The item I promised earlier:
Now, for many years we have used the asterisk ( * ) for its sole use (far
as I know) as a wildcard in LISTSERV list archive searching; the wild card
stands for everything, all, and may be used in only a particular way, such
as:
search * in listname (search myterm* in listname is not a valid use of
asterisk ( * ) as a wildcard)
You could use date limiters if you wished (strongly advised) but you didn't
have to. You could use a starting date, with no end, or an end with no
start:
s * in listname from yy/mm/dd
s * in listname to yy/mm/dd (from = since to = until)
or both, and could get a day's posting using, say, the arbitrary date of
July 14, 2001:
s * in listname from 1/7/14 to 1/7/14 (that's Bastille day, last year)
So, s * in lstown-l from 2/8/3 sent about 5p.m. LSOFT.COM time on
2/8/3, ought to produce an output of everything distributed on LSTOWN-L
since 00:00:01 August 3, 2001, until the time when the command was
received by LISTSERV; but it doesn't.
-----------------------------------------------
s * in lstown-l from 2/8/3
-> 5 matches.
Item # Date Time Recs Subject
------ ---- ---- ---- -------
012733 16/12/24 11:17 25 Re: Bogus Subscription Requests
012869 10/01/07 14:38 23 Re: list names and trademarks
016484 10/11/07 12:26 17 WHOLESALE CONTACT LENSES !!!
022913 36/02/07 02:06 37 Re: Digest-H + Bottom_Banner
027160 02/08/03 01:21 47 Archive search
----------------------------------------------------------
Say what?!? Notice those years? Is that 1916(?) or 2016(?)? The last item
seems to match the search but the rest seem bogus. 1910? 2010? 1936?
2036? Doesn't particularly matter as either way they are wrong, unless
LISTSERV has a true time machine which can tell you what folk *would* have
written or *will* write (what does this do to authorship and copyright?).
And even if that is the case, the output does *not* match the search.
I scratched my head and tried several variations:
s * in lstown-l since 2/8/3
s all in lstown-l from 2/8/3
etc. and got the same reply for all of them.
But s * in lstown-l from 2/8/3 to 2/8/3 (closing date argument)
produced a decent output, the bogus items weren't there.
s * in lstown-l from 2/7/28
produced the items starting from 2/7/8 but the output *began* with the same
bogus items with the weird years.
I did a GETPOST for the one about contact lenses, 16484, just to see, and
sure enough, it claims the date of Sun, 7 Nov 2010 12:26:36 PM
Ummhmm. Right. I would guess the item is actually from back when LSTOWN-L
was at its original home in Indiana, though might have been 'round about
when the list was moved to Sweden. I don't really remember so maybe this
is an item which *will* be posted in 2010 (will Nov 7 be a Sunday in 2010?
I haven't checked).
I tried the open ended search on my lists at Buffalo (still 1.8d) and it
worked fine, exactly as has always worked, none of this weird stuff which
has nothing to do with the search.
Is this a function of the archives at LSOFT being royally mucked up (I
ain't saying they are, but I'ld like to know where those dates came from)?
Is this a function of 1.8e?
Is it a function of modification of preexisting archive structure via
installation of 1.8e?
Why does the wildcard open end date search not work for LSTOWN-L at
peach.ease.lsoft.com (well, it works, kinda, but you also get weird junk)
and why are the dates being nonsensically altered?
If this is the sort of thing, in addition to the other gripes about 1.8e
which have been posted on this list, to be expected from switching to 1.8e,
then I think I might beg Jim S. please to leave well enough alone, let's
stick with 1.8d.
However, I still want to know what is going on here.
Douglas Winship [log in to unmask]
|
