dockapps/wmbiff/TODO

99 lines
4.3 KiB
Plaintext
Raw Normal View History

This is not a prioritized list. Patches that accomplish
these todo items are welcome.
* Config file rewrite
The current configuration file has the following problems
** # comments make #'s unsuitable for passwords.
** : and @ delimiters make such characters unsuitable for
passwords.
** the hierarchical configuration is not exported cleanly,
setting a default and then propagating or overriding it
requires an odd syntax.
** elements with spaces are problematic
** annotations / customizations (such as the authentication
methods) are cumbersome.
** continuation lines would allow richer shell actions without
external shell scripts.
* GNUTLS 0.5.9
** adds gnutls_set_default_priority, which could clean up
a lot of the code to set priorities for ciphers that
don't really matter to wmbiff.
* Font support
** Erase counts > 1000 successfully (redraw the whole line)
** Discover height of the font, scale appropriately.
** Draw relative to symbolic width, height.
* Finish checking .wmbiffrc permissions.
With shell command execution support, it is even more
important to protect against other/group writability.
With passwords in the .wmbiffrc, .wmbiffrc shouldn't be
readable. It might be best to complain all the time, but
only abort when a password is used from a readable file,
or a shell command is executed from a writable file.
* Better timeout handling
Sometimes my IMAP server is slow... well, it's my fault
that I keep a few thousand messages on it. The current
timeout code will cause it to fail prematurely.
* Generic constructors for mboxes.
There's some common code across different Clients;
this is one piece that could use refactoring.
* Add more shell "recipes" besides gicu.
Just added 'finger'. I'm hoping to add something to
integrate 'remind' just as soon as I can figure it out.
* Support on/off status LED's, for, for example,
grep eth0 /proc/net/dev | wc -l
The colored LED's are already in WMBiff's pixmap.
* Support buttons 4 and 5
The mousewheel could be useful for some behaviors, like
controlling volume, or selecting different mailboxes.
* Use a helper program to blink keyboard led's.
Requires killing the 'notify' program when new mail no
longer exists. Keyboard blinking is done via a helper
program because it requires root. ledcontrol seems ok,
but extra complicated, and has a useless default install
that requires root. xbuffy's led program seems quite
good, but depending on another biff program seems odd.
* Add a failure-expect case in tlsComm.
For example, an IMAP status query has a good response (*
STATUS) and a bad response (a003 NO STATUS). As soon as
one of them is received, we're done.
* Reload .wmbiffrc when it changes
This is straightforward, except for IMAP mailboxes that
keep a connection to the server open. A close function
will have to be implemented to shut those down gracefully.
Do not involve sighup or any other signals. Use or
improve fileHasChanged().
* Runtime back-tick expansion for non-mbox
Because interesting mailboxes can be defined by date,
though its not clear any other clients can take advantage.
* Support GSSAPI or SASL authentication
Authentication using krb5 would be sweet.
* Unlikely to get done, unless someone volunteers:
* KDE/Gnome users: Test with KDE or gnome-panel
Does Debian bug #108529 apply to wmbiff? Can wmbiff
be swallowed by gnome-panel? It seems possible. It
appears to be swallowed by FVWM correctly.
* Autoconf / Automake Wizards: fix FromCVS.sh
How best to deal with libgnutls's macros? All
I can do is hack.
* Paranoid Pop People: Support POP over TLS.
tlsComm.c should make it easy to provide TLS support
for POP. Unfortunately, this would only be useful for
someone who has such a server that doesn't support IMAPS.
* MH support
Try "path.n=shell:::ls -1 directoryname | wc -l".
If it works, tell us. Optionally the patch at sourceforge
could be integrated.
* Recursive maildir support
Aggregation in general is an oft-desired feature.
* G-Mail support
Though the business model of free mail forces screen scrubbing,
having someone else write and maintain the interface seems
like it would significantly improve the utility of gmail.
$Id: TODO,v 1.21 2004/10/31 22:08:42 bluehal Exp $
Local Variables:
mode: outline
End: