On 2026-01-24 00:51:57 +0100, Sebastian Ramacher wrote:
> On 2026-01-23 22:48:23 +0100, Aurelien Jarno wrote:
> > Hi,
> > 
> > On 2026-01-23 02:16, Santiago Vila wrote:
> > > severity 1125944 grave
> > > affects 1125944 src:ableton-link src:density-fitness src:dump 
> > > src:dynarmic src:fenics-dolfinx src:fizmo-console src:fizmo-ncursesw 
> > > src:fuzzylite src:gcc-11-cross-ports src:gcc-12-cross-ports 
> > > src:gcc-13-cross-ports src:gnome-system-monitor src:libbgcode src:libmcfp 
> > > src:libodsstream src:libpdb-redo src:libsfml src:lomiri-thumbnailer 
> > > src:metview src:metview src:metview src:mrc src:netgen src:nuspell 
> > > src:oaknut src:openblas src:posixsignalmanager src:rdkit src:snapcast 
> > > src:sqlite-modern-cpp src:swiftlang src:termpaint src:therion src:threeb 
> > > src:tl-expected src:tl-optional src:tuiwidgets src:yuzu 
> > > src:appstream-generator src:augustus src:cli11 src:corectrl src:dssp 
> > > src:endless-sky src:fizmo-sdl2 src:gcc-14-cross-ports 
> > > src:gcc-15-cross-ports src:genomicsdb src:highs src:j4-dmenu-desktop 
> > > src:libcifpp src:libcsfml src:libpappsomspp src:nmodl src:papilo 
> > > src:purify src:rapidfuzz-cpp src:sdpa src:sol2 src:sopt src:spdlog 
> > > src:tortoize src:tuning-library
> > > thanks
> > > 
> > > I'm adding all the packages I found which currently FTBFS with this error:
> > > 
> > > input SFrame sections with different format versions prevent .sframe 
> > > generation
> > > 
> > 
> > Thanks for testing that. After a quick investigation, the problem is 
> > that these try to link with a library that is still in using SFrame v2. 
> > For instance dump FTBFS due to libcatch2-dev which has been built with 
> > gcc-14 14.3.0-6, which enabled SFrame generation.
> > 
> > There are ~650 sources packages that have a ELF binary in /usr/lib/* 
> > with SFrame v2 on amd64. That said the common denominator with the above 
> > failures, seems to be that they try to link against a static library 
> > with SFrame v2. Scanning the archive, I have found that the following 
> > source packages that have a binary package that ships a .a or a .o with 
> > SFrame v2 on amd64:
> > 
> > afflib_3.7.22-1
> > altos_1.9.22-2+b1
> > avahi_0.8-17
> > base16384_2.3.2-1
> > catch2_3.7.1-0.6
> > coinor-csdp_6.2.0-6
> > cpu_1.4.3-16
> > crossfire_1.75.0-9
> > dbus-test-runner_19.04.0-2
> > emoslib_2:4.5.9-12
> > funtools_1.4.8-2
> > gerbv_2.10.0-2
> > ginac_1.8.9-1
> > gmp-ecm_7.0.6+ds-2
> > haskell-bv-sized_1.0.5-2+b3
> > haskell-parameterized-utils_2.1.9.0-2
> > haskell-tasty-checklist_1.0.6.0-2+b3
> > haskell-what4_1.5.1-4+b4
> > ijs_0.35-16
> > irstlm_6.00.05-7
> > jpeg-xl_0.11.1-6
> > lapack_3.12.1-7
> > libdrilbo_0.2.11-2.2
> > libfizmo_0.7.15-2.2
> > libkal_0.9.0-6
> > libprintsys_0.6-14
> > libpyzy_1.0.1-9
> > librelp_1.12.0-1
> > libssh_0.11.3-1
> > libstb_0.0~git20250907.fede005+ds-1
> > libxt_1:1.2.1-1.3
> > litehtml_0.9-1
> > lmbench_3.0-a9+debian.1-10
> > lp-solve_5.5.2.14-4.1
> > lwipv6_1.5a-10.1
> > magics++_4.16.0+really4.16.0-2
> > meschach_1.2b-19
> > mojoshader_0.0~hg1314+git20250606.93d9d45+dfsg-1
> > openblas_0.3.30+ds-3
> > openzwave_1.6.1914+ds2-4
> > osra_2.1.3-4
> > persistent-cache-cpp_1.0.9-1
> > pgplot5_5.2.2-19.8
> > racket_8.18+dfsg1-1
> > rumur_2025.08.31-1
> > skycat_3.1.2+starlink1~b+dfsg-10
> > sqlite3_3.46.1-8
> > stalin_0.11-11
> > starlink-ast_9.2.10+dfsg-3
> > tclx8.4_8.4.1-6
> > triangle_1.6-7
> > wcstools_3.9.7-2
> > xnee_3.19-9.4
> > 
> > I guess binNMUing them should fix the issue.
> 
> Scheduled. ghc is currently undergoing a transition, so haskell-* will
> see uploads and rebuilds anyway.

Santiago, could you retest the packages that failed? I tested a small
sample (dynarmic, oaknut, sopt, sqlite-modern-cpp) of the packages
marked as affected by this bug and they now build again after the round
of binNMUs.

Cheers
-- 
Sebastian Ramacher

Reply via email to