| Message ID | cover.1707775899.git.steve@futurile.net |
|---|---|
| Headers | show |
| Series | Adding clojure.java-time | expand |
Hello, according to our naming convention https://guix.gnu.org/de/manual/devel/en/html_node/Package-Naming.html https://guix.gnu.org/de/manual/devel/en/html_node/Python-Modules.html I think the package should be called clojure-java-time (well, this is not spelt out precisely, but we could argue by analogy, and by consistency with the existing clojure packages). I know nothing about clojure, but am wondering about two things: - There is one other package with #:aot-exclude '(#:all), which gives an explanation why the line is there; could you add one here as well? - Would it make sense to package math.combinatorics first to enable the tests? If we do not test now, I am afraid we will forget it later... Andreas
Hi Steve, Did you have a chance to look into Andreas’ comments about this patch? https://issues.guix.gnu.org/69119 Ludo’. Andreas Enge <andreas@enge.fr> skribis: > Hello, > > according to our naming convention > https://guix.gnu.org/de/manual/devel/en/html_node/Package-Naming.html > https://guix.gnu.org/de/manual/devel/en/html_node/Python-Modules.html > I think the package should be called clojure-java-time (well, this is not > spelt out precisely, but we could argue by analogy, and by consistency with > the existing clojure packages). > > I know nothing about clojure, but am wondering about two things: > - There is one other package with #:aot-exclude '(#:all), which gives > an explanation why the line is there; could you add one here as well? > - Would it make sense to package math.combinatorics first to enable the > tests? If we do not test now, I am afraid we will forget it later... > > Andreas
Hi Ludo, I got blocked as this patch caused me to look at whether we should be compiling to byte-code by default. I have not been able to achieve consensus so far (and have taken some time off from the thread): https://lists.gnu.org/archive/html/guix-devel/2024-02/msg00241.html Having been unable to convince others I also asked the Clojure community: https://clojureverse.org/t/should-linux-distributions-ship-clojure-byte-compiled-aot-or-not/10595 Their perspective is that: a. We should not byte-code compile libraries (what Clojure calls AOT), but we could do so for tools/apps. b. We should not package libraries at all From both interactions I'm unsure if packaging this (or any other Clojure libs/tools) is the right move. Steve On 2 Mar, Ludovic Courtès wrote: > Hi Steve, > > Did you have a chance to look into Andreas’ comments about this patch? > > https://issues.guix.gnu.org/69119 > > Ludo’. > > Andreas Enge <andreas@enge.fr> skribis: > > > Hello, > > > > according to our naming convention > > https://guix.gnu.org/de/manual/devel/en/html_node/Package-Naming.html > > https://guix.gnu.org/de/manual/devel/en/html_node/Python-Modules.html > > I think the package should be called clojure-java-time (well, this is not > > spelt out precisely, but we could argue by analogy, and by consistency with > > the existing clojure packages). > > > > I know nothing about clojure, but am wondering about two things: > > - There is one other package with #:aot-exclude '(#:all), which gives > > an explanation why the line is there; could you add one here as well? > > - Would it make sense to package math.combinatorics first to enable the > > tests? If we do not test now, I am afraid we will forget it later... > > > > Andreas