gub progress

classic Classic list List threaded Threaded
11 messages Options
Reply | Threaded
Open this post in threaded view
|

gub progress

Erik Sandberg-2
Hi,

Now gub builds correctly. However, when I invoke lilypond
(target/linux/system/usr/bin/lilypond), it consistently fails when converting
to pdf.

Additional diagnostics:
$ target/linux/system/usr/bin/gs
GPL Ghostscript 8.50: Can't find initialization file gs_init.ps.
$ find target/linux/system -name gs_init.ps
target/linux/system/usr/share/ghostscript/8.50/lib/gs_init.ps
$

Ideas?

Also, what's the next step once lilypond-HEAD works nicely? I suppose the
first step is to make a local mirror of the lily cvs repository; in that
case, any suggestions for a good choice of version control system? I'd like
something that
- can merge changes from and to the official lily cvs repository nicely and
easily
- presumably works similarly well with the revision control system we will use
after we have ditched cvs.

--
Erik


_______________________________________________
lilypond-devel mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/lilypond-devel
Reply | Threaded
Open this post in threaded view
|

Re: gub progress

Christian Hitz

Am 03.06.2006 um 18:51 schrieb Erik Sandberg:

> Additional diagnostics:
> $ target/linux/system/usr/bin/gs
> GPL Ghostscript 8.50: Can't find initialization file gs_init.ps.
> $ find target/linux/system -name gs_init.ps
> target/linux/system/usr/share/ghostscript/8.50/lib/gs_init.ps
> $
>
> Ideas?

To circumvent the problem set the GS_LIB variable to 'target/linux/
system/usr/share/ghostscript/8.50/lib'

I had to set this variable to use the lilypond supplied gs with  
lilypond-book on MacOSX.

Christian



_______________________________________________
lilypond-devel mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/lilypond-devel
Reply | Threaded
Open this post in threaded view
|

Re: gub progress

Erik Sandberg-2
In reply to this post by Erik Sandberg-2
On Saturday 03 June 2006 22:14, Han-Wen Nienhuys wrote:

> 2006/6/3, Erik Sandberg <[hidden email]>:
> > Now gub builds correctly. However, when I invoke lilypond
> > (target/linux/system/usr/bin/lilypond), it consistently fails when
> > converting to pdf.
> >
> > Additional diagnostics:
> > $ target/linux/system/usr/bin/gs
> > GPL Ghostscript 8.50: Can't find initialization file gs_init.ps.
> > $ find target/linux/system -name gs_init.ps
> > target/linux/system/usr/share/ghostscript/8.50/lib/gs_init.ps
>
> I'm not sure if this is supposed to work; I always run from the installer
> dir,
>
>   target/PLATFORM/installer-BRANCH/usr/bin/lilypond

Doesn't help.

I have GS_LIB=
"/home/erik/lily/gub/gub/target/linux/system/usr/bin/../share/ghostscript/8.50/lib:
/home/erik/lily/gub/gub/target/linux/system/usr/bin/../share/ghostscript/8.50/Resource"

and usr/bin/gs (with any parameters given) just dies silently with a nonzero
exit code. Same thing with s/system/installer-HEAD/g.

--
Erik


_______________________________________________
lilypond-devel mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/lilypond-devel
Reply | Threaded
Open this post in threaded view
|

Re: gub progress

Johannes Schindelin
In reply to this post by Erik Sandberg-2
Hi,

On Sat, 3 Jun 2006, Erik Sandberg wrote:

> Also, what's the next step once lilypond-HEAD works nicely? I suppose the
> first step is to make a local mirror of the lily cvs repository; in that
> case, any suggestions for a good choice of version control system?

I already proposed git, but you can basically pick your choice out of
several programs: AFAIK git, cogito, mercurial, bzr, baz, darcs are
equally feature-wise. If you insist on a single respository, you can also
take Subversion.

My favourite is git, evidently, mainly because it is blazingly fast, and
has good integration with cvs.

> I'd like something that
> - can merge changes from and to the official lily cvs repository nicely and
> easily

from: git-cvsimport. to: git-cvsexportcommit.

> - presumably works similarly well with the revision control system we
> will use after we have ditched cvs.

I'd be very careful ditching cvs. Most open source developers know cvs,
and it is the greatest common denominator.

Also, if you want to ditch cvs, you always leave somebody behind. For
example, from what I understand, Han-Wen likes darcs very much... and I
know plenty people (me included) who don't like it. Same goes for every
version control system, but at least people are _used_ to cvs.

