LilyPond Regression Tests

Here's the output of my LilyPond pixel regtest checker. I checked all the regtests using my latest multi-threaded program, which took 27 minutes elapsed to run. The comparison program itself took 39 minutes cpu time. I also collected the time it took to compile each file for each version. The total time for .40 was 1464 seconds, and for .41 1460 seconds. So we're getting faster...

figured-bass-continuation-center.png. Some small changes - my program circles these and the list below is the pixel that's different (x, y) with the brightness variation in []

 

figured-bass-continuation-forbid.png. Slight variation in the placement of the flat and the length of the continuation.

 

 

figured-bass-continuation.png. Slight variation in the length of the continuation.

figured-bass-durations.png. Slight variation in the length of the continuation.

figured-bass-extenders-markup.png. The markup that I don't understand is in a quite different place - further down in .41

figured-bass-implicit.png: Slight variation in the placement of the accidentals.

 

figured-bass-staff.png. Slight variation in the placement of the staves and notes.

page-breaking-min-systems-per-page2.png. .41shifts this very high on the page. Don't think this looks good, but i've not looked into the details of the test.

page-breaking-page-count3.png. .41 shifts this test high on the page.

part-combine-force-once.png. Changes to the part combine code.

stem-length-estimation-page1.png. This is good news. stem-length-estimation only produces a single page.

tablature-grace-notes.png. Slight variation in the formation of the number.

utf-8.png. Slight text variation.