classification
Title: stop using ranlib
Type: Stage: resolved
Components: Build Versions: Python 3.7
process
Status: open Resolution:
Dependencies: Superseder:
Assigned To: Nosy List: benjamin.peterson, cstratak, kadler, vstinner
Priority: normal Keywords: patch

Created on 2017-09-29 06:15 by benjamin.peterson, last changed 2018-11-09 16:42 by kadler.

Pull Requests
URL Status Linked Edit
PR 3815 merged benjamin.peterson, 2017-09-29 06:24
PR 3824 merged benjamin.peterson, 2017-09-29 16:26
PR 10417 open cstratak, 2018-11-08 16:42
Messages (8)
msg303301 - (view) Author: Benjamin Peterson (benjamin.peterson) * (Python committer) Date: 2017-09-29 06:15
As far as I'm aware, every modern *nix's ar supports an "s" flag that removes the need to run ranlib separately on a a static library. We should just do that and stop running ranlib. That saves us some lines in configure.ac and the Makefile.
msg303332 - (view) Author: Benjamin Peterson (benjamin.peterson) * (Python committer) Date: 2017-09-29 15:42
New changeset d15108a4789aa1e3c12b2890b770550c90a30913 by Benjamin Peterson in branch 'master':
stop using ranlib (closes bpo-31625) (#3815)
https://github.com/python/cpython/commit/d15108a4789aa1e3c12b2890b770550c90a30913
msg303333 - (view) Author: STINNER Victor (vstinner) * (Python committer) Date: 2017-09-29 15:48
Buildbots are unhappy :-( Compilation failed on:

http://buildbot.python.org/all/builders/AMD64%20FreeBSD%20CURRENT%20Debug%203.x/builds/962/steps/compile/logs/stdio

http://buildbot.python.org/all/builders/AMD64%20FreeBSD%20CURRENT%20Non-Debug%203.x/builds/964/steps/compile/logs/stdio
msg303336 - (view) Author: Benjamin Peterson (benjamin.peterson) * (Python committer) Date: 2017-09-29 16:26
New changeset 6fb0e4a6d085ffa4e4a6daaea042a1cc517fa8bc by Benjamin Peterson in branch 'master':
explicitly list objects for the ar command (#3824)
https://github.com/python/cpython/commit/6fb0e4a6d085ffa4e4a6daaea042a1cc517fa8bc
msg303338 - (view) Author: STINNER Victor (vstinner) * (Python committer) Date: 2017-09-29 16:28
Thanks ;-)
msg329001 - (view) Author: Charalampos Stratakis (cstratak) * Date: 2018-10-31 17:44
Would it make sense to backport that to 3.6 as well? Currently this blocks https://bugs.python.org/issue28015 from being backported to 3.6
msg329483 - (view) Author: STINNER Victor (vstinner) * (Python committer) Date: 2018-11-08 17:27
(I reopen the issue.)

> Would it make sense to backport that to 3.6 as well? Currently this blocks https://bugs.python.org/issue28015 from being backported to 3.6

There is a risk of regression. Does bpo-28015 fix really depend on this change?

Benjamin wrote: "As far as I'm aware, every modern *nix's ar supports an "s" flag", but I'm not sure if Python 3.6 is only used on "modern Unix". Some people use AIX and HP-UX: does ar support "s" on these OSes?
msg329534 - (view) Author: Kevin (kadler) * Date: 2018-11-09 16:42
AIX supports the -s flag: https://www.ibm.com/support/knowledgecenter/en/ssw_aix_71/com.ibm.aix.cmds1/ar.htm#ar__row-d3e27561
History
Date User Action Args
2018-11-09 16:42:53kadlersetnosy: + kadler
messages: + msg329534
2018-11-08 17:27:56vstinnersetstatus: closed -> open
resolution: fixed ->
messages: + msg329483
2018-11-08 16:42:33cstrataksetpull_requests: + pull_request9697
2018-10-31 17:44:43cstrataksetnosy: + cstratak
messages: + msg329001
2017-09-29 16:28:16vstinnersetmessages: + msg303338
2017-09-29 16:26:50benjamin.petersonsetstatus: open -> closed
resolution: fixed
messages: + msg303336

stage: patch review -> resolved
2017-09-29 16:26:37benjamin.petersonsetstage: resolved -> patch review
pull_requests: + pull_request3809
2017-09-29 15:48:39vstinnersetstatus: closed -> open

nosy: + vstinner
messages: + msg303333

resolution: fixed -> (no value)
2017-09-29 15:42:44benjamin.petersonsetstatus: open -> closed
resolution: fixed
messages: + msg303332

stage: patch review -> resolved
2017-09-29 06:24:20benjamin.petersonsetkeywords: + patch
stage: patch review
pull_requests: + pull_request3799
2017-09-29 06:15:32benjamin.petersoncreate