Re: Regtest error

[James Bailey]

I always get the documentation, but there are some disadvantages to it. You 
don't see things the way a normal user who is copying and pasting the examples 
would see them. When I went through all of the regtests a couple of months ago, 
there were some issues that only appeared if you compiled the actual code in 
the regtests. I have seen some of the discussion about the regtest comparison 
checker but a) it was a bit too complex for me to follow and b) I don't have 
access to windows. 
On Nov 29, 2010, at 6:45 PM, Phil Holmes wrote:

> Entered as http://code.google.com/p/lilypond/issues/detail?id=1430
> 
> I don't know if this is the way you want to do this, but you can download all 
> the regtests as a PDF from the development documentation page?  That said, 
> there's 288 pages to wade through, and this wouldn't appear a good use of 
> your time.  We have a new volunteer for Bug Squad - Marc Schonbrun - and I've 
> asked him to start with a test-by-test check - seeing that the output is 
> actually as described.
> 
> Dunno if you've picked this up from the thread, but I've just written a local 
> regtest checker that does just what you describe, but also compares the 
> output from the old regtests with the new ones.  It's different from the 
> regtest checker since it uses a pixel-by-pixel comparison.  So if you can run 
> windows you could have a copy.  Failing that I could .zip the images and 
> stick them on a website somewhere.  About 900 images and 13 megs.
> 
> -- 
> Phil Holmes
> Bug Squad
> 
> 
> "James Bailey" <address@hidden> wrote in message news:address@hidden
> Certainly not good. It worked in the original version, 2.13.36, but hasn't 
> since 2.13.37. I usually actually try to compile all of the regtests and look 
> at them for the beta releases, but there has always been a non-alpha released 
> so quickly after an beta release that I only got through about 20 or so.
> 
> If anyone has any tips on how to compile them easily, I'd be more than 
> willing to listen. My previous method was to copy all of the ly files in the 
> regression test folders to a separate folder, remove the first three lines, 
> and then run lilypond on them. I've just gained the ability to compile the 
> documentation on OSX, so if there's a step somewhere in the creation of the 
> docs where I can generate actual files like a normal user would, it would 
> make life a lot easier for me.
> 
> 
> On Nov 29, 2010, at 4:41 PM, Phil Holmes wrote:
> 
>> Just checking this isn't known about before raising a bug report. 
>> stem-length-estimation.ly is described as "Stems with overridden 'length 
>> should not confuse height estimation. This example should fit snugly on one 
>> page.".  That'll be 1 1/2 pages in 2.13.40.  Running it against 2.12.3 gives 
>> an error because of the new spacing names, but actually takes 2 full pages.
>> 
>> -- 
>> Phil Holmes
>> Bug Squad
>> 
>> 
>> 
>> 
>> _______________________________________________
>> bug-lilypond mailing list
>> address@hidden
>> http://lists.gnu.org/mailman/listinfo/bug-lilypond 
> 
> 
> 
> _______________________________________________
> bug-lilypond mailing list
> address@hidden
> http://lists.gnu.org/mailman/listinfo/bug-lilypond