On Fri, Apr 08, 2016 at 09:31:47AM +0300, Efraim Flashner wrote: > On Thu, Apr 07, 2016 at 11:27:01PM +0200, Ludovic Courtès wrote: > > Danny Milosavljevic <dan...@scratchpost.org> skribis: > > > > >>guix substitute: error: download from > > >>'http://hydra.gnu.org/nar/36wqhbch45x055wj6gfsng00zkwfqg6n-texlive-20150523-texmf.tar.xz' > > >> failed: 503, "Service Temporarily Unavailable" > > > > This is because we’ve decided to stop distributing texlive-texmf (4 GiB) > > from Hydra; it’s just too much, at least for now. > > > > You need to use --fallback (or --no-substitutes) to get it. > > > > See Leo’s and Efraim’s messages for the rest. :-) > > > > Ludo’. > > > > Do we want to tag it `#:build-locally? #t` or `#:substitutable? #f` > then?
I don't like the current situation, where building libreoffice (among others) requires --fallback unless texlive-texmf is already in the store. About #:local-build, the manual [0] says: "When local-build? is true, declare that the derivation is not a good candidate for offloading and should rather be built locally (see Daemon Offload Setup). This is the case for small derivations where the costs of data transfers would outweigh the benefits." I think it's orthogonal to substituting binaries. Some discussion: https://bugs.gnu.org/18747 The manual [0] says about #:substitutable?: "When substitutable? is false, declare that substitutes of the derivation’s output should not be used (see Substitutes). This is useful, for instance, when building packages that capture details of the host CPU instruction set." Some people are using their own servers and it could make sense for them to substitute texlive-texmf if they have fast I/O and networking. But it might not make a big difference to them in practice. [0] https://www.gnu.org/software/guix/manual/guix.html#Derivations
