New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
perl -V 1000s of lines long/make_patchnum.pl has unbounded CPU and disk usage #14805
Comments
From @bulk88Created by @bulk88While working on a branch, I noticed make_patchnum.pl was taking 2-3 -----------------------------------------------------------------
Version Number: Windows NT 5.1 (Build 2600) C:\perl521\srcnewb4opt>timeit git.exe cherry b88origin/sockethandlefix Version Number: Windows NT 5.1 (Build 2600) C:\perl521\srcnewb4opt>timeit git.exe cherry b88origin/sockethandlefix Version Number: Windows NT 5.1 (Build 2600) C:\perl521\srcnewb4opt> HEAD-2 is a p5p repo commit for me. The problem seems to be a rebased ----------------------------------------------------------------- ----------------------------------------------------------------- WIP #118127 don't closesocket on non-socket handles on Win32 *********************CUT********************* commit 856d3623f8d51a9a6e6231593eb5d9da23f7d0a8 b88 bld improve debugging of padlist API *********************CUT********************* Characteristics of this binary (from libperl): Perl Info
|
From @tonycozOn Thu Jul 16 22:45:56 2015, bulk88 wrote:
What commit does b88origin/sockethandlefix point at? git rev-parse b88origin/sockethandlefix Tony |
The RT System itself - Status changed from 'new' to 'open' |
From @bulk88On Tue Jul 21 22:51:42 2015, tonyc wrote:
Owner@OWNER-5D26B85EE /C/perl521/srcnewb4opt (sockethandlefix) Owner@OWNER-5D26B85EE /C/perl521/srcnewb4opt (sockethandlefix) WIP don't closesocket on non-socket handles commit 721000cb2bcbd3292ef9171f35acf43985eeea15 b88 bld commit ae103e0 fix multi-eval of Perl_custom_op_xop in XopENTRY : : : (END) Owner@OWNER-5D26B85EE /C/perl521/srcnewb4opt (sockethandlefix) I think I see what is happening, git "upstream" is the ancient branch on my remote, not p5p remote blead. I still wonder if there should be a limit of a couple 100 commits in make_patchnum.pl, maybe with make_patchnum.pl exit 1 and not writing a git_version.h if that limit is hit. -- |
From @tonycozOn Wed Jul 22 04:38:09 2015, bulk88 wrote:
...
Yes, it's the remote for that branch, it has nothing to do with the date of the commit. You can change the remote for that branch with: git branch --set-upstream somebranch upstreambranch or by editing .git/config. As to limiting the number of entries, I'd accept that, but I could see others having a different opinion. For me to accept a patch it would need something to indicate how many commits weren't listed, so: 3480fba Tony |
From @bulk88On Wed Jul 29 20:57:07 2015, tonyc wrote:
Does it have to say "120 other commits" or just a line that says "commit list truncated at 100 commits"? My git cmd lines skills are poor, but to find how many skipped commits there were, the git client would need to dump all the SHA1s to make_patchnum.pl anyway and the 10s of seconds the git client took will still happen, the only point of truncating the list at that point is so the perl binary doesn't grow by 100s of KB to store the text of all those SHA1s but the 10s of seconds spent in the git client during the build process are still there. -- |
From @tonycozOn Sun, Aug 02, 2015 at 09:14:27PM -0700, bulk88 via RT wrote:
I think it's useful to give the user some measure of just how far they How were you planning on limiting the commits? With HEAD at blead, the following: git cherry origin/maint-5.20 HEAD HEAD~100 takes 4 seconds, and lists 187 commits. From reading C<git help cherry> it looks like the limit parameter only Tony |
Given a lack of response, I don't see anything actionable here. I propose closing this case. |
Migrated from rt.perl.org#125630 (status was 'open')
Searchable as RT125630$
The text was updated successfully, but these errors were encountered: