Re: Lilypond crash

[Phil Holmes]

----- Original Message ----- 
From: "Valentin Villenave" <address@hidden>

> How interesting. Of course, LilyPond-win being a 32-bit app, it
> shouldn't be able to use more than roughly 3.6 GB anyway. (Or is it
> just up to the OS?)

You're probably right.  In practice, it never gets anywhere near that.

> As for the Win vs Linux inequality, that's something I noticed a long
> time ago: for instance, at some point I couldn't compile my opera *at
> all* on Windows whereas it only took 12 minutes on GNU/Linux x86_64,
> and about 20 minutes with GNU/Linux x86. Similarly, I couldn't compile
> Nicolas' scores with less than 2 GB RAM, whilst with 4 GB it worked
> fine.

I'd guess this is more to do with the memory allocation mechanisms of the 
compilers than the OSes themselves.  Windoze gets a bad press, but the later 
versions are pretty stabler, TBH.  I rarely reboot my PC (use standby) and 
it's generally as solid as a rock.

> What concerns me more is that according to your data, LilyPond's
> overall capacity might be dropping between the last stable version.

Not significantly - it's pretty marginal.

> Cheers,
> Valentin

The other issue is - all these test cases with multiple repeats are 
artificial - the real question is - does it do long _real_ scores?

--
Phil Holmes