Bug#1133139: erfa: FTBFS on armhf: t_erfa_c validation failed!

From	Aurelien Jarno <aurel32@debian.org>
Newsgroups	linux.debian.bugs.dist, linux.debian.ports.arm
Subject	Bug#1133139: erfa: FTBFS on armhf: t_erfa_c validation failed!
Date	2026-05-11 00:20 +0200
Message-ID	<MTkFb-4ctt-1@gated-at.bofh.it> (permalink)
References	<MIh7X-e3JN-1@gated-at.bofh.it> <MTkFb-4ctt-3@gated-at.bofh.it> <MIh7X-e3JN-1@gated-at.bofh.it> <MTkFb-4ctt-3@gated-at.bofh.it>
Organization	linux.* mail to news gateway

Cross-posted to 2 groups.

Show all headers | View raw

control: reassign -1 src:glibc
control: retitle -1 libc6/armhf: ld.so clobbers VFP registers during runtime linking
control: affects -1 erfa

Hi,

On 2026-05-01 13:21, Ole Streicher wrote:
> I now uploaded two new releases of the package, without success:
> 
> * 2.0.1-2 disabling optimization of armhf
> * 2.0.1-3 using gcc-14 instead of the actual one (gcc-15)
> 
> In both cases, the failure persists. I am out of ideas now. Disabling the
> architecture would remove a large portion of astrophysics packages for armhf
> (Raspberries, often used in smaller telescope setups); just disabling the
> tests and ignoring the errors may lead to wrong results on these platforms.
> 
> Any idea is appreciated.

This is actually an issue in the dynamic linker (ld.so) affecting armhf 
since GCC 15 is used to build glibc. I have sent a patch upstream:
https://inbox.sourceware.org/libc-alpha/20260510220041.169362-1-aurelien@aurel32.net

I am therefore reassigning the bug to the glibc package. On the erfa 
side, the workarounds can be removed, it'll build fine once a fixed glibc 
reaches the archive.

Regards
Aurelien

-- 
Aurelien Jarno                          GPG: 4096R/1DDD8C9B
aurelien@aurel32.net                     http://aurel32.net

Back to linux.debian.bugs.dist | Previous | Next | Find similar

Thread

Bug#1133139: erfa: FTBFS on armhf: t_erfa_c validation failed! Aurelien Jarno <aurel32@debian.org> - 2026-05-11 00:20 +0200

csiph-web