scs@vax3.iti.org (Steve Simmons) (12/07/89)
Having just suffered for the nth time thru an incredibly stupid and verbose .sigfile, I'd really love to see a .sig-filterer for all known newsreaders. The informal (note that word) standard seems to that a .sig is anything following '--' on a line by itself, so that'd be good enough for me. An end to spurious mail: please don't followup or email me that I should complain to the peretrators or their sysops; that's just pissing upwind. Also, followups saying that TuttiFruttiNetNews already has this feature are kind of a waste. My point is that *all* news reading programs would better off with this. I've suggested the feature to the author of my favorite, and am posting this in hopes others feel the same need and will do the work or make the patches.
scs@itivax.iti.org (Steve Simmons) (12/08/89)
scs@vax3.iti.org (Steve Simmons) writes: >Having just suffered for the nth time thru an incredibly stupid >and verbose .sigfile, I'd really love to see a .sig-filterer for >all known newsreaders. The informal (note that word) standard >seems to that a .sig is anything following '--' on a line by >itself, so that'd be good enough for me. Some additional data and a correction on the '--' form: A quick survey of 303 articles from rec.arts.comics gives 16 articles that clearly have no sig whatsoever. Of the remaining 287, 117 (41%) use signatures clearly separated from the body by the following pattern: '\n--[-]*[ ]*\n' (which is supposed to mean a '--' at the beginning of the line, zero or more additional dashes, an optional space, and end of line). In no case did this pattern appear twice in an article. Nobody should take this data as gospel, tho. Sadly, ultra-long sigs were somewhat less likely to have such a marker. -- Steve Simmons scs@iti.org Industrial Technology Institute '"You're not a big name on Usenet until someone puts you in their .sig file." -- Anonymous'
brad@looking.on.ca (Brad Templeton) (12/08/89)
You don't have to change the newsreader, just the paging program. In most newsreaders you can set the pager, although most pagers aren't as good as that of rn. Just make a pager that deletes sigs. (Or filter the article through sed -e '/--/,$d' | more or something) I think there's a PAGER environment variable. If you simply want to ignore articles from people with long sigs, assuming them to be idiots, the line: reject if line_count( signature ) > 5; in Newsclip will do it. (Newsclip will be posted to comp.sources.misc shortly) Or get fancy and do reject if line_count( text ) / (1+line_count( signature )) < 3; which rejects articles which are more than 33% signature. -- Brad Templeton, ClariNet Communications Corp. -- Waterloo, Ontario 519/884-7473
tale@cs.rpi.edu (Dave Lawrence) (12/08/89)
In article <4595@itivax.iti.org> scs@itivax.iti.org (Steve Simmons) writes:
'\n--[-]*[ ]*\n'
(which is supposed to mean a '--' at the beginning of the line, zero or
more additional dashes, an optional space, and end of line).
An optional space or spaces. No brackets needed around the hyphen, either.
In no case did this pattern appear twice in an article. Nobody
should take this data as gospel, tho.
Yes, don't. And before you go rushing off to put it in your article
trimming code, be aware that it will dutifully nuke the rest of any
article that has the form:
beginning-of-line,two-or-more-hyphens,zero-or-more-spaces,end-of-line
For example, this will blow away any standard digest after the
introduction section.
It is my understanding that the traditional pattern is "\n-- \n",
though many people have since come and not known about the space (like
GNUS as distributed) and hence "\n-- ?\n" is a better regular
expression for this. I will be adding it as a variable option to
GNUS, but given that there really isn't very, very standard use of it
I expect to see only limited benefit. I suppose limited benefit is
better than none, though.
Dave
--
(setq mail '("tale@cs.rpi.edu" "tale@ai.mit.edu" "tale@rpitsmts.bitnet"))