<div dir="ltr">FYI on the perf regression fix for CLJ-2484, that patch is self-contained and should apply cleanly to 1.10.0 if you wanted to do so. As of right now, I do not expect to make any further changes wrt that issue in 1.10.1.<div><br></div><div>I'm hopeful that we will be able to release 1.10.1 without many more updates (I do have one more queued up, but we will probably go RC after that). I think 2-4 week expectation is the right ballpark timing for 1.10.1 if that is useful in making decisions.</div><div><br></div><div>Thanks for your work as always!</div><div>Alex</div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Fri, Apr 19, 2019 at 9:26 AM Elana Hashman <<a href="mailto:ehashman@debian.org">ehashman@debian.org</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex">Greetings programs!!<br>
<br>
Clojure is looking good for the buster release. We will be releasing<br>
with:<br>
<br>
- Clojure 1.10.0<br>
- Leiningen 2.9.0<br>
- Various supporting libraries and dependencies<br>
<br>
Additionally, I have successfully gotten clojure1.8 removed from<br>
testing, so we won't have to maintain multiple versions of Clojure in<br>
buster. Huzzah!<br>
<br>
What's next, if anything?<br>
<br>
Rob has pointed out that Clojure 1.10 has experienced a performance<br>
regression on JDK11, which seems to be getting patched in the 1.10.1<br>
release. However, that release is not yet available; currently only the<br>
preview build 1.10.1-beta2 is out. It is possible (likely?) that 1.10.1<br>
will not be released in time for buster.<br>
<br>
Here is some of our IRC banter on the subject for the record:<br>
<br>
13:49 < rlb> ehashman: fwiw, looks like they maybe they did fix it. Invoking <br>
1.10.1-beta2 clojure.main via a lein dummy app uberjar:<br>
13:49 < rlb> $ cljr10<br>
13:49 < rlb> Clojure 1.10.1-beta2<br>
13:49 < rlb> user=> <br>
13:49 < rlb> <br>
13:49 < rlb> $ time cljr10 -e ''<br>
13:49 < rlb> real 0m0.476s<br>
13:49 < rlb> user 0m0.905s<br>
13:50 < rlb> sys 0m0.089s<br>
13:50 < rlb> <br>
13:50 < rlb> For the current debian 1.10 it's 1.8s.<br>
15:16 < ehashman> I think this is the perf regression: <br>
<a href="https://dev.clojure.org/jira/browse/CLJ-2484" rel="noreferrer" target="_blank">https://dev.clojure.org/jira/browse/CLJ-2484</a><br>
15:18 < ehashman> (I'm looking at <br>
<a href="https://clojure.org/community/devchangelog#_release_1_10_1" rel="noreferrer" target="_blank">https://clojure.org/community/devchangelog#_release_1_10_1</a> )<br>
<br>
I think it is likely that we will *not* be able to take the patch for<br>
buster, but wanted to let the team know. If it was possible, this<br>
performance regression fix would be the last Clojure change targeted for<br>
buster, assuming the upstream release comes out on a schedule that works<br>
for the release team. I have no idea when 1.10.1 is coming out, but<br>
based on previous release schedules, since a release candidate isn't<br>
even available yet, I think it will be over a month.<br>
<br>
I do not want to try cherry-picking the performance fix while ignoring<br>
the rest of the upstream changes---while the patch is available on the<br>
JIRA I linked above, I don't know if it will apply cleanly to 1.10.0 and<br>
it's quite involved (>100 lines) so I don't want to accidentally<br>
introduce a regression or incompatibility with upstream.<br>
<br>
<br>
Until next time,<br>
<br>
- e<br>
_______________________________________________<br>
Pkg-clojure-maintainers mailing list<br>
<a href="mailto:Pkg-clojure-maintainers@alioth-lists.debian.net" target="_blank">Pkg-clojure-maintainers@alioth-lists.debian.net</a><br>
<a href="https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/pkg-clojure-maintainers" rel="noreferrer" target="_blank">https://alioth-lists.debian.net/cgi-bin/mailman/listinfo/pkg-clojure-maintainers</a></blockquote></div>