Re: automake/338: AM_GNU_GETTEXT([external]) support

[Charles Wilson]

Alexandre Duret-Lutz wrote:

> > > > "Charles" == Charles Wilson <address@hidden> writes:
>
> [...]
>
>  Charles> Then, aclocal would automatically find
>  Charles> /usr/share/aclocal/gettext.m4 without the need for
>  Charles> -I/usr/share/aclocal -- but the test in tests/defs
>  Charles> would still fail to find it.
>
> In such a setup these tests would be skipped.  
> That's not catastrophic: they'll run elsewhere.


Okay -- as I said, the current behavior seems to be by design.


>  Charles> The real question here is, IF my patch is included,
>  Charles> then should tests/defs also be extended to use the
>  Charles> dirlist paths?
>
> If you want.  According to the size of your mail, it seems you do <wink>.


eh, as the proud papa I do -- but I understand that it may be of only 
limited use, so I won't be offended if it's not accepted.

In a different thread, someone who had posted another patch did say they 
thought this patch "solved a different problem" than their own...and 
implied that they thought mine should be accepted (along with their own, 
of course)


> [...]
>
>  Charles> unless my dirlist patch is accepted,
>
> Would you accept to complete this patch so we can include it?


Sure...


> What's needed is
>   1. that you follow existing spacing convention to format your code
>      (e.g., space before '(', not after)


Hmm...I thought I did; I'll take another look.


>   2. documentation for this new feature in the manual (its purpose
>      and the way it interacts with -I and --acdir) -- and a mention
>      in NEWS


Okay.


>   3. a test case, if possible


Urk.  I've never written one of those before.


>   4. a signed copyright assigment to the FSF (I'll send you the form
>      later today)


Not a problem.

Should the patch be against 1.6.2, or HEAD?  (HEAD is a problem; cvs is 
a PITA on a 56k dailup...)

--Chuck