Re: [Dbmail] [Dbmail-Users] Dbmail and DNS

similar problem today. I kludged around it by setting the transport to lmtp:127.0.0.1:24 instead of lmtp:localhost:24. That shouldn't be required, but it did fix it for me. You should also check the chroot(2) some of the postfix daemons live in and confirm that a usable etc/resolv.conf and etc

Re: [Dbmail] DBmail webmail app

those using pg. At least with 8.1 and up; I can't remember whether 7.4 already had to automatic large object support. (Any large object is automatically stored out of the table, and a pointer to it stored in the table. There may be a config option to specify what constitutes a large object

Re: [Dbmail] Changelog

in the queries. OTOH, since the use of ILIKE and/or ~ invalidates the need for indices, they currenly are providing almost no value, just making INSERTs and UPDATEs more expensive. Only when = is used whenever possible will the indices provide any real value. -JimC -- James Cloos &

Re: [Dbmail] "-" and "_" in folder names

apply to '_'. Please don't. I have *several* (my folderGrep script says 436 right now) folders with _ in them, and use it as a substitute for several difficult characters, such as ' ', in auto-named folders. Living w/o underscore -- and hyphen for that matter -- wou

Re: [Dbmail] Spammer and dbmail

nitive.) So dbmail's code for breaking the From: header up for the dbmail_fromfield table needs some love -JimC -- James Cloos <[EMAIL PROTECTED]> OpenPGP: 1024D/ED7DAEA6 ___ DBmail mailing list DBmail@dbmail.org https://mailman.fastxs.nl/mailman/listinfo/dbmail

Re: [Dbmail] Discursion about bug #465

o contiguous per-folder UIDs will fix it for gnus, maybe also for others. -JimC -- James Cloos <[EMAIL PROTECTED]> OpenPGP: 1024D/ED7DAEA6 ___ DBmail mailing list DBmail@dbmail.org https://mailman.fastxs.nl/mailman/listinfo/dbmail

Re: [Dbmail] db server -- 64 bit, or 32 bit?

enerates better code. And the better ABI should also make a difference -- especially for function calls. I'd only expect to notice the difference under load, though. -JimC -- James Cloos <[EMAIL PROTECTED]> OpenPGP: 1024D/ED7DAEA6

Re: [Dbmail] db server -- 64 bit, or 32 bit?

disk space usage by about half. Definitely compile postgresql with -m64. If you are on amd64 also compile dbmail with -m64. On UltraSPARC you may find dbmail will run faster compiled with -m32, but given how much use it makes of int64_t I'd give both versions a test. -JimC -- James C

Re: [Dbmail] db server -- 64 bit, or 32 bit?

e two INT8s for most users and the current schema for the largest installs. -JimC -- James Cloos <[EMAIL PROTECTED]> OpenPGP: 1024D/ED7DAEA6 ___ DBmail mailing list DBmail@dbmail.org https://mailman.fastxs.nl/mailman/listinfo/dbmail

Re: [Dbmail] Re: DBMail 2.3.0 released

ISO/IEC 10118-3 http://www.incits.org/ref-docs/FDIS_10118-3.pdf -- James Cloos <[EMAIL PROTECTED]> OpenPGP: 1024D/ED7DAEA6 ___ DBmail mailing list DBmail@dbmail.org https://mailman.fastxs.nl/mailman/listinfo/dbmail

Re: [Dbmail] Re: DBMail 2.3.0 released

treading new ground here. That said, debate is good, fun and highly constructive. Let it continue! -JimC -- James Cloos <[EMAIL PROTECTED]> OpenPGP: 1024D/ED7DAEA6 ___ DBmail mailing list DBmail@dbmail.org https://mailman.fastxs.nl/mailman/listinfo/dbmail

Re: [Dbmail] Re: DBMail 2.3.0 released

|> Yes, the birthday paradox is relevant. I should've been more precise. The borthday paraox is in fact the reason why one uses a 2n-bit hash to equal the strenght of an n-bit symmetric cypher. So I did implicitly take the paradox into account in my first post. -JimC -- James Cloos

Re: [Dbmail] 'ERROR: invalid byte sequence for encoding "UTF8": 0x96'

ity — that takes the search string as an argument and does its own grepping of the rows of the messageblks.messageblk and mimeparts.data columns which match is_header=0. And then arrange for dbmail to call that function when doing searches on the body. Doable, but more than a little bit of work. -J

