Re: [Nmh-workers] JLR

[John Romine]

> [2010-10-29 14:22] David Levine <address@hidden>
> > meillo wrote:
> > 
> > > just one question for better understanding:
> > > 
> > > Who or what is ``JLR''?
> > > 
> > > There are several #ifdefs for JLR in code which is related to scan(1).
> > > There is also a comments which appears to be written by JLR and
> > > docs/ChangeLog_MH-3_to_MH-6.6 lists changes done for JLR.
> > 
> > John Romine, I expect.  From docs/README.about:
> > 
> >   In 1982, the University of California, Irvine took up
> >   maintenance of the software, under the direction of Marshall
> >   T. Rose and John L. Romine.
>
> That seems to be reasonable.
>
> Does anyone know some more? Had he had his personal, modified,
> version of MH, because this is how it appears?
>
>
> meillo
Stuff under #ifdef JLR means something experimental I was working on.

In this case (scan, formatsbr) it has to do with an extension to the mh-format
syntax to allow for looping.

The scan format is processed once for each message.  Those #ifdef JLR changes
allowed for the top part of the format file to be processed once, then a second,
looping part to be processed once per message.  As I recall, there were new 
mh-format escape sequences to delimit the loop.  This would have allowed for 
things like per-format column headings in the scan output.

Since existing format files didn't include the scan listing header (it was hard-coded
in scan.c) it would not have been backward-compatible.  All existing format files
(including any local ones) would have needed to be changed to include the format
codes for a header.  The practice at the time was not to introduce incompatible
changes in a minor release, and I never managed to put out a newer major release.

John