Re: guile 2.0?

[David Kastrup]

Werner LEMBERG <address@hidden> writes:

> I'm curious: Is there anything new w.r.t. guile 2.0?  We had a lot of
> talk, but recently there was only silence...  In particular, is
> version 2.0.11 good enough for lilypond?

I differ with "we had a lot of talk".  There was a lot of talk about how
important it would be if somebody bothered to do stuff.  I got rid of
all the compilation errors and load order problems.  All of that is in
master.  There is still some encoding problem with the PostScript
output, and there is a dev/guilev2 branch where this encoding problem is
absent at the cost of rather invasive changes that don't allow for
compilation with GUILEv1 any more (all of the changes in master require
--enable-guile2 to compile with GUILE2).

Either way, the binary crashes frequently.  Garbage collection appears
to be rather hosed, particularly egregiously regarding the actions of
remove-first and remove-empty flags (namely, explicit suicide of
systems).

In spite of all the talk (including those from GUILE maintainers), there
is no indication that anybody but myself ever bothered actually
compiling the GUILEv2 version or contributing to debugging/redesigning
any code at all.

A reasonably simple task, for example, would be figuring out just _what_
change in the dev/guilev2 branch is responsible for fixing the
PostScript generation problem.  That's rather tedious work akin to
bijection (just manual, and with code).  The result would be that a
version compiled using GUILE2.0 would be generating proper PostScript
and one could focus better on fixing the crashes.

However, since debugging the crashes does not really depend on the
PostScript encoding problem, and since nobody could be bothered
investing any work in looking at the GUILEv2 version, let alone making
progress on the crashes, this seems like a pretty academical feat.

Personally, I have a rather low frustration tolerance, and saying about
half a year "haven't you fixed that on your own yet?" is not likely to
lead to fabulous results.

-- 
David Kastrup