unofficial mirror of guile-devel@gnu.org 
 help / color / mirror / Atom feed
* gettext-0.14.2 is released
@ 2005-02-25 12:31 Bruno Haible
  2005-02-25 23:14 ` Kevin Ryde
  0 siblings, 1 reply; 7+ messages in thread
From: Bruno Haible @ 2005-02-25 12:31 UTC (permalink / raw)


Hi,

http://ftp.gnu.org/gnu/gettext/gettext-0.14.2.tar.gz
contains Scheme support:
  - xgettext supports Scheme.
  - messages can be marked as "scheme-format", denoting 'format' strings.
  - There is an example (examples/hello-guile) demonstrating the use of
    gettext with guile. Best used with guile 1.7 or newer. It is a good idea
    to test it with the current guile CVS.

Bruno



_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-devel


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: gettext-0.14.2 is released
  2005-02-25 12:31 gettext-0.14.2 is released Bruno Haible
@ 2005-02-25 23:14 ` Kevin Ryde
  2005-02-26 16:52   ` Bruno Haible
  2005-03-15 14:18   ` Bruno Haible
  0 siblings, 2 replies; 7+ messages in thread
From: Kevin Ryde @ 2005-02-25 23:14 UTC (permalink / raw)
  Cc: guile-devel

Bruno Haible <bruno@clisp.org> writes:
>
>   - messages can be marked as "scheme-format", denoting 'format' strings.

I noticed a couple of forms that provoke errors,

	(format #f (gettext "~/"))
	(format #f (gettext "~q"))
	(format #f (gettext "~8,2i") 1+2i)
	(format #f (gettext "~65c"))
	(format #f (gettext "~4,3,'-t"))
	(format #f (gettext "~k") "fmtstr")

~q is a copyright message
~/ is a tab character (like ~| etc)
~i is complex number fixed-point (up to 5 params)
~c accepts a charnum param
~t accepts a pad char param
~k is an alias for ~?

Nosing around format-scheme.c (if that's the right code) I noticed ~!
described as a clisp call, in format.scm ~! is an output flush (no
parameters or args), though it doesn't seem to cause an error.

I added a bit to "Formatted Output" in the guile cvs reference manual
based on my understanding of what format.scm is supposed to do.
There's only a few differences from CL.


_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-devel


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: gettext-0.14.2 is released
  2005-02-25 23:14 ` Kevin Ryde
@ 2005-02-26 16:52   ` Bruno Haible
  2005-02-27 22:06     ` Kevin Ryde
  2005-03-15 14:18   ` Bruno Haible
  1 sibling, 1 reply; 7+ messages in thread
From: Bruno Haible @ 2005-02-26 16:52 UTC (permalink / raw)
  Cc: guile-devel

Kevin Ryde wrote:
> >   - messages can be marked as "scheme-format", denoting 'format' strings.
>
> I noticed a couple of forms that provoke errors,

Thanks for reporting this. I was confused through the situation in slib
and through the documentations: In SLIB, when I saw that Aubrey Jaffer
had deprecated FORMAT, I told him that FORMAT is essential for i18n (because
the SLIB PRINTF doesn't support reordering of arguments) and explained him
that some elements of his SLIB FAQ about FORMAT are simply wrong expectations.
About the documentation, I found that the older SLIB doc describes it as
  "An almost complete implementation of Common LISP format description
   according to the CL reference book @cite{Common LISP} from Guy L. Steele"
so I referred to that.

Now I see your message in
  http://lists.gnu.org/archive/html/guile-devel/2004-07/msg00044.html
Is it correct to assume that

  - For guile, the code and doc regarding FORMAT in SLIB are not relevant
    any more?

  - The guile FORMAT attempts to be an extension of Common Lisp FORMAT,
    i.e. that all CL format directives are supported without changes
    (except when documented otherwise, such as ~_)?

  - Differences in behaviour between a valid CL format string and the
    same string in guile (except when documented otherwise) are bugs in
    format.scm that I should report?

  - The reference to which I can point translators is the node
    "Formatted Output" in the guile documentation?

> Nosing around format-scheme.c (if that's the right code) I noticed ~!
> described as a clisp call

Oops, that was a copy&paste artefact. I've now removed that.

> 	(format #f (gettext "~/"))
> 	(format #f (gettext "~q"))
> 	(format #f (gettext "~8,2i") 1+2i)
> 	(format #f (gettext "~65c"))
> 	(format #f (gettext "~4,3,'-t"))
> 	(format #f (gettext "~k") "fmtstr")
>
> ~q is a copyright message
> ~/ is a tab character (like ~| etc)
> ~i is complex number fixed-point (up to 5 params)
> ~c accepts a charnum param
> ~t accepts a pad char param
> ~k is an alias for ~?

OK, will be corrected in gettext.

Bruno



_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-devel


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: gettext-0.14.2 is released
  2005-02-26 16:52   ` Bruno Haible
@ 2005-02-27 22:06     ` Kevin Ryde
  2005-02-28 21:29       ` Bruno Haible
  0 siblings, 1 reply; 7+ messages in thread
From: Kevin Ryde @ 2005-02-27 22:06 UTC (permalink / raw)
  Cc: guile-devel

Bruno Haible <bruno@clisp.org> writes:
>
>   - For guile, the code and doc regarding FORMAT in SLIB are not relevant
>     any more?

Oh, well, if it's withdrawn from slib then there's nothing for guile
to stay compatible with.

>   - The guile FORMAT attempts to be an extension of Common Lisp FORMAT,
>     i.e. that all CL format directives are supported without changes
>     (except when documented otherwise, such as ~_)?

