Re: [Spam] request for help: huge score

[David Kastrup]

Thomas Morley <address@hidden> writes:

> 2016-11-23 12:33 GMT+01:00 Rutger Hofman <address@hidden>:
>> You can check if my typeset of Berg's Adagio (Kammerkonzert) works. It is
>> here: http://imslp.org/wiki/Kammerkonzert_(Berg,_Alban) then under the list
>> of parts, there is a zip with the source, with a Makefile to build score and
>> all parts (make -j <num-cores> should work). (This is also a nice test if I
>> really bundled everything :-) ) I think 2.19.48 or so.
>>
>> If you like, I can also submit my ongoing work on the Rondo from the same
>> piece. It has a solo piano, and Berg uses all the typesetting tricks in the
>> book. Even so, it is *far* from finished, but it compiles for me.
>>
>> Rutger
>
>
> Hi Rutger,
>
> many thanks for your score.
> A really nice work!
>
> Because the Makefile needs to be changed in order to have my various
> LilyPond-versions work with it, it's a little uncomfortable for me.
>
> But here the results:
> with 2.19.51 (from the installer) compiled in ~5 minutes
> with 2.19.52-guilev2 compiled in ~17 minutes
>
> Some warnings with bot versions.
>
> Maybe of interest for the current guile2-migration, I found a guile-message:
> "
> Omitting the destination on a call to format is deprecated.
> Pass #f as the destination, before the format string.
> "
> I wasn't aware of this guile change.

It's not new.  Quite the opposite:

address@hidden:~$ guile-1.8 
guile> (format "~s" "x")

Backtrace:
In current input:
   1: 0* [simple-format {"~s"} "x"]

<unnamed port>:1:1: In procedure simple-format in expression (format "~s" "x"):
<unnamed port>:1:1: Wrong type argument in position 1: "~s"
ABORT: (wrong-type-arg)
guile> address@hidden:~$ guile-2.0 
GNU Guile 2.0.11
Copyright (C) 1995-2014 Free Software Foundation, Inc.

Guile comes with ABSOLUTELY NO WARRANTY; for details type `,show w'.
This program is free software, and you are welcome to redistribute it
under certain conditions; type `,show c' for details.

Enter `,help' for help.
scheme@(guile-user)> (format "~s" "x")
;;; <stdin>:1:0: warning: "~s": wrong port argument
$1 = "\"x\""
scheme@(guile-user)> 

But LilyPond _overrides_ the default format definition to get the old
behavior which did not complain.

Your error message is from some location where this override stops
working under Guile2.

So you are seeing a module system outlier here rather than a semantics
change (the semantics change happened way before Guile 1.8.8).

There might have been some patches trying to use an explicit #f
everywhere but I don't even remember how many of them got committed
since it seemed more important to fix the problem of different versions
of format getting used.

-- 
David Kastrup