Release notes for MOZART 7 [Each service also pack implements the improvements of all previous ones.] ================================================================ Version 7.1.6 October 2004 -Denis Mitchell has supplied an example where changing the pitch of one note on a chord did not cause the notes to be correctly aligned either side of the stem. Thanks Denis - this is now fixed. -Denis's example also demanded that extra space should be inserted between staves, when in fact it was not necessary. This is now fixed too. Darrel Hoffman has also noted the spurious extra space warning and hopefully this fix will cure his case too. -Raymond Robijns has sent in a .mz file showing that if a 1st time bar starts at the start of a line and ends somewhere on the following page, then the first part is not correctly drawn. Thanks Raymond; this is also now fixed. -It was reported on the MOZART mailing list that just clicking on a scroll button causes the music to scroll a little way, even if the scroll bar is not moved. This problem is more or less eliminated in this service pack. -Thanks to Bengt Johnsson for pointing out that the "BarPad" command sometimes expands a selected block in an infelicitous way. This is now fixed. -While fixing the above command it became clear that it also does not work correctly on all parts when all parts are selected. This too is now fixed. -Thanks to Jeroen Neve for finding a bug when entering a key change on a line change at the end of the piece, when there is no space between the barline and the end-of-part marker. This is now fixed. -Thanks also to Jeroen for finding a case where inserting a beat before grace notes in a standard strand caused misalignment of bars if another strand contained chord symbols. This is now fixed too. -There have been some problems with 8va sections mysteriously disappearing, and possibly hairpins too. It has been a difficult problem to reproduce consistently, but this is now fixed in all the circumstances we have been able to test. Thanks to Simon Crombie, Darrel Hoffman, Stephen Martin, and John Burgess for persevering with reporting this. -8va symbols correctly caused notes within their span to play back at the octave. However ornaments (trills, mordents, turns, and grace notes) have been steadfastly ignoring the 8va. They have now been coerced into compliance. ================================================================ Version 7.1.5 July 2004 -When using backspace with the caret beneath a sequence of stacked chords, MOZART was correctly deleting the bottom note from each, until it entered a triplet, whereupon it started deleting the top note. Thanks to Philip Hawkins for noticing this and reporting it. It is now fixed. -Thanks to Bengt Johnsson for pointing out that the visibilty of following accidentals was not being corrected in some cases when notes were merged. This is now fixed. -Thanks also to Bengt for showing that if you use a repeat-the-bar command on a chord symbol stave the results arepretty much random. I can only assume that neither I nor anyone else has ever tried this before (as there are other ways of dealing with repeated chord symbols). But that is no excuse. It is now fixed. -MOZART 7 has been crashing when trying to read files saved by MOZART 1. Thanks to Werner Marti for finding this one and reporting it. The problem turned out to be from a change in behaviour between the previous versions of Microsoft's C++ compiler and the new one used to compile MOZART 7, and very specific to the module of MOZART which reads the oldest file formats. It is now fixed. -A small problem when inserting notes at the end of a piece after an extra barline has now also been corrected. -A bigger problem has crept in as a side effect of a correction made in version 7.1.3. In some cases entering a sequence of short notes in a part in parallel with another part which already contains longer notes has been resulting in misalignment of notes. This was not a problem in 7.1.2, but has been one in 7.1.3 and 7.1.4 If you are using either of those versions it is absolutely essential to install service pack 7.1.5 which fixes it. ================================================================ Version 7.1.4 July 2004 -Jeroen Neve found a MIDI file - SultansOfSwing.mid which left MOZART in dire straits. Importing with the swing option left a rather uneven play-back. The bug which caused this is now fixed. -In chord names with flattened 5ths and 9ths the flat sign was not appearing correctly (or indeed at all). This is now fixed too. (I found this one by myself: if someone has told me about it, my apologies I missed it.) -Theodore Stathis was finding that the File/Open command did not always open the file selector box. Thanks particularly to Theodore for being very patient indeed while I bombarded him with questions. It took a while to find the problem as I couldn't