[trunk] irregular measures not saved
Ubuntu 10.04
Mscore 3944M
Make two irregular measures (3 and 4 in the attached png file). They have respectively 3/4 and 1/4 . Others measures are reported as 4/4, regular and 4/4 in the irregular fields
Save, reload The irregularity has been destructed. Any measure of the 12 is reported as 4/4 with 1/1 in irregular fields.
I think that the trick was possible with some previous version....I'll have a look to.
These irregular measures I use for lyrics edition purposes.
Comments
Not able to reproduce r. 3944 trunk, Windows 7
"my" 3944 is the nightly build one which in the about windows post "unstable prerelease to version 2.0, so I understand that it is not the "trunk" version. In http://bernsteinforpresident.com/musescore-nightly I don't see any "trunk" version....
is there a place to find one?
meanwhile trying to find the trunk version I installed a 3746M instable prerelease of version 1.0.0 and obtained the same phenomenon
then I just did
leleu@portable-leleu:~/musescore$ svn update
'.' omis
leleu@portable-leleu:~/musescore$ svn update '.'
'.' omis
leleu@portable-leleu:~/musescore$ cd mscore
leleu@portable-leleu:~/musescore/mscore$ ls
build Makefile Makefile.mingw Makefile.osx mscore wine.bat wrc.bat
leleu@portable-leleu:~/musescore/mscore$ make revision
leleu@portable-leleu:~/musescore/mscore$ sudo make install
[sudo] password for leleu:
cd build; make install
and obtained again the 3746M version
So at the time I'm not able to test (under Ubuntu 10.04) the same version as you
A clarification on the terminology: the trunk is the very latest development version (currently called 2.0 pre). The upcoming release of version 1.0 comes from the same stable branch as the 0.9.6 series. In preparation for a stable release, the prereleases for a stable version comes from a branch (where code changes are limited) not the development trunk (where large improvements to the code are encouraged).
so (see attached) my 3944M is a "trunk" version, isn't it?
to-morrow I have the opportunity to test on Windows seven, so I'll test the last windows nightly
Yes. Labelled 2.0 is a trunk version.
Tested the same bug with 3950M windows nightly.
attached capture show two "regular" measures, but in fact they should be of 3 and 1 beats.
attached also the mscz file which give 1/1 for the real
robert leleu, could you attached the files again--this time without pressing "Preview".
(There is a bug on the website that means the links to the files attached to a comment do not work if you press preview before you press save)
turning around the bug, here they are
in fact I attached 3, to be clearer.
also I noticed that the mscz file claims "save before closing" even as there has been non modification..
doing the same irregular measures and saving in xml, the resulting file, when opened again, is correct.
however the same operation with complex scores yield xml files.....who crash mscore when opening (see http://musescore.org/fr/node/8898)
I attach the xml file
tested 9.6.3 (windows), which is not affected
Succeeded the following test:
Open a score with last nightly version 2.00 (Ubuntu), reduce measures, save as XML
The XML file crashes mscore
Open the xml with 0.9.6.3 windows: correctly viewed; save as mscz
Open tis mscz with version 2.00: the score is OK, irregular measures are correctly seen as 1/4 or 3/4
Confirmed using nightly trunk r. 3955, Windows 7
seems solved
I opened a 0.9.6 3072 score (with irregular measures) under 2.0 3969 and it was OK
not solved!
I have a clear exemple, whose file I attach.
The first measure of the first sung staff should be irregular 1/4: it seems correct at first configuration, but after reload is no longer, and is reported 1/1, and configuring again 1/4 make disappear the first measure... the graphical effect is not good, see attached png
Similarly the last mesasure of the same staff should be 3/4
Not only irregular measures are not saved, they are not managed correctly while they are empty.
My setup: trunk SVN rev. 3954; Windows XP, Qt SDK 10.04.
Steps: The attached PNG (one image is really worth one thousand words!) describes 2 processes showing that measures are corrupted if made irregular while empty, but are managed correctly if made irregular when not empty (.mscx files referred to in the image are also attached).
Hoping it helps,
M.
P.S.: priority increased, as this leads to data corrupted or lost.
Ops, forgot to increase priority!
perhaps not so critical
I attach another score with irregular measures, integrally made with the last nightly build (trunk 3969)
Irregular measures don't disappear when reload.
Previous scores were mor or less "updated" from previous versions
I again made a score with irregular measures. It worked perfectly.
But pay attention if you have empty staffs: you must transform "default whole rests" in actual rests in each staff for the measure you want to make irregular.
So I think the bug is no longer critical, and change it to minor.
Isn't this fixed?
M.
Appears fixed to me.
Tested using f5dc176 Windows 7
Automatically closed -- issue fixed for 2 weeks with no activity.