Ciao,
Dscho



_______________________________________________
lilypond-devel mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/lilypond-devel
Reply | Threaded
Open this post in threaded view
|

Re: gub progress

Pedro Kröger
Johannes Schindelin <[hidden email]> writes:

> Also, if you want to ditch cvs, you always leave somebody behind. For
> example, from what I understand, Han-Wen likes darcs very much... and I
> know plenty people (me included) who don't like it. Same goes for every
> version control system, but at least people are _used_ to cvs.

Could you tell why you don't like darcs? I'm just curious to hear. I've
been using darcs for some time but I'm considering to try git as well.

Pedro


_______________________________________________
lilypond-devel mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/lilypond-devel
Reply | Threaded
Open this post in threaded view
|

Re: gub progress

Johannes Schindelin
Hi,

On Sun, 4 Jun 2006, Pedro Kröger wrote:

> Johannes Schindelin <[hidden email]> writes:
>
> > Also, if you want to ditch cvs, you always leave somebody behind. For
> > example, from what I understand, Han-Wen likes darcs very much... and I
> > know plenty people (me included) who don't like it. Same goes for every
> > version control system, but at least people are _used_ to cvs.
>
> Could you tell why you don't like darcs? I'm just curious to hear. I've
> been using darcs for some time but I'm considering to try git as well.

<disclaimer>I am not trying to offend anybody. Just stating my opinions
here, so no need to get mad at me, although you are free to call me an
idiot.</disclaimer

I don't like darcs for several reasons:

- I need to install a completely new compiler to patch darcs (Haskell
does not bother me as much; I could learn it).

- the whole bruhaha about the "theory of patches" and that it is connected
with some concepts from quantum mechanics is _soooo_ seagull consultant.
It has written BS all over it.

- It is _slow_. I had the pleasure of having to work with tailor, which is
a program to convert between different Version Control formats, and which
is written in darcs. My experience is that git beats it any time of the
day.

- It is buggy. When trying to convert the repository of tailor itself with
tailor into git format, I hit a bug in darcs early on.

- It does not seem to focus on version control, but on changing the order
of patches. Thus, it is more similar to quilt than to a version control
system.

- It lacks a nice GUI where I can see what happened in a particular
branch. In fact, I am not quite sure that you actually get the whole
history, what with all the patch reordering.

- The repository format itself is fragile, just as cvs'. It is based on
patches, so if there is one single patch corrupt, you loose the history
from then on.

- AFAIK darcs has not even started to handle complicated merges
gracefully. In contrast, git handles modifications in one branch, and a
rename in the other quite well.

- git makes it _really_ easy to access older version without checking them
out. For example, I can get the differences in lily/music.cc between
version 1.3.108 and 2.9.7 with "git diff lilypond_1_3_108:lily/music.cc  
lilypond_2_9_7:lily/music.cc".

If I thought long and hard, I could probably come up with more reasons I
don't like darcs, but I think you got an idea.

BTW, the git repository I set up, tracking lilypond's cvs, is still
online. Containing all history of the cvs repository, the first download
is about 55MB. Feel free to play with it.

Ciao,
Dscho

_______________________________________________
lilypond-devel mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/lilypond-devel
Reply | Threaded
Open this post in threaded view
|

Re: gub progress

Pedro Kröger
Johannes Schindelin <[hidden email]> writes:

> although you are free to call me an idiot.

no need to do that :-)

> - It does not seem to focus on version control, but on changing the
> order of patches.

this is kind of true (I don't know enought about darcs to say it's
totaly true)

> - It lacks a nice GUI where I can see what happened in a particular
> branch.

me too

> In fact, I am not quite sure that you actually get the whole history,
> what with all the patch reordering.

this is true. the darcs documentation even mention it someplace.

> - The repository format itself is fragile, just as cvs'. It is based on
> patches, so if there is one single patch corrupt, you loose the history
> from then on.

hum, I haven't tought of that.

> - git makes it _really_ easy to access older version without checking them
> out. For example, I can get the differences in lily/music.cc between
> version 1.3.108 and 2.9.7 with "git diff lilypond_1_3_108:lily/music.cc  
> lilypond_2_9_7:lily/music.cc".

this is something that's a little bit annoing about darcs.

Thanks for your email. I think I'll give git a try.

