lilypond-devel
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Fwd: Binaries of LilyPond 2.23.5 with Guile 2.2


From: Michael Käppler
Subject: Re: Fwd: Binaries of LilyPond 2.23.5 with Guile 2.2
Date: Fri, 17 Dec 2021 11:00:56 +0100
User-agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:91.0) Gecko/20100101 Thunderbird/91.4.0

Am 15.12.2021 um 14:38 schrieb Michael Käppler:
Am 15.12.2021 um 14:25 schrieb Jonas Hahnfeld:
Am Montag, dem 13.12.2021 um 22:09 +0100 schrieb Jonas Hahnfeld:
Am Mittwoch, dem 08.12.2021 um 21:51 +0100 schrieb Jonas Hahnfeld via
Discussions on LilyPond development:
If that still doesn't work, there are two more possibilities:
1. We're missing an option in building some of the dependencies that I
had figured out for the proof-of-concept. I payed attention to a few
important ones, but I'm going to recheck.
To give a quick update on this, I think I know why it works for me but
for almost nobody else: I have my Windows configured with an English
language pack, but German regional settings (Monday is the first day of
the week! And you write dates as DD.MM.YYYY, not some other weird
order!). I suspect this is sufficiently unusual that libraries don't
load translated message strings. As soon as I switch to full English or
full German, it also breaks for me.

The Python scripts build gettext / libintl and libiconv differently
(more correctly) than the shell scripts did, so this might explain why
the older binaries worked. I'll try to figure out how I need to change
the process to make it work.
I think I got it, can you give the newly built binaries in
https://cloud.hahnjo.de/index.php/s/3gpCEQnjZZAT8FJ a try? This uses
https://gitlab.com/lilypond/lilypond/-/merge_requests/1061 and works on
my system with a more common locale configuration, including properly
showing translated messages.
Works nicely in a short test. Thanks, Jonas!
After working a couple of hours with this build, I noticed that it
segfaults from time to time
during bootup, i.e. before actually parsing the file. For some reason I
cannot reproduce it currently
even if running x times in a row.

Since it happens before processing the file it might be unrelated to the
recent GC crashes, not sure, however.

Jonas, could you build the binaries again without stripping debugging
symbols? And maybe one edition with
'--disable-optimisation' in addition to that?

Michael



Thanks!
Jonas






reply via email to

[Prev in Thread] Current Thread [Next in Thread]