I got this reply off-list, so I'll put it back on-list :)
Begin forwarded message:
> From: Bec and John Silva <[hidden email]>
> Date: 28 augustus 2005 3:27:16 GMT+02:00
> To: Maurits Lamers <[hidden email]>
> Cc: Han-Wen Nienhuys <[hidden email]>
> Subject: Re: lily rO><Or!
> Hi Maurits,
> I have nothing to do with the development of Lilypond and I have done
> very little more than a super-quick scan through a few of the source
> files, so what I say is a complete guess.
> I suspect that Lilypond could be made to layout a single measure if
> requested, but I think one of the main advantages of Lilypond's output
> is that it takes the entire score into consideration when calculating
> the layout. If a feature was added which could isolate calculations to
> just differences in the score, a user would run the risk of forgetting
> to recalculate the entire score for optimum layout (and personally,
> I'm always against letting users mess up their own work).
> I know the time to calculate can be a bit long for this computer-age
> of "it must happen right now!", but I've found that it causes me to do
> more work between previews. I've also found that the more proficient I
> get at Lilypond the less bothered I am with non-instantaneous
> Just my thoughts.
> - John
>> I was wondering...
>> if you cannot optimize the code itself very much. maybe it is
>> possible to improve the way Lilypond parses the ly file.
>> When I am editing and tweaking a file, I need to change one little
>> thing and recompile it.
>> As far as I can see it, lilypond then always runs the complete file
>> from scratch...
>> Maybe it is possible to provide lilypond with a kind of diff function
>> that enables it to determine the changes and change only what is
>> necessary ?
>> (by saving states for example...) Or is this impossible ?
>> If this could work it could mean quite a speed improvement...
>> On 24-aug-05, at 22:00, Han-Wen Nienhuys wrote:
>>> Nicolas Sceaux wrote:
>>>> Wahou. This used to be around 40 minutes, in an slightly older
>>>> computer, with multiple lilypond invocations.
>>>> Lily is getting so cool. Kuddos to Han-Wen and Jan!
>>> the downside to this is that there is little left to optimize, I
>>> still think Lily is a bit slow, but with painstaking optimization,
>>> I've only been been able to get some 10 to 30 % during the 2.7
>>> Han-Wen Nienhuys - [hidden email] - http://www.xs4all.nl/~hanwen >>>
>>> lilypond-user mailing list
>>> [hidden email] >>> http://lists.gnu.org/mailman/listinfo/lilypond-user >>>
>> lilypond-user mailing list
>> [hidden email] >> http://lists.gnu.org/mailman/listinfo/lilypond-user >>