Pedro


_______________________________________________
lilypond-devel mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/lilypond-devel
Reply | Threaded
Open this post in threaded view
|

Re: gub progress

Han-Wen Nienhuys-2
In reply to this post by Erik Sandberg-2
Erik Sandberg schreef:

>>> Additional diagnostics:
>>> $ target/linux/system/usr/bin/gs
>>> GPL Ghostscript 8.50: Can't find initialization file gs_init.ps.
>>> $ find target/linux/system -name gs_init.ps
>>> target/linux/system/usr/share/ghostscript/8.50/lib/gs_init.ps
>> I'm not sure if this is supposed to work; I always run from the installer
>> dir,
>>
>>   target/PLATFORM/installer-BRANCH/usr/bin/lilypond
>
> Doesn't help.
>
> I have GS_LIB=
> "/home/erik/lily/gub/gub/target/linux/system/usr/bin/../share/ghostscript/8.50/lib:
> /home/erik/lily/gub/gub/target/linux/system/usr/bin/../share/ghostscript/8.50/Resource"

that's really strange. Can you investigate further? Maybe you can use
strace to see what it is looking for,

FWIW,

lilydev@muurbloem:~/vc/gub-split$
GS_LIB=/home/lilydev/vc/gub-split/target/linux/system/usr/share/ghostscript/8.50.lib/
target/linux/system/usr/bin/gs

GPL Ghostscript 8.50 (2005-12-31)
Copyright (C) 2005 artofcode LLC, Benicia, CA.  All rights reserved.
This software comes with NO WARRANTY: see the file PUBLIC for details.
GS>

> and usr/bin/gs (with any parameters given) just dies silently with a nonzero
> exit code. Same thing with s/system/installer-HEAD/g.


--

Han-Wen Nienhuys - [hidden email] - http://www.xs4all.nl/~hanwen

LilyPond Software Design
  -- Code for Music Notation
http://www.lilypond-design.com



_______________________________________________
lilypond-devel mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/lilypond-devel
Reply | Threaded
Open this post in threaded view
|

Re: gub progress

Johannes Schindelin
In reply to this post by Pedro Kröger
Hi,

On Sun, 4 Jun 2006, Pedro Kröger wrote:

> Johannes Schindelin <[hidden email]> writes:
>
> > although you are free to call me an idiot.
>
> no need to do that :-)

;-)

> > - The repository format itself is fragile, just as cvs'. It is based on
> > patches, so if there is one single patch corrupt, you loose the history
> > from then on.
>
> hum, I haven't tought of that.

I actually had a fsck-up, because I accidentally issued a "rm -rf ." in
one of my private projects. I was not really fast enough with Ctrl-C,
but git could still access most of the revisions.

> > - git makes it _really_ easy to access older version without checking them
> > out. For example, I can get the differences in lily/music.cc between
> > version 1.3.108 and 2.9.7 with "git diff lilypond_1_3_108:lily/music.cc  
> > lilypond_2_9_7:lily/music.cc".
>
> this is something that's a little bit annoing about darcs.
>
> Thanks for your email. I think I'll give git a try.

Please do. And if you run into problems, do not hesitate to ask me, or the
list. There are even some people on irc...

Ciao,
Dscho

_______________________________________________
lilypond-devel mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/lilypond-devel
Reply | Threaded
Open this post in threaded view
|

Re: gub progress

Erik Sandberg-2
In reply to this post by Han-Wen Nienhuys-2
On Monday 05 June 2006 01:48, Han-Wen Nienhuys wrote:

> Erik Sandberg schreef:
> > I have GS_LIB=
> > "/home/erik/lily/gub/gub/target/linux/system/usr/bin/../share/ghostscript
> >/8.50/lib:
> > /home/erik/lily/gub/gub/target/linux/system/usr/bin/../share/ghostscript/
> >8.50/Resource"
>
> that's really strange. Can you investigate further? Maybe you can use
> strace to see what it is looking for,
>
> FWIW,

Looks like it's crashing while reading the second 4096-byte block from
gs_init.ps. That file is identical to the working copy that my gs-gpl-8.50
from debian/testing uses.

