|
Grex > Helpers > #138: Grex System Problems - Winter 2004/2005 |  |
|
| Author |
Message |
| 25 new of 870 responses total. |
keesan
|
|
response 352 of 870:
| 
Jan 12 03:14 UTC 2005 |
Many thanks for talk. Yesterday I had two long exchanges of info via tel
(which is even more confusing than usual when grex is running at load average
60 or the language of communication is Russian in the Roman alphabet).
|
aruba
|
|
response 353 of 870:
|
Jan 12 03:16 UTC 2005 |
Is it your birthday, Anne?
|
keesan
|
|
response 354 of 870:
|
Jan 12 03:51 UTC 2005 |
Lynx still offers me the option of sending a downloaded file to my computer
via kermit/telnet but I don't think that ever worked. Could lynx.cfg be
changed to reflect reality? Or grex be changed to allow kermit file transfer
via telnet? Maybe it once used to be slower than ftp but it is just as speedy
now and often more convenient than saving a file to a home directory, and then
starting a new connection to grex in order to ftp it. I was downloading a
few photos of someone I had a one-hour tel exchange with yesterday who had
ideas on how to improve procmail filtering. (His idea was to throw out all
mail from aol, yahoo, msn, earthlink - he said this gets 90% of the spam).
|
mooncat
|
|
response 355 of 870:
|
Jan 12 03:53 UTC 2005 |
Re#353- Yup. :) I'm 30 now!
|
mfp
|
|
response 356 of 870:
|
Jan 12 04:15 UTC 2005 |
You're more than ten years older than me!
|
sholmes
|
|
response 357 of 870:
|
Jan 12 04:23 UTC 2005 |
Happy Birthday !
|
gelinas
|
|
response 358 of 870:
|
Jan 12 04:24 UTC 2005 |
The "no route to host" message is often a result of the packet filter
blocking outbound packets.
OpenBSD's packet filters don't look at all the groups a person belongs to, so
we have to link programs to "inet_wrap", which then sets the group
appropriately and calls the real program.
wget is not on the 'approved' list. links is, but it does not seem to be set
up properly.
The problem with DTE appears to be malformed html: they don't properly
separate the file name from the host name. I don't know why links on sdf can
parse the link. If I recall correctly, links reports the name of the host it
cannot reach. If you look at that domain name, the problem should be
obvious.
|
keesan
|
|
response 359 of 870:
|
Jan 12 04:32 UTC 2005 |
Links at sdf works on the dteenergy site on the second try (but the site has
other problems so I may give up on it).
I will use straight ftp until lynx and wget work again with ftp sites.
DTE says they asked the company doing the website to fix some problems but
it just keeps getting worse instead of better. Thanks Joe for spending so
much time on fixing these small problems.
|
gelinas
|
|
response 360 of 870:
|
Jan 12 04:47 UTC 2005 |
I fixed the birthday-greeter. Thanks for reporting its failure, Anne.
And happy birthday. :)
|
mooncat
|
|
response 361 of 870:
|
Jan 12 05:25 UTC 2005 |
re #356- Yeah, nyah nyah! ;)
re#357- thanks. :)
Re #360- you're welcome, and thanks! :)
|
other
|
|
response 362 of 870:
|
Jan 12 06:13 UTC 2005 |
Backtalk authentication problem:
Using the vanilla interface, some of the links at the bottom of a page
when browsing new responses to an item are formatted such that even if
the user is connecting via https, the link will be formatted for http,
which results in a browser asking for cleartext authentication before
submitting a "forget" request, for example.
|
other
|
|
response 363 of 870:
|
Jan 12 06:14 UTC 2005 |
Backtalk authentication problem:
Also, response 'Post' buttons submit via http rather than https.
|
davel
|
|
response 364 of 870:
|
Jan 12 16:45 UTC 2005 |
Re way back (out of my scrollback buffer) re my problem with ft "read since":
Joe, I did read the ft help. (It didn't differ from my memory of Picospan
help on this issue, but it's been a few years since I looked at that.)
"help read" referred me to (I think) "help range", which gave the format I was
trying as an example.
|
keesan
|
|
response 365 of 870:
|
Jan 12 17:52 UTC 2005 |
My mailbox this morning was full because a friend sent me two photos, one of
which was about 400K and the other a bit smaller. Please could the mail size
limit be given a high priority? McNally, until then, could you explain
precisely how to set up a .procmailrc that will dump such mails? I did not
follow the last explanation. For the benefit of all email users here, please
also include the first three lines of .procmailrc and the last two.
