Re: garbled output, error

[Urs Liska]

Am 27.06.2015 um 11:34 schrieb David Kastrup:
> Urs Liska <address@hidden> writes:
> 
>> Am 27.06.2015 um 11:04 schrieb N. Andrew Walsh:
>>> Well, I've now loaded this up on my laptop, and re-cloned the
>>> openlilylib directory, and now the problem no longer appears. I have
>>> other issues -- namely that my production box is running lilypond
>>> 2.19.22, which I can't seem to install from ubuntu -- but those are minor.
>>
>> 2.19.22 isn't released yet AFAICS. Do you have a custom built LilyPond?
>>
>>>
>>> Urs, did you update something on the git tree, or was this simply a
>>> problem on my side that I'll run into again when I get back home?
>>
>> It's not clear what you say. Do you have a laptop with something <
>> 2.19.22 and a home pc with a custom build from the 2.19.22 line?
>>
>> If that's the case then probably you'll experience the problem at home
>> again because the changes in LilyPond have been introduced there.
>>
>> I assume I'll have to change the implementation (and look for other
>> parts where this might matter) and introduce a version switch so the
>> code executed from openLilyLib depends on the LilyPond version.
> 
> Unfortunately, 2.19.22 has not yet been released so there is "2.19.22
> before change" and "2.19.22 after change" rather than "2.19.22 at the
> point of its release".
> 
> It might be more reliable to use something like
> 
> (catch 'wrong-number-of-args
>   (lambda () (something bla))
>   (lambda _ (something parser bla)))
> 
> which just tries the new calling convention and, failing that, reverts
> to the old call.

That's surely cleaner. But I'll stick to my simpler approach because if
I add this switch which will execute the new code when version >=
2.19.22 this will only fail for someone who has a custom build from not
the latest master. *If* someone should run in that corner case and
really needs to use that functionality that person could afford pulling
the latest master and building LilyPond IMO

Urs