Regression 2.2: File with MMRest and vertical frame created in 2.1 doesn't display correctly in 2.2
- Open the attached file created with MuseScore 2.1 in MuseScore 2.2 e37f2513
- The file is supposed to have 32 measures and look as follow, but the vertical frame and the following measures are completely missing !
| Attachment | Size |
|---|---|
| mmrest_section.mscz | 5.55 KB |
Comments
This issue occurs on December 20, 2017.
This nighlty works: a09f8e2
So, a side effect of this last mentioned nighlty/commit.118ede5
Or, of the previous commit: https://github.com/musescore/MuseScore/commit/7fd14491138bbc4db1d5cfd4f…
Well, both commits are basically the same, the former is merging the PR for the latter. And indeed it does also something with mm rests, @cadiz1: great job!
My bet is on https://github.com/musescore/musescore/commit/118ede5#diff-fd94ca9d9d96…
"Well, both commits are basically the same"
Indeed, this it was what I had just seen, without really understanding how!
In reply to "Well, both commits are… by cadiz1
Would you mind telling us how you're doing this? To me this seems like either black magic or a huge amount of work. Might be worth a page in the developers' handbook...
"Would you mind telling us how you're doing this?"
How to do what exactly?
@cadiz1 Finding the guilty commit ;)
And jojo, you are right, that's the guilty line.
Good job everyone. Let's fix it.
Identifying the culprit build
In reply to Identifying the culprit build by Jojo-Schmitz
As usual, nothing special. With archives, I go back in time, and I locate where the change occurs. It's take time of course, but this last one was easy.
One more on my hunting board! :)
You do use binary search right ? https://fr.wikipedia.org/wiki/Recherche_dichotomique
Like not trying them in sequence but using intervals ?
So no magic, just blood, sweat and tears ;-)
I wonder whether
git bisectcould help here?For 2.2 https://github.com/musescore/MuseScore/pull/3525
Checking master
Binary search? I read the article, it's Hebrew for me :)
"Nothing magic" indeed. Pragmatically, I try a nightly at the beginning of each month, March, February, etc. and I quickly realized that it was in December. It worked at first, but not at the end.
Then, just tighten the vise, by week interval, and finally on the day.
In one word: starting from a big circle, and get closer to reach the target.
More or less a binary search ;)
Master is totally different regarding MMrest and doesn't need the fix. Waiting for travis-ci to do its stuff and we are good to go ! Thank you everyone, great team work !
_it's Hebrew for me _ is that the idiom in French? In English it is "this seems Greek to me", in German it is "das kommt mir Spanisch vor" ;-)
In reply to _it's Hebrew for me _ is… by Jojo-Schmitz
French familiar expressions :) https://fr.wiktionary.org/wiki/c%E2%80%99est_de_l%E2%80%99h%C3%A9breu
or, synonym, https://fr.wiktionary.org/wiki/c%E2%80%99est_du_chinois
"More or less a binary search "
Glad to learn this! :)
Fixed in branch 2.2, commit dd9e1e39cf
fix #270136: Regression 2.2: File with MMRest and vertical frame created in 2.1 doesn't display correctly in 2.2
Fixed in branch 2.2, commit 5f6b17bab8
Merge pull request #3525 from lasconic/fix-270136_2.2
fix #270136: Regression 2.2: File with MMRest and vertical frame crea…
Automatically closed -- issue fixed for 2 weeks with no activity.