.forward is, I think |/usr/bin/procmail right? This sends all incoming
mail to procmail to be filtered. Maybe someone could run a script to set up
this sort of mail size filter for everyone until the exim one is working?
Why would 'cpl' dump a returned message from me?
|
tsty
|
|
response 366 of 870:
|
Jan 12 18:57 UTC 2005 |
is there any sort of filtering mechanism creatable (by user or system)
that can mostly de-spam the stuff in 'plain ol mail'?
same as dpc and keesan, i am *really* getting slammmmmmmed/spammmmmed!
|
rcurl
|
|
response 367 of 870:
|
Jan 12 19:00 UTC 2005 |
If I could just filter on "cum", it would take care of about half my spam.
|
mcnally
|
|
response 368 of 870:
|
Jan 12 19:10 UTC 2005 |
> My mailbox this morning was full because a friend sent me two photos,
> one of which was about 400K and the other a bit smaller. Please
> could the mail size limit be given a high priority?
>
> McNally, until then, could you explain precisely how to set up a
> .procmailrc that will dump such mails?
Try adding a recipe like this to test with:
:0fw
* > 50000
/a/k/e/keesan/bigmail
Then send yourself two messages, one small one, one with at least 50 kbytes
of content. The larger message should wind up in ~keesan/bigmail and not
in your inbox. Once you're satisfied that it filters as you expect you can
change the destination file to /dev/null.
|
other
|
|
response 369 of 870:
|
Jan 12 19:38 UTC 2005 |
367: Do you get a lot of spam in Latin?
|
keesan
|
|
response 370 of 870:
|
Jan 12 19:40 UTC 2005 |
I get a lot of spam with utf or iso on subject lines and just now one of my
two latest spams was about a cumshot but that is only about 5% of the spam.
Is there some way to set a filter to forward mails over 100K to another
address instead of a folder here? They are usually photos from friends.
|
keesan
|
|
response 371 of 870:
|
Jan 12 19:46 UTC 2005 |
Thanks Mike, that worked and I am now sending mails over 100K to /dev/null.
I notice that in my filter I am sending Larry Nigeria spam with
the string ! polygon@.. WOuld I just put something like that in instead
of the /dev/null on the third line to forward big mails to myself at some
other address?
If anyone else wants to use Mike's filter, be sure to also include the first
three lines of my .procmailrc and the last three lines (along with this filter
which is now in there) so that the rest of your mail will go to your inbox,
and change keesan to your login.
The only problem with this method is that the mail does not bounce back to
the sender so they don't write and ask how come the mail bounced, but you can
check the log file daily and write people you know whose large mails bounced.
You also need the .forward file so that mail will go through procmail.
|
keesan
|
|
response 372 of 870:
|
Jan 12 19:48 UTC 2005 |
Minor problem with Mike's filter: Extraneous filter-flag ignored. Does this
mean one of the characters in :0fw should be removed? The filter works
anyway, letting through a small mail but diverting a big one.
|
blaise
|
|
response 373 of 870:
|
Jan 12 19:55 UTC 2005 |
From the Procmail Quick Start
(http://www.ii.com/internet/robots/procmail/qs):
Matching a Word
If you want a recipe to match only the word test, rather than the string
test, surround test with the Procmail word delimiters \< and \> as in
the following:
:0:
* ^Subject:.*\<test\>
IN-testing
[end quote]
So to do what Rane is looking for would be
:0:
* ^Subject:.*\<cum\>
/dev/null
as a Procmail rule.
I highly recommend that anyone who wants to use Procmail for filtering
(whether spam or otherwise) read the Quick Start; it is very well
arranged to walk people new to Procmail through setting it up.
|
keesan
|
|
response 374 of 870:
|
Jan 12 20:02 UTC 2005 |
Thanks for that info. I have not needed to filter on cum because mail
containing it includes so many other easily filtered strings.
I removed fw from Mike's filter and it still works and no longer tells me
there is an extraneous filter-flag: First line :0, second line * > 100000
third line /dev/null (I cannot type a colon at line beginning here).
|
blaise
|
|
response 375 of 870:
|
Jan 12 20:03 UTC 2005 |
The best I can tell, you don't need either the f or the w flag in the
size-catching recipe. Those are only needed when the action of the
recipe uses an external program, from what I can find.
|
blaise
|
|
response 376 of 870:
|
Jan 12 20:04 UTC 2005 |
Note that the recipe I gave will match "cum" *in the subject line only*,
and *only as a word by itself*. (For example, I didn't want to match
"cumulative"...)
|