Summary: | dev-libs/jemalloc-5.3.0: re-keyword | ||
---|---|---|---|
Product: | Gentoo Linux | Reporter: | Sergei Trofimovich (RETIRED) <slyfox> |
Component: | Keywording | Assignee: | Mozilla Gentoo Team <mozilla> |
Status: | RESOLVED FIXED | ||
Severity: | normal | CC: | hydrapolic, matoro_gentoo |
Priority: | Normal | Keywords: | CC-ARCHES, PullRequest |
Version: | unspecified | Flags: | nattka:
sanity-check+
|
Hardware: | All | ||
OS: | Linux | ||
See Also: | https://github.com/gentoo/gentoo/pull/25243 | ||
Whiteboard: | |||
Package list: |
dev-libs/jemalloc-5.3.0-r1 ~amd64 ~arm ~arm64 ~hppa ~ppc ~ppc64 ~x86 ~ia64 ~riscv ~s390 ~alpha ~m68k ~mips
|
Runtime testing required: | --- |
Bug Depends on: | 856061 | ||
Bug Blocks: |
Description
Sergei Trofimovich (RETIRED)
2020-07-12 08:56:02 UTC
Possibly related: /var/tmp/portage/dev-libs/jemalloc-5.2.1/work/jemalloc-5.2.1/include/jemalloc/internal/quantum.h:65:6: error: #error "Unknown minimum alignment for architecture; specify via " 65 | # error "Unknown minimum alignment for architecture; specify via " | ^~~~~ This is on sparc32. I've now keyworded this for ~m68k. 2 tests fail with USE=prof where it was expecting a larger backtrace count, but this isn't surprising. Looks good on sparc64. sparc done Let's start over here. Keywords for dev-libs/jemalloc: | | u | | a a p s r a l | n | | m r h p p i i s l o m m | e u s | r | d a m p p c a x a s 3 p o 6 i | a s l | e | 6 r 6 p p 6 r 8 6 c 9 h n 8 p | p e o | p | 4 m 4 a c 4 c 6 4 v 0 a g k s | i d t | o ------------+-------------------------------+---------+------- [I]5.2.1-r1 | + + + + + + o + ~ ~ ~ ~ o ~ o | 7 o 0/2 | gentoo 5.3.0 | o o o o o o ~ o o o o o ~ o o | 8 o | gentoo After testing jemalloc on amd64 & x86 (and some random rdeps on both) I'd have no problems just keywording everything to 5.3.0. But... since the AT teams are much more efficient nowadays we can do it properly too. CMP: dev-libs/jemalloc-5.2.1-r1/image with dev-libs/jemalloc-5.3.0/image ABI: libjemalloc.so.2(32) func(+10) ABI: libjemalloc.so.2(64) func(+10) SIZE: 5.08MiB -> 7.91MiB, 13 -> 13 files ------> ABI(+20) SIZE(+55.58%) could be worth a subslot bump though... The bug has been referenced in the following commit(s): https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=c35c6ebdb309e11471606e91c00d8b623daab2e2 commit c35c6ebdb309e11471606e91c00d8b623daab2e2 Author: Joonas Niilola <juippis@gentoo.org> AuthorDate: 2022-06-30 05:54:10 +0000 Commit: Joonas Niilola <juippis@gentoo.org> CommitDate: 2022-06-30 05:54:10 +0000 dev-libs/jemalloc: keyword 5.3.0 for ~x86 Bug: https://bugs.gentoo.org/732316 Signed-off-by: Joonas Niilola <juippis@gentoo.org> dev-libs/jemalloc/jemalloc-5.3.0.ebuild | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) https://gitweb.gentoo.org/repo/gentoo.git/commit/?id=4b06025cb56aa2105b4e6a0580573fa7ed35f301 commit 4b06025cb56aa2105b4e6a0580573fa7ed35f301 Author: Joonas Niilola <juippis@gentoo.org> AuthorDate: 2022-06-30 05:53:55 +0000 Commit: Joonas Niilola <juippis@gentoo.org> CommitDate: 2022-06-30 05:53:55 +0000 dev-libs/jemalloc: keyword 5.3.0 for ~amd64 Bug: https://bugs.gentoo.org/732316 Signed-off-by: Joonas Niilola <juippis@gentoo.org> dev-libs/jemalloc/jemalloc-5.3.0.ebuild | 2 +- 1 file changed, 1 insertion(+), 1 deletion(-) riscv done (In reply to Joonas Niilola from comment #6) > CMP: dev-libs/jemalloc-5.2.1-r1/image with dev-libs/jemalloc-5.3.0/image > ABI: libjemalloc.so.2(32) func(+10) > ABI: libjemalloc.so.2(64) func(+10) > SIZE: 5.08MiB -> 7.91MiB, 13 -> 13 files > ------> ABI(+20) SIZE(+55.58%) > > could be worth a subslot bump though... Only an ABI break "going backwards". Introducing new functions in a new version is okay (happens all the time - there's nothing to break), but dropping them isn't (because built binaries may depend on them), nor is changing structs, function signatures, ... That's why if you downgraded libjemalloc now after building pkgs against the new versoin, those pkgs may well be broken, but we turn a blind eye to it because backwards compatibility (or forwards, whatever you want to call it) is kind of impossible to support properly. (If we were to worry about that, we'd have to handle a LOT more packages.) TL:DR: all fine b/c just added symbols, not modified or dropped. arm64 done s390 done hppa done ppc done arm done ppc64 done m68k done ia64 done alpha done mips done all arches done |