Am Montag, den 25.05.2015, 18:00 +0200 schrieb Mattia Rizzolo:
> > Why does it build ghc-dbgsym at all? I do not define that in
> > debian/rules. And
> > https://wiki.debian.org/ReproducibleBuilds/ExperimentalToolchain does
> > not mention a change in that direction either.
> We use the patched version of debhelper that builds .ddeb packages. That for
> either test debhelper at generating them and (TTBMK) helped understanding some
> differences (given that it contains some informations that would have been
> stripped otherwise).


> > Also, is there any way to find out for what binary this debug file is?
> > Or would it be visible if readelf would not fail on these files?
> you can use the build-id.
> The differences in that file comes from
> ./usr/lib/ghc/unix_G4Yo1pNtYrk8nCq1cx8P9d/libHSunix-

it seems that the GHC build builds some auxiliary binaries (e.g. unlit,
hp2hs) with the bootstrapping compiler, so for a chance of a fully
reproducible build, the ghc package used needs to be fixed as well. That
should be the case with ghc-7.8.4-7, uploaded to unstable today.

It does not explain the differences in build-id of the the various
lib...so files. Maybe the path name of the temporary file is taken into
account as well. But shouldn’t that show up in the ghc-dbgsym package?


Joachim "nomeata" Breitner
Debian Developer
  nome...@debian.org | ICQ# 74513189 | GPG-Keyid: F0FBF51F
  JID: nome...@joachim-breitner.de | http://people.debian.org/~nomeata

Attachment: signature.asc
Description: This is a digitally signed message part

Reproducible-builds mailing list

Reply via email to