On Sun, Apr 24, 2016 at 02:43:26PM +0300, Niko Tyni wrote: > I wonder if it would be useful to prefix the patch (file)names with > something else than the current 'debian/' or 'fixes/', perhaps '5.20.3/', > to highlight their origins? This might at least make 'perl -V' output > a bit clearer.
Dominic proposed fixes/5.20.3/ which is fine by me assuming git-dpm copes (I expect it does). > Another thing: we need to make sure the fixes are in unstable + testing > before pushing them to stable. In this case, given 5.20.3 was released > before 5.22.1 (see perlhist.pod), I expect it's all fine but it shouldn't > hurt to check. I guess I'm sort of volunteering :) OK, I've had a quick look at these. They are all in 5.22.1 so fixed in sid/stretch, no worries about that. Here are a few things that caught my eye while at it. +1 from me to all the others on the list. d40f1ca59e9f4eb4e0e717b5304072636b24a62a lib/perl5db.pl: Restore noop lock prototype this is already in jessie as fixes/perldb-threads.diff 0f19268d11e0e7ce38b5a449f9f53461a0dc3226 [perl #123748] - Add test case for possible getenv/putenv/setenv stomping in perl.c (also squash 63209b393029691eb62065536b4aaab4ade1ad7b into it) This is not a functional change, just a new test. Do we want that? Also, I'm not sure if squashing is the right thing to do but I don't feel strongly about that. 8d89c0509dd5eb1de58dc6617f6e08599eb24792 [PATCH] [perl #123786] don't leak the temp utf8 copy of namep This is the only one that's not a cherry-pick from another branch so it's worth a mention, though it definitely needs to be included. The ticket says: "This isn't an issue in blead, since the leaking code is no longer present - all pad names are in utf-8 so there's no need to convert." Blead at the time (Feb 2015) was pre-5.22 so it should be OK on sid. I've verified that the test case leaks on jessie but not sid. ccafce1bfd59cdbdd7af0ad68f7557a7471d6c64 Perl_save_re_context(): re-indent after last commit Whitespace only but would be nice to group with 17d9707d444517764c7bcb479c236a8c58a1d605 Again, to squash or not to squash? 0cb29f21ba1db04416bb1a2dbc32cd5f7c75bc84 lib/h2ph.t to test generated t/_h2ph_pre.ph instead of the system one Another test-only commit. 41575f0a39566008526e7e38eab84cba82cf372b Add pthread to libswanted This is not a doc fix though it was grouped there. We override libswanted in debian/prune_libs.diff so maybe not that useful? -- Niko