ludo@gnu.org (Ludovic Courtès) writes: > Kei Kebreau skribis: > >> ludo@gnu.org (Ludovic Courtès) writes: >> >>> Kei Kebreau skribis: >>> >>>> ludo@gnu.org (Ludovic Courtès) writes: >>>> >>>>> Hi Kei, >>>>> >>>>> The subject is truncated. :-) Caused by which commit? >>>>> >>>>> Kei Kebreau skribis: >>>>> >>>>>> * gnu/packages/bioinformatics.scm (seek)[arguments]: Modify unpack phase. >>>>>> * gnu/packages/ci.scm (cuirass)[arguments]: Add patch-/bin/sh phase. >>>>> >>>>> This is weird, normally the ‘patch-source-shebangs’ phase, which happens >>>>> right before ‘configure’, should take care of these, no? >>>>> >>>> >>>> Normally the 'patch-source-shebangs' does just that, but in this case >>>> the script needs to be patched because the bootstrapping phase depends >>>> on perl. >>> >>> OK. I don’t understand at what point this became necessary though. >>> >> >> It's probably a result of moving the 'bootstrap' phase before the >> 'patch-source-shebangs' phase. Without this change, the seek package >> builds fine. I have no preference either way, though I notice that the >> same thing is done for guile-for-guile-emacs. > > Since this commit was just about changing the syntax, not the semantics, > I would expect it to keep phases in the same order. That would be more > concise than adding custom code to patch things here and there, no? > > WDYT? > I agree in the case of seek. >>>> From 87ea8addad973568d8599fb93c8df9766b96b6a1 Mon Sep 17 00:00:00 2001 >>>> From: Kei Kebreau >>>> Date: Sat, 29 Jul 2017 11:28:42 -0400 >>>> Subject: [PATCH] gnu: Use 'modify-phases' syntax. >>>> To: guix-patches@gnu.org >>> >>> Wrong patch. :-) >>> >>> Ludo’. >> >> Fixed! >> >> From d9eee184cf80252092c34864d492527ee768a7fb Mon Sep 17 00:00:00 2001 >> From: Kei Kebreau >> Date: Thu, 7 Sep 2017 09:36:28 -0400 >> Subject: [PATCH] gnu: Fix build failures caused by commit >> d10092b849153dc27cfed0a9601fde6c7bdec918. >> >> * gnu/packages/bioinformatics.scm (seek)[arguments]: Modify unpack phase. >> * gnu/packages/ci.scm (cuirass)[arguments]: Add patch-/bin/sh phase. > > That’s fine with me, even though I have a slight preference for simply > restoring the order of phases as it was before. > I'm most comfortable reverting seek to its last successful state without the substitute* call and patch leave the cuirass patch as is. If that's okay with you I can push this to master. > Thanks! > > Ludo’. Thanks for helping to sort this out!