Re: [Nmh-workers] Second release candidate for nmh 1.6 now available

[David Levine]

> >  the dud "ISO-8859-" is the buffer initializer which i
> >  believe is caused by norm_charmap being asked to work on its own output.
> >
> >  i've fixed this by malloc()ing that buffer; see attached patch.
> 
> Thanks!  However ... that will result in a memory leak, yes?  Okay,
> you'll point out that nmh is terrible on memory management and I won't
> disagree, but I don't think that's an excuse for making it worse.
> Fixing this properly either means we need to be much more clever in
> norm_charmap(), or always allocate a buffer and fix every caller of
> norm_charmap().  Sigh.

Right.  I just committed (locally) a fix for mhshow.  I looked
at other uses and they look OK.  It turns out that other callers
(get_charset(), write_charset_8bit()) pass along the result of
norm_charmap().  Fixing every caller will be invasive.

I'll push the fix in a few minutes (I'm trying to track down the
sendfiles test problem with lzma).

David