[Erc-help] a couple of issues (services-module; doc string; buffer creation; encoding)

[Tobias C. R. <tc...@fr...>]

Hi,

* I wonder how nicely the 'nickserv module plays with the usual
  authentification mechanism (I'm speaking of `erc-password' and
  `erc-prompt-for-password'.) Do I still have to set the latter variable
  to NIL when I specified a password to be used via the 'nickserv
  module?  After a short glance over the code, I'd say yes---although I
  didn't really grasp the hook mechanisms spooking in `erc-nickserv.el'.
 

* As I read through the erc-<module>.el files, I got confused by theirs
  commentaries because, for instance, `erc-stamp.el' says there: ``In
  order to actually activate this, you must call `erc-timestamp-mode'.''

  Of course, this is absolutely correct when looking at the commentary
  from a module-oriented perspective. That's why I don't suggest to
  change it. However, the defcustom `erc-modules', which I consulted for
  clarification, wasn't as much a relief as I wished it was, because I
  misunderstood the first sentence of the doc string which goes:

    A list of modules which erc should enable

  I mixed up what ``enable'' means in this context, ie. I thought the
  modules specified in `erc-modules' were just loaded but not activated
  and I had to invoke the ERC-<foo>-MODE functions manually. Even though
  not wrong, I suggest to change the doc string to something like
  
    A list of modules which erc load and activate automatically. This
    loading and activation is performed by `erc-update-modules', hence
    remember to call \(erc-update-modules), if you set the value of this
    without using `customize'.
    
    Note: When using `customize', modules removed from the list will be
          disabled.


* If I'm queried (or if generally a new buffer ought to be created), I'd
  like the newly created buffer to replace the current buffer of a
  specific frame. I usually have erc in its own frame and Gnus, too, and
  when I'm writing a mail and someone queries me, my mail buffer pops
  off (but instead erc should do so with the most recently used buffer in
  the erc frame.) Is this possible? Or how do you deal with this problem?


* I intend to change the format of `erc-encoding-coding-alist' to
  introduce an additional server level, so that the encoding is
  determined not only by matching a channel, but also by matching an
  irc-server (analogously to `erc-autojoin-channels-alist'.) I think
  this can be done backwards-compatibly by testing if the CDR of an
  entry in the encoding alist is a symbol or not. (I think, this is easy
  enough that I'll do that myself when I have some spare time.


--tcr.