On Sunday 02 May 2004 23.00, Han-Wen Nienhuys wrote:
> Couldn't we just have the entire collection rendered just like
> input/test in the tips & tricks? Putting binary blobs, which are also
> compiler output (by lilypond), into a source version control system
> isn't the way to go I.M.O.

I see the pngs only as an illustration to the bug description. Different pngs 
may be created in different ways, see e.g. grace-accidental-spacing.png. The 
reason why most pngs are direct lilypond output, is that this often is the 
easiest & clearest way of doing it (but it would work just as well e.g. to 
draw a picture of lilypond's output by hand)

There are 2 reasons they are there:
1. To make it more obvious to you developers what I mean
2. That I will be able to easily compare the .ps output of new versions to the 
old buggy output, so I quickly can decide whether the bug has been fixed.

What you describe is AFAICU some kind of history of how layout bugs evolve 
between lily versions. I have not seen any use for this myself yet. Bugs tend 
to be binary; first it doesn't work, then it works; so in this case the only 
relevant history is contained in the date the bug was removed from CVS.

I agree that using cvs for uploading png may seem a bit strange. But it works 
and is very simple. "cvs add -kb foo.png" is much easier than using ftp. And 
the file will be downloaded automatically to . during cvs update, which is 
much easier than having to open mozilla.

I'll be happy to change the system once there is a really good reason for it, 
though.

Erik


_______________________________________________
bug-lilypond mailing list
[EMAIL PROTECTED]
http://mail.gnu.org/mailman/listinfo/bug-lilypond

Reply via email to