duanev@kauai.ACA.MCC.COM (Duane Voth) (03/03/90)
I've been trying to pick and sort messages (letters? articles?)
in a folder but I keep getting this:
Range not broad enough to sort anything
my command is:
pick -t <pattern> | sort a
pick -t <pattern> by itself finds about 30 out of 94 messages.
sort with 's' and 'd' do the same thing. what am I missing?
also, is there a way to "partition" the current folder so that
all unread messages can be kept in the same file but with all
messages addressed to a common place in one partition, messages
to another address in another partition, and remaining messages
in a third? a blank line or an empty message would make a
reasonable partition divider which would then reduce this request
to a complicated sorting problem. any ides?
(I'm running (6.5 4/17/89) [5/12/89])
duane
--
--- duane voth duanev@mcc.com
---- {uunet,harvard,gatech,pyramid}!cs.utexas.edu!milano!pp!duanev
--- ALL systems are arbitrary! Effectiveness is the measure of Truth
--schaefer@ogicse.ogi.edu (Barton E. Schaefer) (03/03/90)
In article <619@kauai.ACA.MCC.COM> duanev@kauai.ACA.MCC.COM (Duane Voth) writes: } I've been trying to pick and sort messages (letters? articles?) } in a folder but I keep getting this: } Range not broad enough to sort anything } my command is: } pick -t <pattern> | sort a } pick -t <pattern> by itself finds about 30 out of 94 messages. } sort with 's' and 'd' do the same thing. what am I missing? What you are missing is that the messages to be sorted must be in a contiguous block, that is, their numbers must be consecutive. So you can sort 1-5, or 3-12, or 7-17, but not 1,3,5-7,12,17. You can only pipe "pick" into "sort" if you have pre-sorted on the same field you will be selecting with pick. Since there isn't any option to sort on the "To" field, "pick -t" is most likely NOT returning a consecutive list, so sort gets upset. The closest you can come with only pick and sort is sort a ; pick -t <pattern> which sorts ALL messages by author and then returns the ones "to" whoever in the right order. However,there were some "cut" and "paste" cmds posted here a while back. The idea is to use "save" (followed by possible delete/update) to cut and "merge" to paste. So: pick -t <pattern> | save pick.file | delete update # optional merge pick.file | sort a sh rm pick.file } also, is there a way to "partition" the current folder Not really. You can bundle all the related messages into a digest and use "undigest" to unpack them when you want to look at them separately. There isn't any good way to build the digest except to save all the messages into a file, change all the "From " (no colon) lines to ">From " and add a line with 8 hyphens (--------) above each >From. I could try to explain how to do this from automatically from mush, but it's tricky. Maybe digest-building would be a good thing for mush to support ... At some point mush will support cut/paste and multiple open folders, which could be used to implement your partitioning. But not yet; at least not in the next release ... -- Bart Schaefer "February. The hangnail on the big toe of the year." -- Duffy schaefer@cse.ogi.edu (used to be cse.ogc.edu)
duanev@kauai.ACA.MCC.COM (Duane Voth) (03/06/90)
In article <7682@ogicse.ogi.edu>, schaefer@ogicse.ogi.edu (Barton E. Schaefer) writes: > In article <619@kauai.ACA.MCC.COM> duanev@kauai.ACA.MCC.COM (Duane Voth) writes: > } ... is there a way to "partition" the current folder > > [digest idea...] [future cut/paste idea...] Actually, I've found something I can live with that works fairly well. I've created several bogus mail articles of the form: From ----[gwm-]-------------------- Sun Jan 1 00:00:00 1900 To: gwm-bugs@mirsa.inria.fr Subject: ----[gwm-]----------------------------- Date: Sun Jan 1 00:00:00 1900 Status: OR separator article where [gwm-] is the title of my "partition", and a cmd of the form: collect 'pick -t !* | copy ! mail.trash | delete; update; merge mail.tras h; sh rm mail.trash' Now, when I say "collect gwm-*", all articles recieved by me on the gwm-* mailing lists get placed at the end of the folder with the separator article preceeding them. It's aesthetically appealing and fairly easy to use! Possible improvements are: a) a shell script (?) that can detect my seperator articles and do a "collect" on each thus eliminating my argument to collect and making it a "sort the whole folder" command. b) some way to merge at the beginning of the folder instead of the end (so that ungrouped articles are all at the end of the folder where new articles appear). Any ideas here? One last thing, sort d couldn't correctly sort dates with years above 1999 (I tried putting the separator at the end of a partition to start with). I'm using (6.5 4/17/89) [5/12/89]. Is sort looking only at the last two digits of dates? -- --- duane voth duanev@mcc.com ---- {uunet,harvard,gatech,pyramid}!cs.utexas.edu!milano!pp!duanev --- ALL systems are arbitrary! Effectiveness is the measure of Truth --
schaefer@ogicse.ogi.edu (Barton E. Schaefer) (03/06/90)
In article <639@kauai.ACA.MCC.COM> duanev@kauai.ACA.MCC.COM (Duane Voth) writes: } In article <7682@ogicse.ogi.edu>, schaefer@ogicse.ogi.edu (Barton E. Schaefer) writes: } > In article <619@kauai.ACA.MCC.COM> duanev@kauai.ACA.MCC.COM (Duane Voth) writes: } > } ... is there a way to "partition" the current folder } > } > [digest idea...] [future cut/paste idea...] } } Actually, I've found something I can live with that works fairly well. [Description of separator article deleted] } [I presume this should be preceded by the keyword "cmd":] } collect 'pick -t !* | copy ! mail.trash | delete; update; } merge mail.trash; sh rm mail.trash' ^^ there was a space there, I deleted it } } Now, when I say "collect gwm-*", all articles recieved by me on the ^^^^^ I bet you mean "gwm-.*" (regex syntax) } gwm-* mailing lists get placed at the end of the folder with the } separator article preceeding them. It's aesthetically appealing } and fairly easy to use! This is essentially the same thing that happens when you do "undigest", if you substitute the digest article itself for the separator article and the messages burst from the digest for the other messages in the partition created by this "collect" cmd. } Possible improvements are: a) a shell script (?) that can detect my } seperator articles and do a "collect" on each thus eliminating my } argument to collect and making it a "sort the whole folder" command. This isn't hard. (I presume that by "shell script (?)" you mean a mush script.) The main drawback is that !* isn't expanded in scripts. So you'd have to change it to cmd collect 'pick -t $to ....' and then use set to='gwm-.*' collect set to='next mailing list pattern, whatever that is' collect etc. (Eventually mush may have a looping construct for this kind of thing.) } b) some way to merge at the beginning of the folder instead of the end } (so that ungrouped articles are all at the end of the folder where new } articles appear). Any ideas here? Sure. You do pick -r * -s --gwm----- | from - # move to first separator article copy ! 1-. {.} mail.trash | delete # copy+delete everything above it update; merge mail.trash # and pull it in sh rm -f mail.trash } One last thing, sort d couldn't correctly sort dates with years above 1999 This will be handled correctly in 7.1 -- along with time zones. It has the usual signed-long-integer-seconds-since-1970 limitation of around 2038, but that should suffice for a while at least. :-) } Is sort looking only at the last two digits of dates? You guessed it ... actually, mush is only *storing* the last two digits of the date; sort would handle it right if it had the rest of the info. -- Bart Schaefer "EARTH: Surrender IMMEDIATELY or we PICKLE DAN QUAYLE" "THPPFT!" schaefer@cse.ogi.edu (used to be cse.ogc.edu)