work out why only he apparently had he problem, and only after installing a service pack. Having ruled out possible complications from the Greek version of Windows, the answer eventually became clear. A bug crept in in Service Pack 7.1.0 when the open file selector was made resizable. MOZART 7.1.0 to 7.1.3 have problems with File/Open after you save a .mz file to the root directory of any drive - eg a:\. This bug hunt was something of a marathon, but it is now fixed! ================================================================ Version 7.1.3 May 2004 -Gary Burke has reported a case where DeleteBeat fails. This was where the notes from the following bar which moved back to fill the deleted space were triplets. Thanks to Gary's clear illustration of the problem this is now fixed. -Jeroen Neve has observed that if you edit the score to create one with no strands at all, then disaster ensues. As far as we are aware, no-one has tried this before, but henceforth it will be safe if anyone does. Nice one Jeroen! :-) -The GoToBar command moves the caret to the specified bar and scrolls so that the caret becomes visible in the window. Raymond Robijns has pointed out that if the caret is already in that bar, but is off screen, then MOZART complacently fails to scroll to it. This is now improved and MOZART now always scrolls to render the desired bar visible even if the caret is already there. -Raymond has also discovered a clever way to unbalance hairpins. If you moved to the end of a hairpin and enter a pedal marking, MOZART would replace the end of the hairpin with the pedal mark, leaving the other end of the hairpin without its twin, resulting in rather random drawing. This is now fixed too: the pedal mark is added without the hairpin end being replaced. ================================================================ Version 7.1.2 March 2004 -Jeroen Neve has found problems with the BeatInsert command if the beat to be inserted is exactly equal to te length of the bar. This is now fixed- MOZART explicitly prevents it. (In any case it woule be the same as inserting a bar - you just have to use the BarInsert command.) -Greg Dunn has discovered a bug if a chord symbol stave is given a transposition (to provide a capo). The bug kicked in if the transposition was far enough that it required an enharmonic replacement - in thiscase Bbb by A. This is now fixed too. -Jeroen Neve has also found a problem (well two in fact) merging motes in the vicinity of beams which straddle rests. Sometimes a note could lose its stem, and sometimes the beam angle could end up rather odd. No longer! -MOZART tries to avoid naming chords with double flats or double sharps in the names. On our travels around the above bugs we found a small inconsistency which meant a double flat name might occur where it shouldn't. This is now less likely. ================================================================ Version 7.1.1 January 2004 -Jeroen Neve has noticed that sometimes Ctrl+Spacebar does not allways leavre the caret on the note just entered as it is supposed to. This is now fixed. -Jeroen has also discovered that while one is completing the final bar, the default accidental signs are not properly represented if one completes a second (or later) strand on tha stave first. This is fixed too. -Jeroen has also found a systematic way of creating orphaned hairpin ends and in the process getting an invalid number of beats in the bar. His wonderfully clear report of how to to this has enabled it to be fixed here. ================================================================ Version 7.1.0 December 2003 -The instrument dialogue was not always selecting the existing instrument in the list on the left. Thanks to Stephen Martin for finding this one. It is now fixed. -Thanks to Denis Mitchell for noting that text items were not always getting redrawn correctly after editing. This is now fixed too. -Thanks to Jeroen Neve for finding a nasty bug which appeared when he tried to change a passage from 4/4 into 2/4. I am still trying to work out how the code has ever managed to work at all since this command was introduced, but it is fixed now. -Also fixed is the fact that different Mozart version running simultaneously open files starting from the directory which was last used by any of them. MOZART 7 now remembers its own last used directory. (Though earlier versions will not.) -The open file dialogue for *.mz files is now resizable. -"Copy page image" no longer puts a black line down the right hand size. Thanks to Jon McNamara for prompting me most recently about this one. The image width was getting rounded up to a multiple of 32 pixels and the black area was the "extra" bit. I still haven't worked out why this is happening, but the problem is bypassed for the moment by painting the extra bit white. -Support is added for a possible future "chromakey" 3rd