Re: [nmh-workers] Reply to Ralph Corderoy

[Ralph Corderoy]

Hi Valdis,

> > ISTM it's also a bug to ignore problems opening a file that exists,
> > and to ignore a file given in an environment variable not existing.
>
> I agree on the first.  The second probably need to be more nuanced,
> and "a *specific* file given in the environment".  Consider how *PATH
> variables work, and it's possible that none of the directories contain
> the file in question and it's still OK.

Yes, I meant only environment variables intended for nmh and defined by
it to have a path, absolute or relative, to a file, e.g. MHMTSCONF.
mh_profile(5) lists many.  I don't think any of them are PATH-like;
I agree, when searching a PATH-like it's not an error if a directory
doesn't exist.

I've opened https://savannah.nongnu.org/bugs/index.php?56192 so it
doesn't get forgotten.

-- 
Cheers, Ralph.