bug#36747: Official MesCC bootstrap binaries differ from my locally built ones

[Ludovic Courtès]

Hello,

Mark H Weaver <address@hidden> skribis:

> Ludovic Courtès <address@hidden> writes:
>
>> Mark H Weaver <address@hidden> skribis:
>>
>>> You can reproduce them with the following command from a git checkout at
>>> commit 9e6256ba0f32ab12d61c914a3fed879dac881762, which is the tip of the
>>> 'wip-binaries' branch, based on recent 'master':
>>>
>>>   ./pre-inst-env guix build --system=i686-linux bootstrap-tarballs
>>>
>>> mhw@jojen ~/guix-wip-binaries$ git describe
>>> v1.0.1-2479-g9e6256ba0f
>>> mhw@jojen ~/guix-wip-binaries$ ./pre-inst-env guix build 
>>> --system=i686-linux bootstrap-tarballs
>>> /gnu/store/rdwyr8mh7dvhfkb5g4cws6q40hp23rbi-bootstrap-tarballs-0
>>> mhw@jojen ~/guix-wip-binaries$ cd 
>>> /gnu/store/rdwyr8mh7dvhfkb5g4cws6q40hp23rbi-bootstrap-tarballs-0
>>> mhw@jojen /gnu/store/rdwyr8mh7dvhfkb5g4cws6q40hp23rbi-bootstrap-tarballs-0$ 
>>> sha256sum *
>>> 3e50c070a100b6bcf84c4bf5c868f9cd0a9fd1570f5d82fbfb78f8411959091b  
>>> guile-static-stripped-2.2.4-i686-linux.tar.xz
>>> 1acd8f83e27d2fac311a5ca78e9bf11a9a1638b82469870d5c854c4e7afaa26a  
>>> linux-libre-headers-stripped-4.14.67-i686-linux.tar.xz
>>> 021543d9bb6af55f39e68d69692e3cb74646ced2cad0bb9ac0047ef81e9d7330  
>>> mescc-tools-static-stripped-0.5.2-i686-linux.tar.xz
>>> fb32090071b39fc804fb9a7fba96f0bc5eb844a0efd268fb24c42e6bfa959de0  
>>> mes-minimal-stripped-0.19-i686-linux.tar.xz
>>> c80cdd17b0a24eebdd75570ff72c4ec06e129bd702ac008186b57f6301c448e7  
>>> static-binaries-0-i686-linux.tar.xz
>>>
>>> Do you get the same results?
>>
>> I built it on bayfront and got the exact same result:
>
> I'm glad to hear it!  Marius and Janneke got these exact results as
> well, albeit at a different commit which should be equivalent.

Great.

>> (I’m actually surprised Guile is 100% reproducible; ISTR there were
>> still issues with gensyms in .go files, but maybe I’m wrong.)
>
> It's apparently not reproducible unless built sequentially.  I did some
> preliminary investigation, and my guess is that the gensyms produced by
> Guile's compiler sometimes depends on whether dependent modules were
> previously compiled or not.

I see.

>> So, green light, right?
>
> Yes!
>
>>> To match what 'core-updates-next' expects, it would be good to upload
>>> the files above, along with digital signatures, to the following new
>>> directory:
>>>
>>>   https://alpha.gnu.org/gnu/guix/bootstrap/i686-linux/20190815/
>>
>> Unless you disagree, I’ll upload these to ftp.gnu.org (not
>> alpha.gnu.org, but same directory as you wrote) tomorrow.
>
> That's fine, although I guess %bootstrap-base-urls will need to be
> extended to include the new base URL.

Uploaded to
<https://ftp.gnu.org/gnu/guix/bootstrap/i686-linux/20190815/>.  I
adjusted ‘%bootstrap-base-urls’ and the branch is now being evaluated
for real:

  https://ci.guix.gnu.org/jobset/core-updates-next

Thanks everyone for verifying and “fixing” these bootstrap binaries, and
apologies for delaying the process!

Ludo’.