Re: Check for null pointer

[Dan Eble]

On 2011-08-20, at 16:00 , Reinhold Kainhofer wrote:

> So, basically, you are saying that whenever you run "make X" through color, 
> it 
> will crash, but it will not crash in any other cases? 
> Are you sure this is not a problem of color together with make (I had some 
> strange problems that only occured when something was called by make, because 
> the locale settings were messed up)?

I've been using color with make for years without any problems (really that's 
the only thing I use it with!); however, I just recently started using OS X 
10.7, so I wouldn't rule it out.  On the other hand, I did observe a very 
similar crash running 2.15.9 in gdb without either of them.

>> I tried Graham's experimental 2.15.9 (from last week) and it worked fine at
>> first, but then I tried running it via gdb and it crashed with similar but
>> not identical backtrace as 2.14.2.  (The caller of kill_mmrest is
>> Part_combine_iterator::process instead of Part_combine_iterator::unisono).
> 
> Hmm, do you have any example file where this appears (even if it's not 
> minimal, for now it's just important that we can reproduce it somehow).

Be my guest.

1. extract http://faithful.be/sheet-music/dfe-books-20110820.tar.bz2
2. edit the top-level Makefile to set the LILYPOND variable
3. make PraiseAndPrayer PARTS=piano

Once the output directories are created you should be able to run it again 
without make, if you want.  (Running in gdb is when I see 2.15.9 crash.)  The 
command would be

lilypond -Iinclude -dno-point-and-click -e '(begin (primitive-load 
"src/scm/part-analyzer.scm") (primitive-load "src/scm/part-combiner.scm"))' 
-Iinclude/part/piano -Isrc -o piano src/books/PraiseAndPrayer.ly

Thanks for your help,
-- 
Dan