Sun, 28 Feb 1993 22:19:59 +0100
|
Beta 1.7f-3 will be distributed tomorrow. It contains a number of minor
fixes, a change to the digest/index processor, and a change to the LISTS
database driver.
The subject field listed in indexes will no longer have 'Re:' prefixes
stripped. This was making it difficult to see what messages were new and
which ones were replies. The "table of contents" in the digest is
unchanged, with the 'Re:' prefixes removed.
An optional "digest header" text can be inserted between the end of the
table of contents and the beginning of the digest itself. If a LISTSERV
(not CMS) file by the name 'listname DIGEST-H' is found, its contents
will be inserted at this point in the digest. This file, like list
welcome messages, can be maintained by the list owner.
The instructions and job skeleton at the end of indexes can be replaced
(not added to) by shorter, customized text from the file 'listname
INDEX-H'. Again, if that file exists it replaces EVERYTHING after the
list of messages. If you want the instructions to remain, just copy them
to the INDEX-H file. The purpose of this addition is to replace the
instructions with a version better tailored to your audience (no point
explaining about REPLY TEXT if your audience is 100% PROFS).
A "farewell message", similar to the welcome message, has been added. The
file is called 'listname FAREWELL' and must start with a 'Subject:' line.
The message is sent when the user signs off or during a DELETE without
the QUIET option.
The change to the LISTS database driver makes it possible to detect
inconsistencies between the contents of the LISTS database and GLOBLIST
FILE. While the checksum/refresh mechanism used to synchronize this
distributed database ensures that GLOBLIST FILE discrepancies are
corrected within at most a month, it doesn't address the case where the
information in GLOBLIST FILE is correct (hence no checksum error) but
some data is missing from the LISTS database. This condition cannot
happen under normal circumstances, but the installation of the 1.7f LISTS
database driver artificially introduces such a discrepancy, as GLOBLIST
FILE is left unchanged while the old LISTS database files are no longer
seen/used. The new driver forces a refresh when the amount of lists (for
a given node) differs between the two files.
The new driver also obsoletes LUPDTIME FILE, moving the information to
PERMVARS FILE (a file added by 1.7f which is similar to LASTING GLOBALV,
but without the length restrictions). Because the data is not migrated,
this will reset the "last date" information.
Eric
|
|
|