Re: Possible bug in registerRecordDeviceDriver.pl

[Andrew Johnson <anj@aps.anl.gov>]

Kay-Uwe Kasemir wrote:
> I had a similar problem caused by simplistic parsing in registerRecordDeviceDriver.pl.

The format for .db and .dbd files used to be extremely free format - you 
could write
	driver(a) driver(b) driver(c)
and it would understand you perfectly.  However the parser in 
registerRecordDeviceDriver.pl is as you say extremely simplistic, because 
in practice people lay their files out to be human readable, and it wasn't 
tested on anything unusual.  It will not understand the above line, 
ignoring the entries for b and c.

I'm about to commit a slightly better version for R3.14.5 which can still 
be beaten, but won't suffer from quite so many faux-pas issues.

For R3.15 I'm working on a much better parser in perl which will follow 
the originally documented free-format rules.

- Andrew
--
Dear God, it is great the way you always get the stars in the
right place.  Why can't you do that with the moon? - Jeff