Re: maintaining advanced power-user Scheme functions

[Jan-Peter Voigt]

Hi there,

well, I didn't read all messages of this thread, but I see two points of 
discussion:
1. we already have LSR
2. who likes to use git?

1.: Well, we have LSR, but the current implementation needs a 
tomcat-server with mysql. I did use tomcat sometimes and created some 
webapps, but as a born pacifist I really don't like to call files *.WAR 
;) ok, no more joking!
This part is maintained by one(?) person and my impression is, that its 
not that easy to get in touch?
I know, every lily user can start the update, but wouldn't it be easier, 
if one, submitting a snippet, is upgrading it to the current stable version?
hm, but who is contacting whom to kick off the upgrade?
IMO we have something, but it would be very nice to optimize it.

2.: a) I like git ;)
b) thinking of a web front end, which uses GIT as a storage backend is 
(IMHO) not that bad.

3.: (a third point of disussion? Well, I just had a thought, which does 
not exactly fit the two former points ...)
I was thinking of building yet another snippet repository (YASR ;)
* The repository shall be searchable - like the current LSR
* The snippets shall be tagged, for which lily-version they work!
If one finds a snippet, not running in the lily-version he is actually 
using, he is able to post an updated version.
* If there is a scheme-snippet, it might not be useful, to produce any 
PDF/PNG but maybe text

But to create this, a lot of things have to work:
* Website with Database (searching) - this is made up easy, but can mean 
a lot of work
* ... and different running versions of lilypond - this propably forces 
a dedicated server, which means more administration

Just some thoughts

Cheers, Jan-Peter




Am 14.08.2013 14:34, schrieb Janek Warchoł:
> Harm and David N. (and some other people) write lots of very advanced
> (and very helpful!) Scheme functions.  These funcitons are improved
> over time, and there is a problem related to that: it's easy to get
> lost in all the email threads about them, and it's not always obvious
> where the most recent version is.
>
> I think that such functions should be tracked by version control, and
> i see two approaches:
> - include them in official LilyPond source as soon as they are
> created.  Upside: there's bigger chance that they will be updated when
> there are some changes. Downside: one has to write documentation and
> go through official patch submitting channels.
> - use another repository.  What about OpenLilyLib?http://www.openlilylib.org/
>
> At any rate, we better do something about it - at current state, all
> these funcitons can easily become lost or forgotten, and that would be
> a grave loss.
+1
> Thoughts?