GNU bug report logs - #67921
[PATCH haskell-team 1/3] gnu: ghc-next: Update to version 9.4.8

Previous Next

Full log

Message #295 received at 67921 <at> debbugs.gnu.org (full text, mbox):

From: Nicolas Graves <ngraves <at> ngraves.fr>
To: Divya Ranjan <divya <at> subvertising.org>
Cc: dev <at> jpoiret.xyz, Lars-Dominik Braun <lars <at> 6xq.net>, saku <at> laesvuori.fi,
 67921 <at> debbugs.gnu.org
Subject: Re: [bug#67921] [PATCH v3 01/24] gnu: ghc: More robust build with
 binutils ≥ 2.39.
Date: Fri, 15 Nov 2024 09:38:20 +0100

On 2024-11-14 23:13, Divya Ranjan wrote:

>> The upside is more consistent and easier to understand code in Guix,
>> the downside is rebuild times. Only depends on how you weigh each
>> other.
>
> Any estimation for the ballpark within which which it might fall? Here
> do you mean rebuild times of each version? Like, this build time isn't
> going to affect user installation and creation of drvs, or would it?

I'm not sure I understand your question well.  Once these derivations
are built on the server and available as substitutes, users will replace
them by downloading their binaries.  But I'm not sure this is your
question.

Using 15 cores on x86-64, it takes basically one day (~10h) to rebuild
versions 8.6 to 9.4.

This first commit is not strictly necessary.  If you wich to avoid these
rebuild costs, you can instead just drop the patch, and use this option
passed to the linker on 9.4 only (necessary, there are similar build
failures otherwise).  This is actually a good alternative, since I know
other versions will build.  If you can fix 9.6, we can inject my first
patch only for server evaluation before merging.

-- 
Best regards,
Nicolas Graves

Previous Next

GNU bug tracking system
Copyright (C) 1999 Darren O. Benham, 1997,2003 nCipher Corporation Ltd, 1994-97 Ian Jackson.