On May 31, 2001 at 10:06, J C Lawrence wrote:
Since you mentioned that you have a MH-style mailstore, how
difficult would it be to have the filename of each message be the
actual message-id? Therefore, you can access the original mail
message just by a simple file open operation using the message-id
as the filename (with whatever pathname prefix).
Hurm. Yeah, It would be trivial to write a little tool to just
symlink each message against its MessageID. (I can't do renames s
the mailstore is regularly processed by real MH tools) <testing>
Yup, that'll work. Thanks.
To save inodes, you could use a BerkeleyDB, ndbm, or similiar type
of data store to map message-ids to filenames. There are modules
in Perl for interfacing with these type of files with a hash interface.
Other popular scripting languanges probably provide the same. The key
is to use a database format that does not have key and value length
limitations that could cause you problems.