PR 7765 and internationalization

PR 7765 and internationalization

[Gabriel Dos Reis]

Under some circumstances, the internationalization machinary returns a
null pointer for a diagnostic that is not translated, causing an ICE
where there is none as the compiler is properly concerned --
that seems to happen only with some $LANG != C.  PR 7765 concerns such
a behaviour.  Personnaly, I don't know what to such with a report. 
Close? ...?

-- Gaby

Re: PR 7765 and internationalization

[Geoff Keating]

Gabriel Dos Reis <gdr@integrable-solutions.net> writes:

> Under some circumstances, the internationalization machinary returns a
> null pointer for a diagnostic that is not translated, causing an ICE
> where there is none as the compiler is properly concerned --
> that seems to happen only with some $LANG != C.  PR 7765 concerns such
> a behaviour.  Personnaly, I don't know what to such with a report. 
> Close? ...?

I thought that this was never supposed to happen; if the string is not
translated, the original string should be returned.

-- 
- Geoffrey Keating <geoffk@geoffk.org>

Re: PR 7765 and internationalization

[Gabriel Dos Reis]

Geoff Keating <geoffk@geoffk.org> writes:

| Gabriel Dos Reis <gdr@integrable-solutions.net> writes:
| 
| > Under some circumstances, the internationalization machinary returns a
| > null pointer for a diagnostic that is not translated, causing an ICE
| > where there is none as the compiler is properly concerned --
| > that seems to happen only with some $LANG != C.  PR 7765 concerns such
| > a behaviour.  Personnaly, I don't know what to such with a report. 
| > Close? ...?
| 
| I thought that this was never supposed to happen; if the string is not
| translated, the original string should be returned.

That was also my understanding until I went throught it and it was
independently reported.  I have no theory to offer as to why it
doesn't work as documented, however :-)

-- Gaby

Re: PR 7765 and internationalization

[Fergus Henderson]

On 09-Oct-2002, Gabriel Dos Reis <gdr@integrable-solutions.net> wrote:
> 
> Under some circumstances, the internationalization machinary returns a
> null pointer for a diagnostic that is not translated, causing an ICE
> where there is none as the compiler is properly concerned --
> that seems to happen only with some $LANG != C.  PR 7765 concerns such
> a behaviour.  Personnaly, I don't know what to such with a report. 
> Close? ...?

So there is a bug in a library that GCC is using, which is not part of GCC?

I would advise the following procedure:
	1. Report the bug to the external library provider.
	2. [Optional] If the library source is available, fix it! :-)
	3. Add a work-around in the GCC sources
	   (this is desirable even if the bug is fixed --
	   many users might still be using the old version of the library).
	4. If the problem can't be worked around in a manner that is completely
	   transparent to the end user, document the problem in the "Trouble"
	   section of the GCC manual.
	5. Once steps 1 and either 3 or 4 have been done,
	   close the bug report.

-- 
Fergus Henderson <fjh@cs.mu.oz.au>  |  "I have always known that the pursuit
The University of Melbourne         |  of excellence is a lethal habit"
WWW: <http://www.cs.mu.oz.au/~fjh>  |     -- the last words of T. S. Garp.