Marius Bakke <address@hidden> writes: > Alex ter Weele <address@hidden> writes: > >> Marius Bakke <address@hidden> writes: >> >>> Alex ter Weele <address@hidden> writes: >>> >>>> From 2774cc719fa132ed8fd34a1d71f2c5a0f8697645 Mon Sep 17 00:00:00 2001 >>>> From: Alex ter Weele <address@hidden> >>>> Date: Fri, 20 Jul 2018 21:35:14 -0500 >>>> Subject: [PATCH 1/2] gnu: agda: Compile .agda files. >>>> >>>> * gnu/packages/agda.scm: (agda)[arguments]: Compile .agda files. >>> >>> [...] >>> >>>> + (arguments >>>> + `(#:modules ((guix build haskell-build-system) >>>> + (guix build utils) >>>> + (srfi srfi-26)) >>>> + #:phases >>>> + (modify-phases %standard-phases >>>> + (add-after 'compile 'agda-compile >>>> + (lambda* (#:key outputs #:allow-other-keys) >>>> + (let* ((out (assoc-ref outputs "out")) >>>> + (agda-compiler (string-append out "/bin/agda")) >>>> + (agda-files-path (string-append >>>> + out >>>> + >>>> "/share/x86_64-linux-ghc-8.0.2/Agda-" >>> >>> This is unfortunate, since it hard-codes both architecture and GHC >>> version. Can you think of another method to find these files? >>> >>> Are there ".agda" files elsewhere under "/share"? Could the find-files >>> invokation below simply search "out/share", for example? >>> >> >> That works! >> >>>> + ,version >>>> + "/lib/prim/Agda/")) >>>> + (agda-files (append >>>> + (find-files agda-files-path "\\.agda$") >>>> + (find-files (string-append >>>> + agda-files-path >>>> + "Builtin") >>>> + "\\.agda$")))) >>> >>> (find-files ...) recurses through subdirectories, are you sure adding an >>> extra pass for "Builtin/*.agda" makes a difference here? >>> >> >> Removed. >> >>>> + (for-each (cut invoke agda-compiler <>) agda-files) >>>> + #t)))))) >>>> (home-page "http://wiki.portal.chalmers.se/agda/";) >>> >>> Otherwise LGTM. >>> >>> [...] >>> >>>> From ff441cb0b500d247de9a54c3212f80b4fcfaf04f Mon Sep 17 00:00:00 2001 >>>> From: Alex ter Weele <address@hidden> >>>> Date: Sat, 21 Jul 2018 10:57:35 -0500 >>>> Subject: [PATCH 2/2] gnu: Add emacs-agda2-mode >>>> >>>> * gnu/packages/agda.scm (emacs-agda2-mode): New variable. >>> >>> LGTM! >> >> I found a few potential issues with it. The variable and package names >> were mismatched, the inputs were inherited, and the home page was >> misspelled. Corrected in the following patch series. >> >> Thanks for the review! > > Thanks for catching my review mistake! All LGTM. Whoops, Ludovic helpfully reminded me on IRC that you don't yet have commit access. I've pushed these patches. (note that I added a #t after the chdir for emacs-agda2-mode, since the return value of chdir is unspecified)
signature.asc
Description: PGP signature
--- End Message ---