Re: [Dbmail] mail encryption

want to invest in some kind of true random number generator for the server to be sure you don't deplete /dev/random and DOS the incoming mail. -JimC -- James Cloos <[EMAIL PROTECTED]> OpenPGP: 1024D/ED7DAEA6 ___ DBmail mailing list DBmail@dbmail.org https://mailman.fastxs.nl/mailman/listinfo/dbmail

Re: [Dbmail] Size of dbmail DB in postgreSQL

ng the list of headers saved in headervalue would have a significant and beneficial impact on performance. -JimC -- James Cloos <[EMAIL PROTECTED]> OpenPGP: 1024D/ED7DAEA6 ___ DBmail mailing list DBmail@dbmail.org https://mailman.fastxs.nl/mailman/listinfo/dbmail

[Dbmail] Re: [Dbmail-dev] Announce: subversion is a goner

tar files, but it is useful for those who use the tar files rather than git. -JimC -- James Cloos <[EMAIL PROTECTED]> OpenPGP: 1024D/ED7DAEA6 ___ DBmail mailing list DBmail@dbmail.org https://mailman.fastxs.nl/mailman/listinfo/dbmail

Re: [Dbmail] DBMAIL error on SELECT

ven name and argument type(s). You might need to add explicit type casts. ] I don't know whether it is possible to do full text searches on a BYTEA w/o retrieving the whole thing; perhaps one of the extensions might add a function which can do it? -JimC -- James Cloos OpenPGP: 1024D/ED7DAEA6 ___ DBmail mailing list DBmail@dbmail.org http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail

Re: [Dbmail] DBMAIL error on SELECT

ail_messageblks | where is_header=0 and | convert_from(messageblk,'SQL_ASCII') ~ 'dbmail-dev' ) |order by physmessage_id, is_header desc; ` Generating an index of the output of convert_from(messageblk,'SQL_ASCII') would be painful at best. -JimC -- J

Re: [Dbmail] PostgreSQL 8.4.x or 9.0.x as mailstore

lso using the E'' syntax. I think that was the only issue I encountered. I've not yet tried importing the dbmail db into 9.0. -JimC -- James Cloos OpenPGP: 1024D/ED7DAEA6 ___ DBmail mailing list DBmail@dbmail.org http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail

[Dbmail] postgresql 9.0

duced shared pressure to allow larger sites to retain fully cached indices for a given ram size. Postfix and dspam also work well with 9.0. -JimC -- James Cloos OpenPGP: 1024D/ED7DAEA6 ___ DBmail mailing list DBmail@dbmail.org http://mailman.fast

Re: [Dbmail] postgresql 9.0

>>>>> "LM" == Lasantha Marian writes: LM> By the way, can you confirm whether the DBMail used was 2.2 or 2.3 ? I use my patched version of 2.2 (the archives have a pointer to the current version of my patches. -JimC -- James Cloos

Re: [Dbmail] DBMail insert email via SQL?

X. Cf dbmail-smtp --help for the details on its argument syntax. -JimC -- James Cloos OpenPGP: 1024D/ED7DAEA6 ___ DBmail mailing list DBmail@dbmail.org http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail

[Dbmail] ilike

in terms of total runtime than the ILIKE sequential scan. That could be a significant advantage on busy databases. -JimC -- James Cloos OpenPGP: 1024D/ED7DAEA6 ___ DBmail mailing list DBmail@dbmail.org http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail

Re: [Dbmail] ilike

LOWER instead of ILIKE is still the best solution. -JimC -- James Cloos OpenPGP: 1024D/ED7DAEA6 ___ DBmail mailing list DBmail@dbmail.org http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail

Re: [Dbmail] ilike

d benefit from eliminating ILIKE in favour of LOWER() and LIKE. DBmail already added in the index needed to make this change useful, presumably thinking that ILIKE would use it. But ILIKE does not. -JimC -- James Cloos OpenPGP: 1024D/ED7DAEA6 ___ DBmail mailing list DBmail@dbmail.org http://mailman.fastxs.nl/cgi-bin/mailman/listinfo/dbmail

Re: [Dbmail] (imap) performance for list folders

gle query should be needed to grab any given message's messageblks. The rest of the metadata should have been grabbed by the enter-group query. Then dbmail would be efficient. But that only works with sql servers which support enough of ansi sql. CTEs, in particular, are required for dbmail&