[bug#56604] closed (Re: bug#56604: [PATCH 0/8] Update Clojure to 1.11.1.)

[Roman Scherer]

Hi Maxim and Ludo,

thanks for your replies. Sorry that I was not answering earlier. The heat killed my laptop and I am having trouble to get set up again.

I follow up when I have my new laptop.

Thanks

On Sat, Jul 23, 2022, 00:12 GNU bug Tracking System <help-debbugs@gnu.org> wrote:

Your bug report

#56604: [PATCH 0/8] Update Clojure to 1.11.1.

which was filed against the guix-patches package, has been closed.

The explanation is attached below, along with your original report.
If you require more details, please reply to 56604@debbugs.gnu.org.

--
56604: https://debbugs.gnu.org/cgi/bugreport.cgi?bug=56604
GNU Bug Tracking System
Contact help-debbugs@gnu.org with problems



---------- Forwarded message ----------
From: "Ludovic Courtès" <ludo@gnu.org>
To: r0man <roman@burningswell.com>
Cc: 56604-done@debbugs.gnu.org
Bcc: 
Date: Sat, 23 Jul 2022 00:11:05 +0200
Subject: Re: bug#56604: [PATCH 0/8] Update Clojure to 1.11.1.
Hi,

r0man <roman@burningswell.com> skribis:

> This phase makes sure the timestamp of compiled class files is set to a later
> point in time than the timestamp of the corresponding Clojure source files. If
> the timestamps of the class and source files are the same, the Clojure
> compiler will compile the sources again which can lead to issues. This problem
> has been discussed here [1]. The suggested solution was to keep/adjust the
> timestamps of the class files.

Sounds reasonable.  It’s a bummer though that the whole phase is pasted
from ant-build-system.scm, the only difference being the timestamps
(1980 instead of 1970).

I added a TODO comment in clojure-build-system.scm when applying the
patch.  Could you follow up with a patch to factorize that?

> Btw, I was a bit surprised that in Guix Clojure packages are AOT compiled. The
> general wisdom in the Clojure community seems to be to avoid AOT compilation
> when distributing libraries, and only AOT compiling Uberjars for final
> deployment. Due to issues like I mentioned in clojure-instaparse.
>
> Are we sure that AOT compiling all Clojure source files by default is a good
> idea, instead of just compiling user declared namespaces which Leiningen and
> friends are doing? WDYT?

Not much, but as you might have seen in ./etc/teams.scm, the project is
finally being structured as teams.  There’s an opportunity for you to
start a Clojure team and to take the lead!  :-)

As a first step, I’d recommend getting in touch with people who have
worked on ‘clojure-build-system’ and packaged things in the past.

>   gnu: clojure-tools-cli: Update to 1.0.206.
>   gnu: clojure-tools-gitlibs: Update to 2.4.181.
>   gnu: clojure-tools-deps-alpha: Update to 0.14.1212.
>   gnu: clojure-tools: Update to 1.11.1.1149.
>   gnu: clojure: Update to 1.11.1.
>   gnu: clojure-algo-generic: Fix test failing under AOT in Clojure 1.11.1.
>   gnu: clojure-core-match: Update to 1.0.0.
>   gnu: clojure-instaparse: Update to 1.4.12 (disabled AOT).

I adjusted all the commit logs to follow our conventions; please
consider doing this next time:

  https://guix.gnu.org/manual/devel/en/html_node/Submitting-Patches.html

The instaparse patch missed the hash update so I did that too.

Thanks!

Ludo’.




---------- Forwarded message ----------
From: r0man <roman@burningswell.com>
To: guix-patches@gnu.org
Cc: 
Bcc: 
Date: Sat, 16 Jul 2022 20:17:34 +0200
Subject: [PATCH 0/8] Update Clojure to 1.11.1.
Hello Guix,

This patch series updates Clojure and it's packages to 1.11.1.

It also adds the 'reset-class-timestamps phase to the Clojure build system.

This phase makes sure the timestamp of compiled class files is set to a later
point in time than the timestamp of the corresponding Clojure source files. If
the timestamps of the class and source files are the same, the Clojure
compiler will compile the sources again which can lead to issues. This problem
has been discussed here [1]. The suggested solution was to keep/adjust the
timestamps of the class files.

[1] https://www.mail-archive.com/clojure@googlegroups.com/msg99928.html

Could you please review this?

In the patch for clojure-instaparse, I disabled the AOT compilation because of
a known issue. Is that ok?

Btw, I was a bit surprised that in Guix Clojure packages are AOT compiled. The
general wisdom in the Clojure community seems to be to avoid AOT compilation
when distributing libraries, and only AOT compiling Uberjars for final
deployment. Due to issues like I mentioned in clojure-instaparse.

Are we sure that AOT compiling all Clojure source files by default is a good
idea, instead of just compiling user declared namespaces which Leiningen and
friends are doing? WDYT?

Thanks Roman.

r0man (8):
  gnu: clojure-tools-cli: Update to 1.0.206.
  gnu: clojure-tools-gitlibs: Update to 2.4.181.
  gnu: clojure-tools-deps-alpha: Update to 0.14.1212.
  gnu: clojure-tools: Update to 1.11.1.1149.
  gnu: clojure: Update to 1.11.1.
  gnu: clojure-algo-generic: Fix test failing under AOT in Clojure 1.11.1.
  gnu: clojure-core-match: Update to 1.0.0.
  gnu: clojure-instaparse: Update to 1.4.12 (disabled AOT).

 gnu/packages/clojure.scm            | 185 +++++++++++++++++++---------
 guix/build/clojure-build-system.scm |  53 +++++++-
 2 files changed, 178 insertions(+), 60 deletions(-)

--
2.36.1