Re: Use a stable sort when ordering MIDI items (issue 353790043 by mrbobo1...@gmail.com)

2018-11-03 Thread Carl . D . Sorensen
On 2018/11/03 12:39:41, Dan Eble wrote: The ticket for this review is https://sourceforge.net/p/testlilyissues/issues/5434/ . Carl, it sounds like James needs clarification as to whether you are still pressing for changing more sort calls to stable_sort calls. MHO is that this change

Re: Use a stable sort when ordering MIDI items (issue 353790043 by mrbobo1...@gmail.com)

2018-11-03 Thread nine . fierce . ballads
The ticket for this review is https://sourceforge.net/p/testlilyissues/issues/5434/ . Carl, it sounds like James needs clarification as to whether you are still pressing for changing more sort calls to stable_sort calls. MHO is that this change stands fine on its own.

Re: Use a stable sort when ordering MIDI items (issue 353790043 by mrbobo1...@gmail.com)

2018-11-01 Thread nine . fierce . ballads
On 2018/11/01 03:23:21, c_sorensen wrote: On 10/31/18, 7:32 PM, "Dan Eble" wrote: But I still wonder if we should use both sort and stable_sort. Seems like we have no reason (unless it's a performance issue) to use both. Using a stable_sort where it isn't

Re: Use a stable sort when ordering MIDI items (issue 353790043 by mrbobo1...@gmail.com)

2018-11-01 Thread MrBobo1239
On 2018/11/01 01:50:38, Dan Eble wrote: > (I'm surprised this hasn't been fixed until now.) Welcome to LilyPond, where you supply the fixes. Your change looks good to me, but judging from your questions, it sounds like it could be tested better. Have you read the chapter on Regression

Re: Use a stable sort when ordering MIDI items (issue 353790043 by mrbobo1...@gmail.com)

2018-10-31 Thread Carl Sorensen
On 10/31/18, 7:32 PM, "Dan Eble" wrote: On Oct 31, 2018, at 14:00, carl.d.soren...@gmail.com wrote: > > Why not always have our sort use stable_sort? Because when someone finally clears away std-vector.hh, they will be more likely to mess it up if sort() needs to be

Re: Use a stable sort when ordering MIDI items (issue 353790043 by mrbobo1...@gmail.com)

2018-10-31 Thread nine . fierce . ballads
(I'm surprised this hasn't been fixed until now.) Welcome to LilyPond, where you supply the fixes. Your change looks good to me, but judging from your questions, it sounds like it could be tested better. Have you read the chapter on Regression Tests in the Contributor's Guide? You should

Re: Use a stable sort when ordering MIDI items (issue 353790043 by mrbobo1...@gmail.com)

2018-10-31 Thread Dan Eble
On Oct 31, 2018, at 14:00, carl.d.soren...@gmail.com wrote: > > Why not always have our sort use stable_sort? Because when someone finally clears away std-vector.hh, they will be more likely to mess it up if sort() needs to be transformed into stable_sort() than if the functions parallel those

Re: Use a stable sort when ordering MIDI items (issue 353790043 by mrbobo1...@gmail.com)

2018-10-31 Thread MrBobo1239
I've built two versions of lilypond to measure the performance impact of always using `stable_sort`. The baseline (reference) is commit 964722f804 without any modifications and the second build just changes `vector_sort` to use `sort_stable`. The test file is from Mutopia-2015/11/04-2050

Re: Use a stable sort when ordering MIDI items (issue 353790043 by mrbobo1...@gmail.com)

2018-10-31 Thread Carl . D . Sorensen
Why not always have our sort use stable_sort? Have you tried with a large score (e.g. one from mutopia) to see what the resource implications are? https://codereview.appspot.com/353790043/ ___ lilypond-devel mailing list lilypond-devel@gnu.org

Use a stable sort when ordering MIDI items (issue 353790043 by mrbobo1...@gmail.com)

2018-10-31 Thread MrBobo1239
Reviewers: , Message: Hi all, this is a patch to fix the issue described here: https://lists.gnu.org/archive/html/lilypond-user/2015-02/msg00035.html (I'm surprised this hasn't been fixed until now.) I'm not familiar with lilypond internals at all so I don't know whether this is correct way to