject.org/wiki/Fedora_38_Mass_Rebuild
~~~
There are also packages which shrank their size:
~~~
Package: bullet-3.08-6.fc38
Old package: bullet-3.08-5.fc38
Summary: 3D Collision Detection and Rigid Body Dynamics Library
RPMs: bullet bullet-devel bullet-devel-doc bullet-extras
bullet-extras-devel
Size: 764.79 MiB
Size change: -3.96 MiB
Changelog:
* Wed Jan 18 2023 Fedora Release Engineering
<releng(a)fedoraproject.org> - 3.08-6
- Rebuilt for
https://fedoraproject.org/wiki/Fedora_38_Mass_Rebuild
~~~
although the saving looks more decent. Generally, it seems that after
mass rebuild, our package set is bigger by ~0,36%:
~~~
Size of upgraded packages: 147.51 GiB
Size change of upgraded packages: 542.39 MiB
~~~
But hard to tell what is the reason. If the GCC 13 or frame pointers or
something else. I for one would be interested in such analysis.
Vít
>
> If frame pointers blow up the distribution so much, I guess the best course
> of action would be to revert the Change and do another mass rebuild.
Can you stop grinding your axe against a decision you don't agree with?
You're just "guessing" with zero evidence.
If you look at the bugs, they date back well before the mass rebuild, so
there's no justification to guess it's the frame pointer change.