Re: [Lilypond-auto] Issue 3194 in lilypond: crash when too many bars requested

[lilypond]

Comment #6 on issue 3194 by address@hidden: crash when too many  
bars requested
http://code.google.com/p/lilypond/issues/detail?id=3194

The failure occurs whenever the working-set approaches the available  
physical memory.  On the system I used at the time I ran these tests (2GB  
laptop running Windows Vista SP2) the available physical memory was c.  
1.3GB.

For the tests I used \repeat unfold 1000 { c'1 c  \break }

Starting with 2.17.6 the failure occured quite quickly  
during "Preprocessing graphical objects".  The working-set size hit 1.3GB  
from its initial value of around 130MB during a very rapid rise.

I then switched to 2.13.10.  This compiled the test with a rep count of 700  
successfully but also failed with 1000 when the working-set reached 1.3GB,  
but around 30 mins later, during "Drawing systems...".  The profile of the  
working set over time looks something like this:

                                        -- crash here
    |-----|                  |---------|
    |     |        |---------
    |     |--------
----
---------------------------------------------------

I then tried 2.13.14, which behaved the same way, but crashed just a little  
earlier.  2.13.23 crashed quickly, on the first rapid rise in the  
working-set, during "Solving 1 page-breaking chunks..."

My conclusion is that we are not seeking a specific bug as the cause of  
this failure, but a gradual increase in the working set size as the code  
has been developed since 2.12.  I fear this is a fundamental problem in the  
design rather than a specific bug, but I hope I am wrong.



--
You received this message because this project is configured to send all  
issue notifications to this address.
You may adjust your notification preferences at:
https://code.google.com/hosting/settings