party plug in module as explained at http://www.chromakey.net. This new feature is the reason for the version increase to 7.1. ================================================================ Version 7.0.10 November 2003 -A run of grace notes on a line following one which ended in a run of grace notes was not being drawn correctly. This is now also fixed. -When importing a NIFF file, MOZART did not always recognise a triplet (or other multiplet) if it occured at the end of a bar. This came to light by chance when we were testing the scanning software PhotoScore 3 which now exports NIFF. It is now fixed. -Thanks to John Cox for providing bitmap, and NIFF files obtained form scanning a piece with an anacrusis and various balancing bars with repeat dots. MOZART had real problems importing it but all should be well now. -Thanks also to Jon McNamara for sending bitmap and NIFF files with a very similar problem. His image as analysed with SharpEye is now also importing very nicely and so maybe I have cracked it? ================================================================ Version 7.0.9 October 2003 -Thanks to Phil Pendleton for exposing a bug which appeared when he tried to transpose a piece with 4 strands on two staves. This is now fixed. -Thanks to Jeroen Neve for reporting that phrase marks didn't always end where he expected them to after he inserted them, and that they weren't always straightforward to delete. This is now improved. -Thanks also to Jeroen for noticing that when you enter accidentals on the second strand on a stave before entering any music on the first strand on the stave, then they didn't show up. They do now. -Thanks yet again to Jeroen for finding a MIDI file which defeated MOZART's defences against MIDI files with unnotatable note values. MOZART's defences are now more robust. -A minor memory leak in one branch of the program has been exposed and fixed. ================================================================ Version 7.0.8 July 2003 -It turns out that there has been a problem with previous service pack installers - in particular the updated NIFF import filter was not getting transferred to the hard disk. This can cause severe problems when importing from NIFF. Thanks again to Davor Juric for helping me locate the problem. This version creates a complete set of compatible files of version 7.0.8. ================================================================ Version 7.0.7 July 2003 -It turns out that the NIFF import filter improvements in 7.0.6 do not always work. If clef changes occur in the middle of bars instead of at barlines, and on staves containing more than one strand, where one of the strands needs ghosts adjacent to the clef change, then they may be missed by the import filter. Thanks to Davor Juric for providing a NIFF file with exactly these characteristics and enabling the problem to be fixed in this service pack. -Thanks to Jeroen Neve for discovering some unpleasant behaviour if you have a hard line feed on the final bar of the piece and no space between that and the end of tune marker. This is now fixed too. -Thanks also to Jeroen for finding a case where rounding errors caused different pagination on printer and screen. This should only have happened very rarely; it shouldn't happen at all from this service pack onwards. -Thanks yet again to Jeroen for noting that after switching between portrait and landscape formats and closing the window, MOZART didn't always prompt you to save it. It does now. -The keyboard shortcuts on the chord symbol selector dialogue were not always working in the nice way which MOZART 6 did. Thanks to Alex Ingram for pointing this out. It is now fixed. -In rhythm parts on a chord stave, it was not always possible to tie chords together. Thanks to John Mowad for pointing this out. It is now fixed. -Colin Opie likes three beams on his groups of grace notes, and has said "wouldn't it be nice if MOZART remembered how many you set on the previous group you edited, instead of always offering two as a default?" I have to agree with him. So now it does - within one session at least. -Copying an pasting sections of tune where the copied bloc started/ended inside a hairpin or 8va section, were giving oproblems with mismatched hairpin/8va ends. This seems to have crept in recently, but thanks to Raymond Robijns who sent a file with instructions to reproduce the problem, it is now fixed. -Thanks also to Raymond for finding and illustrating some peculiar behaviour when hairpins and/or 8va's are thrown on to the away stave of a Grand Staff. This is now fixed too. -Thanks to Jeroen Neve for pointing out that bar numbers (after about two lines) don't behave themselves consistently when you add/remove an anacrucis. They do now. -Following discussion on the mailing list more room has been made for longer directory names on the global-preferences/folders