Yes, basically, I think.  There's some things missing though, like ~< ~>.

>   - Differences in behaviour between a valid CL format string and the
>     same string in guile (except when documented otherwise) are bugs in
>     format.scm that I should report?

Yep.  I think there's some doubtful treatment of digits in some of the
floating point output, that's the sort of thing that really ought to
be the same as CL.

>   - The reference to which I can point translators is the node
>     "Formatted Output" in the guile documentation?

For guile, yes.  What other schemes are doing is another matter.

srfi-28 and srfi-48 specify versions of format too.  In the latter ~t
is different, and ~h is an addition.


_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-devel


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: gettext-0.14.2 is released
  2005-02-27 22:06     ` Kevin Ryde
@ 2005-02-28 21:29       ` Bruno Haible
  2005-02-28 22:29         ` Kevin Ryde
  0 siblings, 1 reply; 7+ messages in thread
From: Bruno Haible @ 2005-02-28 21:29 UTC (permalink / raw)
  Cc: guile-devel

Kevin Ryde wrote:
> >   - For guile, the code and doc regarding FORMAT in SLIB are not relevant
> >     any more?
>
> Oh, well, if it's withdrawn from slib then there's nothing for guile
> to stay compatible with.
>
> >   - The guile FORMAT attempts to be an extension of Common Lisp FORMAT,
> >     i.e. that all CL format directives are supported without changes
> >     (except when documented otherwise, such as ~_)?
>
> Yes, basically, I think.  There's some things missing though, like ~< ~>.
>
> >   - Differences in behaviour between a valid CL format string and the
> >     same string in guile (except when documented otherwise) are bugs in
> >     format.scm that I should report?
>
> Yep.  I think there's some doubtful treatment of digits in some of the
> floating point output, that's the sort of thing that really ought to
> be the same as CL.
>
> >   - The reference to which I can point translators is the node
> >     "Formatted Output" in the guile documentation?
>
> For guile, yes.  What other schemes are doing is another matter.

Thanks for these answers. I have updated gettext's format-scheme.c accordingly.

> srfi-28 and srfi-48 specify versions of format too.  In the latter ~t
> is different, and ~h is an addition.

Hmm, the differences between guile's 'format' and SRFI-48 are going to be a
problem in the long run.

Also a problem is that some programs might be using the simple-format, but
xgettext doesn't know about this (since the function name is still 'format'),
and when the translator puts in translations with ~:* and the like, the
simple-format will not support it. I see no solution for it so far, other
than mentioning in the doc that the simple format is unsuitable for
internationalization.

Bruno



_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-devel


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: gettext-0.14.2 is released
  2005-02-28 21:29       ` Bruno Haible
@ 2005-02-28 22:29         ` Kevin Ryde
  0 siblings, 0 replies; 7+ messages in thread
From: Kevin Ryde @ 2005-02-28 22:29 UTC (permalink / raw)
  Cc: guile-devel

Bruno Haible <bruno@clisp.org> writes:
>
> Also a problem is that some programs might be using the simple-format, but
> xgettext doesn't know about this (since the function name is still 'format'),
> and when the translator puts in translations with ~:* and the like, the
> simple-format will not support it.

It might be worth adding ~* to simple-format.  (ice-9 format is pretty
big and not everyone is going to want to use it.)


_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-devel


^ permalink raw reply	[flat|nested] 7+ messages in thread

* Re: gettext-0.14.2 is released
  2005-02-25 23:14 ` Kevin Ryde
  2005-02-26 16:52   ` Bruno Haible
@ 2005-03-15 14:18   ` Bruno Haible
  1 sibling, 0 replies; 7+ messages in thread
From: Bruno Haible @ 2005-03-15 14:18 UTC (permalink / raw)
  Cc: guile-devel

Kevin Ryde wrote:
> Bruno Haible <bruno@clisp.org> writes:
> >   - messages can be marked as "scheme-format", denoting 'format' strings.
>
> I noticed a couple of forms that provoke errors,
>
> 	(format #f (gettext "~/"))
> 	(format #f (gettext "~q"))
> 	(format #f (gettext "~8,2i") 1+2i)
> 	(format #f (gettext "~65c"))
> 	(format #f (gettext "~4,3,'-t"))
> 	(format #f (gettext "~k") "fmtstr")
>
> ~q is a copyright message
> ~/ is a tab character (like ~| etc)
> ~i is complex number fixed-point (up to 5 params)
> ~c accepts a charnum param
> ~t accepts a pad char param
> ~k is an alias for ~?
>
> Nosing around format-scheme.c (if that's the right code) I noticed ~!
> described as a clisp call, in format.scm ~! is an output flush (no
> parameters or args), though it doesn't seem to cause an error.

This is now all fixed in gettext 0.14.3.

Bruno



_______________________________________________
Guile-devel mailing list
Guile-devel@gnu.org
http://lists.gnu.org/mailman/listinfo/guile-devel


^ permalink raw reply	[flat|nested] 7+ messages in thread

end of thread, other threads:[~2005-03-15 14:18 UTC | newest]

Thread overview: 7+ messages (download: mbox.gz follow: Atom feed
-- links below jump to the message on this page --
2005-02-25 12:31 gettext-0.14.2 is released Bruno Haible
2005-02-25 23:14 ` Kevin Ryde
2005-02-26 16:52   ` Bruno Haible
2005-02-27 22:06     ` Kevin Ryde
2005-02-28 21:29       ` Bruno Haible
2005-02-28 22:29         ` Kevin Ryde
2005-03-15 14:18   ` Bruno Haible

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).