ids in received lines (was: lexer size [was: Much simplified lexer])

[David Relson]

On Fri, 14 Nov 2003 11:53:29 +0100
Boris 'pi' Piwinger <3.14 at logic.univie.ac.at> wrote:

> David Relson wrote:
> 
> > Looking back at older versions of bogofilter, I found the following:
> > 
> > E?SMTP\ id\ .*	;
> > 
> > which means that ESMTP and SMTP lines were being ignored.
> > 
> > The current code simply returns the token if header_line_markup is
> > enabled.  Since header_line_markup is the default state, the normal
> > action is just to return the token.  Sicne that is the same thing
> > that bogofilter would do without the special rule, I've removed the
> > rule.
> 
> make check fails for me with this. Anyhow, the line
> <INITIAL>[:blank:]*id\ {ID}
> look wrong. Why would we want zero or more blanks here? I
> think we need at least one and for our purpose exactly one
> should do it.
> 
> pi

pi,

It would be helpful if you test changes "make check" before posting
them.  If the changes cause tests to FAIL, then examine the files
involved to see what changed and see if it's something that should
happen or that should not happen.  With that info you can tell if the
change should be considered further.

In the case of [:blank:]+ vs. [:blank:]*, "make check" reports:

	want: 1784.4029, have: 1805.4160
	FAIL: t.lexer.mbxPASS: t.folded_rfc2047

(and other failures)

Comparing the output files from a successful and failed runs of
t.lexer.mbx shows lots of message IDs (see attachment).

So, please test changes before suggesting them.

Thank you.

David
-------------- next part --------------
A non-text attachment was scrubbed...
Name: diff.spam.2
Type: application/octet-stream
Size: 768 bytes
Desc: not available
URL: <http://www.bogofilter.org/pipermail/bogofilter/attachments/20031114/5162af89/attachment.obj>