dialogue. =============================================================================================== Version 7.0.6 June 2003 -Separating staves which were part of a Grand Staff could result in them being unrecombinable. Thanks to Neil Woolford, Ben Neville, and Vincent Sanzotti for finding this one. It is now fixed. -Deleting bars from within a multiple bars rest symbol only deleted one bar. Thanks to Jeroen Neve for finding this. It is now fixed. -Thanks also to Jeroen for noticing that beamed groups which go from one stave to the other of a grand staff were not getting reformatted correctly when the stave spacing is changed. This too is now fixed. -Stave label creep - a mysterious phenomenon by which stave labels gradually moved to the right, overlapping the stave, is now a thing of the past. This has plagued MOZART and its author since at least the release of MOZART 6 but until now, no-one has actually been able to see it happening. The reason became clear when Jeroen found out that every time a file was saved and reopened, the stave labels crept a very tiny amount to the right, and that this becomes clear if you do [save then open] 10 times in quick succession. This most elusive of bugs thanks to Jeroen, took only ten minutes to fix. -Glissandi were not aligning correctly with notes when the notes were o different score lines. This is now fixed. Thanks to Jeroen for this one too. -Bengt Johnsson noticed that appending chords at the end of a piece was not working properly in some circumstances. This is now fixed too. -8s were not appearing where desired under clefs in parts which transpose an octave down. Thanks to Jill Surkin and Robert Charlesworth for finding this one. It is now fixed. -Copying any part with lyrics unless it was the top one, was causing omiting lyrics when trying to paste. Thanks to Colin Fine for picking this one up, and sending an easily reproducible example. It is now fixed. -Merging adjacent notes must be done with the caret between them. If attempted with the caret actually on a note, this was causing severe problems, which were probably at the root of one or two bug reports, including another one from Colin Fine. Thanks to Colin and all who reported related problems; MOZART now just beeps at you if you try this illegal operation. -Colin has also noticed that, with some arrangements of lyric lines, the Shift+Ctrl+up/down arrow keys do not run through staves and lyric lines in the desired order. Thgis has now been refined. -Thanks to Jeroen Neve for noticing that toglling between phrase marks above and below the notes wasn't always working. This is now fixed. -Robert Charlesworth discovered cases where editing time signatures needs to add a final barline but didn't. One example is 3n bars of 4/4 followed by an incomplete bar of 3 beats. This can be edited to 4n+1 bars of 3/4 and a bar line needs to bee appended. Thanks Robert - it is now. -The NIFF import filter was not always picking up clef changes within the music. Thanks to Jean-Paul Mouris for discovering this and sending a sample NIFF file to test it on. It is now fixed. -Another problem (or two) with NIFF import has been found by Greg Dunn. Importing text items was not always happening properly. Thanks Greg, this is now fixed (though there is not yet provision for importing lyrics). -Thanks also to Jeroen Neve for noticing some subtle enharmonic infelicities in NIFF imports. MIDI does not know the difference between (say) Eb and D# and so in post- processing the import MOZART has to make choices. NIFF does know the difference and the import filter is now correctly reporting to MOZART that it does, and so MOZART knows it can safely omit this post-processing step for NIFF imports. -Thanks to Michael Bednarek for noticing that after changing the bracket style on a multiplet the default block selection mode always defaulted to all staves. It shouldn't do that of course, and now it doesn't. -Thanks also to Michael for suggesting that the Paste operation sometimes takes long enough to make the hourglass cursor desirable. This is hereby introduced. =============================================================================================== Version 7.0.5 May 2003 -Michael Bednarek found a file with misaligned tempo changes which MOZART 7 completely failed to read. This is fixed now. -Transposing pieces was sometimes deleting chord shapes, rather than entering new ones. This is now fixed. -MOZART 7 was not running under Windows NT. This is now fixed and it runs under Windows NT SP6a. Thanks to Cliff Silvernail for finding this and reporting it in a way which made it straightforward to fix. =============================================================================================== Version 7.0.4 was the April 2003 release version. ===============================================================================================