I have dug even deeper in the source code of Prayer, added IPv6 support, got a
reply from upstream maintainer David Carter, and thought about it. The
question is: Will the (additional) effort needed to produce and maintain a
Debian package of good enough quality be worth it, considering the following?
1. Prayer is geared towards large-scale, perhaps even *very* large-scale
installations. It offers speed and low resource usage at the expense of
flexibility. Most admins probably will want to use an Apache-based package,
such as Squirrelmail.
1a. About the flexibility: Changing the appearance is rather hard and, except
for some colours, requires recompilation. Prayer produces HTML4 Transitional,
full of <BODY color=... bgcolor=... >, <FONT> etc. It should preferably be
changed to use CSS.
2. Support for other character sets than ISO-8859-1 is non-existant.
Conversion of various mail text to UTF-8 has to be added.
3. Prayer isn't prepared for l10n. All UI strings have to be gone through and
wrapped in gettext calls.
4. The code is a bit messy in the sense that there are many almost-similar
variants of the same functions. It would do well with some restructuring.
5. At least minimal man pages have to be written.
6. And last but not least, Prayer is practically dead upstream. David Carter
says that the purpose of Prayer was to fit on top of UW-based mail systems
which really weren't designed to run Webmail. Now, after 5 years, they don't
need it anymore. No more releases are planned.
To all this there is the problem with combined folders/directories already
mentioned.
Comments are welcome, especially from the original requester. How many do you
think will find Prayer useful?
(It's not that I want to give up, but it's a bit silly to maintain a package
nobody uses.)
--
Magnus Holmgren holmgren@lysator.liu.se
(No Cc of list mail needed, thanks)
Attachment:
pgpR665Ezd7ob.pgp
Description: PGP signature