I'm still rather clueless. strace output:

GS_LIB="/home/erik/lily/gub/gub/target/linux/system/usr/bin/../share/ghostscript/8.50/lib:/home/erik/lily/gub/gub/target/linux/system/usr/bin/../share/ghostscript/8.50/Resource"
strace usr/bin/gs
[...]
gettimeofday({2744403647623793, 586627592864653996}, {4294967176,
2744403647623793}) = 0
open("gs_init.ps", O_RDONLY)            = -1 ENOENT (No such file or
directory)
open("./gs_init.ps", O_RDONLY)          = -1 ENOENT (No such file or
directory)
open("/home/erik/lily/gub/gub/target/linux/system/usr/share/ghostscript/8.50/lib/gs_init.ps",
O_RDONLY) = 3
fstat64(0x3, 0xffffaf5c)                = 0
fstat64(0x3, 0xffffae08)                = 0
mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0x1000)
= 0x555c7000
_llseek(3, 0, [0], SEEK_CUR)            = 0
_llseek(3, 0, [0], SEEK_SET)            = 0
read(3, "% Copyright (C) 1989-2004 artofc"..., 4096) = 4096
read(3, "t /DISKFONTS known   /DISKFONTS "..., 4096) = 4096
close(3)                                = 0
munmap(0x555c7000, 4096)                = 0
exit_group(255)                         = ?


Debian's gs-gpl has an alternative happy ending:
[...]
lseek(3, 0, SEEK_CUR)                   = 0
lseek(3, 0, SEEK_SET)                   = 0
read(3, "% Copyright (C) 1989-2004 artofc"..., 4096) = 4096
read(3, "t /DISKFONTS known   /DISKFONTS "..., 4096) = 4096
read(3, " mod (-)\n   revisiondate 10000 i"..., 4096) = 4096
fstat(1, {st_mode=S_IFIFO|0600, st_size=0, ...}) = 0
mmap(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0) =
0x2aaaaaac5000
write(1, "GPL Ghostscript 8.50 (2005-12-31"..., 128GPL Ghostscript 8.50
(2005-12-31)
Copyright (C) 2005 artofcode LLC, Benicia, CA.  All rights reserved.
This software comes with ) = 128


--
Erik


_______________________________________________
lilypond-devel mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/lilypond-devel
Reply | Threaded
Open this post in threaded view
|

Re: gub progress

Han-Wen Nienhuys-2
Erik Sandberg schreef:

> On Monday 05 June 2006 01:48, Han-Wen Nienhuys wrote:
>> Erik Sandberg schreef:
>>> I have GS_LIB=
>>> "/home/erik/lily/gub/gub/target/linux/system/usr/bin/../share/ghostscript
>>> /8.50/lib:
>>> /home/erik/lily/gub/gub/target/linux/system/usr/bin/../share/ghostscript/
>>> 8.50/Resource"
>> that's really strange. Can you investigate further? Maybe you can use
>> strace to see what it is looking for,
>>
>> FWIW,
>
> Looks like it's crashing while reading the second 4096-byte block from
> gs_init.ps. That file is identical to the working copy that my gs-gpl-8.50
> from debian/testing uses.
>
> I'm still rather clueless. strace output:
>
> GS_LIB="/home/erik/lily/gub/gub/target/linux/system/usr/bin/../share/ghostscript/8.50/lib:/home/erik/lily/gub/gub/target/linux/system/usr/bin/../share/ghostscript/8.50/Resource"
> strace usr/bin/gs
> [...]
> gettimeofday({2744403647623793, 586627592864653996}, {4294967176,
> 2744403647623793}) = 0
> open("gs_init.ps", O_RDONLY)            = -1 ENOENT (No such file or
> directory)
> open("./gs_init.ps", O_RDONLY)          = -1 ENOENT (No such file or
> directory)
> open("/home/erik/lily/gub/gub/target/linux/system/usr/share/ghostscript/8.50/lib/gs_init.ps",
> O_RDONLY) = 3
> fstat64(0x3, 0xffffaf5c)                = 0
> fstat64(0x3, 0xffffae08)                = 0
> mmap2(NULL, 4096, PROT_READ|PROT_WRITE, MAP_PRIVATE|MAP_ANONYMOUS, -1, 0x1000)
> = 0x555c7000
> _llseek(3, 0, [0], SEEK_CUR)            = 0
> _llseek(3, 0, [0], SEEK_SET)            = 0
> read(3, "% Copyright (C) 1989-2004 artofc"..., 4096) = 4096
> read(3, "t /DISKFONTS known   /DISKFONTS "..., 4096) = 4096
> close(3)                                = 0
> munmap(0x555c7000, 4096)                = 0
> exit_group(255)                         = ?

and from GDB ?  (the binaries include debugging info, so if you unpack
the GS sources in target/linux/src/ , you get full-fledged debugging info).

--

Han-Wen Nienhuys - [hidden email] - http://www.xs4all.nl/~hanwen

LilyPond Software Design
  -- Code for Music Notation
http://www.lilypond-design.com



_______________________________________________
lilypond-devel mailing list
[hidden email]
http://lists.gnu.org/mailman/listinfo/lilypond-devel