Re: Enhanced language support for Modula-2

[Eli Zaretskii <eliz at gnu dot org>]

> Cc: gdb-patches@sources.redhat.com
> From: Gaius Mulley <gaius@glam.ac.uk>
> Date: 21 Feb 2006 10:24:46 +0000
> 
> here is a reworking of my patch which includes a new version of the
> set reading code in gdb/dwarf2read.c. Also contained is a patch to
> gdb/doc/gdb.texinfo reflecting the enhancements to the Modula-2 mode.

Thanks.  My comments are below.

> +      error (_("type not handled in m2_type_print_varspec_prefix()"));

Please don't append "()" to indicate that you are talking about a
function; just name the function.  "foo()" looks like a call to `foo'
with no arguments, which is not what you want to say.

> +      error (_("type not handled in m2_type_print_varspec_suffix()"));

Same here.

Btw, isn't it better to tell what type was not handled?

> +      fprintf_filtered(stream, "SET OF %s", _("<unknown>"));

It is better to include "SET OF" in the string to be translated:

      fprintf_filtered(stream, "%s", _("SET OF <unknown>"));

This way, the translator will see the whole phrase, and will be able
to translate "unknown" to a correct grammatical form (e.g., wrt
gender).  If you think "SET OF" should not be left verbatim, you can
put an i18n comment there:

    /* i18n: Do not translate the "SET OF" part!  */

These specially formatted comments will be extracted to the message
catalog, so the translators will see them right next to the string
they need to translate.

> +	fprintf_filtered (stream, "CASE %s OF\n", _("<variant>"));

Same here.

> +      fprintf_filtered (stream, "struct %s", _("<unknown>"));

And here.

> +	  fprintf_filtered (stream, "<%s %d>",
> +			    _("invalid type code"), TYPE_CODE (type));

And here.

> --- latest-cvs-gdb/src-cvs/gdb/doc/gdb.texinfo	2006-02-10 03:54:33.000000000 +0000
> +++ latest-cvs-gdb/src-m2/gdb/doc/gdb.texinfo	2006-02-21 10:10:41.000000000 +0000

The documentation patch is approved, but please take care of the
following minor issues:

> +@cindex Modula-2 types
> +@node M2 Types

The @cindex entry should come _after_ the @node, not before it.

> +@subsubsection Types

"Types" is too general a name here: this is just about Module-2 types;
there are quite a few other languages supported by GDB that have
types.  So I suggest to rename to "Modula-2 Types".

> +Values of these types may also be printed. So for example, if the
                                            ^^
Two spaces after each period that ends a sentence, please.

> +@example
> +VAR
> +   s: SET OF CHAR ;
> +   r: [20..40] ;
> +@end example
> +
> +@value{GDBN} could be requested to interrogate the type and value of

This last line is a continuation of the sentence from before the
@example, so please put @noindent before this line, to prevent its
indentation as if it were a new paragraph.

> +@example
> +VAR
> +   s: SET ['A'..'Z'] ;
> +@end example
> +
> +the user can query the type of @code{s} via:

Same here (and elsewhere, there are a few more like this).

> +It is advisable to use the following compile flags on the @code{gm2}
> +command line @code{-g -gdwarf-2}.

First, if there's only one command-line option you want to advise, why
use plural ``flags''?  And second, command-line options should be in
@option, not in @code.

> +(gdb) print s

Please use (@value{GDBP}) instead of a literal (gdb) (it's for the
same reason we use @value{GDBN} instead of a literal GDB.

I also agree with Jim's comments.