Re: [Nmh-workers] Bug in 1.7 mh-format numeric output

[Ralph Corderoy]

Hi Ken,

> Looking at the previous version of cpnumber() (the offending function)
> it seems the issue was when the '?' marking truncation was inserted,
> it simply REPLACED the last character inserted, but now it's just
> appended to the string.
>
> This seems like easy enough to fix

I had a look at a simple fix, but charstring_t doesn't have a popchar(),
and altering the loop by one to stop it early didn't look obvious whilst
reading in, so I rewrote the function.  :-}  On the plus side, no
temporary charstring_t is required, with its malloc, and the result
isn't reversed at the end.  Plus, a negative width switches
justification, just like with strings.  I've updated mh-format(5) a bit,
but that whole section was and still is a bit unclear and a tangle.

a034d7e5 pushed to master.  That still leaves the option of a quick-fix
to 1.7, or cherry-picking this.

A slight whoopsy in trying to `git push origin a034d7e5...:master'.  I
tried a --dry-run with `master:a034d7e5...' and it didn't complain so
did it for real.  Unfortunately, it appears to have created a new branch
called, yep, `a034d7e5...'.  And well as polluting the public repo, it
makes local git unhappy because a 40-long hex-string is ambiguous.  How
does that get fixed?

-- 
Cheers, Ralph.
https://plus.google.com/+RalphCorderoy