Bug#973480: transition: llvm-defaults

2020-12-07 Thread Sylvestre Ledru

Hello,

Le 04/12/2020 à 18:44, Sebastian Ramacher a écrit :

Control: tags -1 + confirmed

On 2020-10-31 14:35:36 +0100, Sylvestre Ledru wrote:

Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: transition

Hello,

One more time, it is time to have upgrade llvm.
I would like to ship bullseye with 11 and skip -10 (it had some performance 
regression in the
binary generated).

I am building a version with the ppc64el fix.
llvm-defaults has been pointing to 11 in experimental for quite sometime.
Happy to upload it in unstable when you give me the go.

Rust already relies on -11 in experimental.


Please go ahead.




Thanks!
Gianfranco uploaded it in unstable!

Thanks
S



Bug#973480: transition: llvm-defaults

2020-12-04 Thread Sebastian Ramacher
Control: tags -1 + confirmed

On 2020-10-31 14:35:36 +0100, Sylvestre Ledru wrote:
> Package: release.debian.org
> Severity: normal
> User: release.debian@packages.debian.org
> Usertags: transition
> 
> Hello,
> 
> One more time, it is time to have upgrade llvm.
> I would like to ship bullseye with 11 and skip -10 (it had some performance 
> regression in the
> binary generated).
> 
> I am building a version with the ppc64el fix.
> llvm-defaults has been pointing to 11 in experimental for quite sometime.
> Happy to upload it in unstable when you give me the go.
> 
> Rust already relies on -11 in experimental.

Please go ahead.

Cheers
-- 
Sebastian Ramacher


signature.asc
Description: PGP signature


Bug#973480: transition: llvm-defaults

2020-12-01 Thread Adrian Bunk
On Wed, Nov 11, 2020 at 04:36:06PM +0200, Adrian Bunk wrote:
> On Sat, Oct 31, 2020 at 03:18:52PM +0100, Sebastian Ramacher wrote:
> >...
> > I'm not sure what .depends ~ "~" is supposed to catch here and it causes
> > a parser failure. So I've merged this ben file with the one from the
> > llvm-defaults 9 transition. Once the tracker is available, please check
> > if that looks sane.
> 
> Please remove the second part from the Affected, these packages are not 
> part of the llvm-defaults transition.

I've rebuilt the 13 bad packages from the ben file below:
ccls OK
clazy OK
codelite OK
faust OK
ghdl FTBFS #976233
ldc OK
libclang-perl OK
qttools-opensource-src OK
sparse OK
ycmd OK
kdevelop OK
qtcreator OK
gnome-builder OK

This is not a normal transition where old package(s) disappear from 
testing since both LLVM 9 and LLVM 11 are already in testing and will 
stay in testing after this transition.

This means rebuilds could be scheduled at any time after an 
llvm-defaults upload without actual conflict with other transitions.

Removal of LLVM 9 and/or LLVM 10 from bullseye are topics unrelated
to this llvm-defaults transition.

cu
Adrian


title = "llvm-defaults 11";
is_affected = .build-depends ~ /\b(libclang|liblldb|llvm)-dev\b/;
is_good = .depends ~ /\b(lib|)(clang|lldb|llvm)1?-?11\b/;
is_bad = .depends ~ /\b(lib|)(clang|lldb|llvm)1?-?9\b/;



Bug#973480: transition: llvm-defaults

2020-11-11 Thread Adrian Bunk
On Sat, Oct 31, 2020 at 03:18:52PM +0100, Sebastian Ramacher wrote:
>...
> I'm not sure what .depends ~ "~" is supposed to catch here and it causes
> a parser failure. So I've merged this ben file with the one from the
> llvm-defaults 9 transition. Once the tracker is available, please check
> if that looks sane.

Please remove the second part from the Affected, these packages are not 
part of the llvm-defaults transition.

> Cheers

Thanks
Adrian



Bug#973480: transition: llvm-defaults

2020-10-31 Thread Sebastian Ramacher
Control: tags -1 + moreinfo
Control: block -1 by 971415

On 2020-10-31 14:35:36 +0100, Sylvestre Ledru wrote:
> Package: release.debian.org
> Severity: normal
> User: release.debian@packages.debian.org
> Usertags: transition
> 
> Hello,
> 
> One more time, it is time to have upgrade llvm.
> I would like to ship bullseye with 11 and skip -10 (it had some performance 
> regression in the
> binary generated).
> 
> I am building a version with the ppc64el fix.
>
> llvm-defaults has been pointing to 11 in experimental for quite sometime.
> Happy to upload it in unstable when you give me the go.

This transition is at least blocked by the ocaml transition as that
needs to finish before llvm-toolchain-11 can migrate.

> Rust already relies on -11 in experimental.

What about the other reverse dependencies? Do they all build fine with
llvm-toolchain-11?

> 
> Thanks
> Sylvestre
> 
> Ben file:
> 
> title = "llvm-defaults";
> is_affected = .depends ~ "~" | .depends ~ /(clang|llvm)1?-(?[3456789]|?10)/ | 
> .depends ~ "~" | .depends ~ /(clang|llvm)1?-?[11]/;
> is_good = .depends ~ "~" | .depends ~ /(clang|llvm)1?-?[11]/;
> is_bad = .depends ~ "~" | .depends ~ /(clang|llvm)1?-(?[3456789]|?10)/;

I'm not sure what .depends ~ "~" is supposed to catch here and it causes
a parser failure. So I've merged this ben file with the one from the
llvm-defaults 9 transition. Once the tracker is available, please check
if that looks sane.

Cheers

> 
> 
> (not certain about the is_bad regexp :/)
> 
> 
> -- System Information:
> Debian Release: bullseye/sid
>   APT prefers testing
>   APT policy: (900, 'testing'), (600, 'unstable'), (500, 'oldstable'), (300, 
> 'experimental')
> Architecture: amd64 (x86_64)
> Foreign Architectures: i386
> 
> Kernel: Linux 5.4.0-4-amd64 (SMP w/28 CPU threads)
> Kernel taint flags: TAINT_PROPRIETARY_MODULE, TAINT_OOT_MODULE, 
> TAINT_UNSIGNED_MODULE
> Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8), LANGUAGE not 
> set
> Shell: /bin/sh linked to /bin/dash
> Init: systemd (via /run/systemd/system)
> 

-- 
Sebastian Ramacher


signature.asc
Description: PGP signature


Bug#973480: transition: llvm-defaults

2020-10-31 Thread Sylvestre Ledru
Package: release.debian.org
Severity: normal
User: release.debian@packages.debian.org
Usertags: transition

Hello,

One more time, it is time to have upgrade llvm.
I would like to ship bullseye with 11 and skip -10 (it had some performance 
regression in the
binary generated).

I am building a version with the ppc64el fix.
llvm-defaults has been pointing to 11 in experimental for quite sometime.
Happy to upload it in unstable when you give me the go.

Rust already relies on -11 in experimental.

Thanks
Sylvestre

Ben file:

title = "llvm-defaults";
is_affected = .depends ~ "~" | .depends ~ /(clang|llvm)1?-(?[3456789]|?10)/ | 
.depends ~ "~" | .depends ~ /(clang|llvm)1?-?[11]/;
is_good = .depends ~ "~" | .depends ~ /(clang|llvm)1?-?[11]/;
is_bad = .depends ~ "~" | .depends ~ /(clang|llvm)1?-(?[3456789]|?10)/;


(not certain about the is_bad regexp :/)


-- System Information:
Debian Release: bullseye/sid
  APT prefers testing
  APT policy: (900, 'testing'), (600, 'unstable'), (500, 'oldstable'), (300, 
'experimental')
Architecture: amd64 (x86_64)
Foreign Architectures: i386

Kernel: Linux 5.4.0-4-amd64 (SMP w/28 CPU threads)
Kernel taint flags: TAINT_PROPRIETARY_MODULE, TAINT_OOT_MODULE, 
TAINT_UNSIGNED_MODULE
Locale: LANG=fr_FR.UTF-8, LC_CTYPE=fr_FR.UTF-8 (charmap=UTF-8), LANGUAGE not set
Shell: /bin/sh linked to /bin/dash
Init: systemd (via /run/systemd/system)