commit 3ba3b53b27833b63d44a31df3b874ee144dcbd60
parent ef322a05db1749e7a4861ec95484ad219415421c
Author: wmorgan <wmorgan@5c8cc53c-5e98-4d25-b20a-d8db53a31250>
Date: Sat, 6 Jan 2007 16:15:48 +0000
comment updates
git-svn-id: svn://rubyforge.org/var/svn/sup/trunk@216 5c8cc53c-5e98-4d25-b20a-d8db53a31250
Diffstat:
2 files changed, 40 insertions(+), 18 deletions(-)
diff --git a/lib/sup/mbox/loader.rb b/lib/sup/mbox/loader.rb
@@ -106,7 +106,7 @@ class Loader < Source
@f.gets # now we're at a BREAK_RE, so skip past it
else # case 2
returned_offset = cur_offset
- ## we've already skipped past the BREAK_RE, to just go
+ ## we've already skipped past the BREAK_RE, so just go
end
while(line = @f.gets)
diff --git a/lib/sup/source.rb b/lib/sup/source.rb
@@ -3,26 +3,48 @@ module Redwood
class SourceError < StandardError; end
class Source
- ## dirty? described whether cur_offset has changed, which means the
- ## source needs to be re-saved to disk.
+ ## Implementing a new source is typically quite easy, because Sup
+ ## only needs to be able to:
+ ## 1. See how many messages it contains
+ ## 2. Get an arbirtrary message
+ ## 3. (optional) see whether the source has marked it read or not
##
- ## broken? means no message can be loaded, e.g. IMAP server is
- ## down, mbox file is corrupt and needs to be rescanned.
-
- ## When writing a new source, you should implement:
+ ## In particular, Sup doesn't need to move messages, mark them as
+ ## read, delete them, or anything else. (Well, maybe delete at some
+ ## point.)
##
- ## start_offset
- ## end_offset
- ## load_header(offset)
- ## load_message(offset)
- ## raw_header(offset)
- ## raw_full_message(offset)
- ## next (or each, if you prefer)
-
- ## you can throw SourceErrors from any of those, but we don't catch
- ## anything else, so make sure you catch all non-fatal errors and
- ## reraise them as source errors.
+ ## On the other hand, Sup assumes that you can assign each message a
+ ## unique integer id, such that newer messages have higher ids than
+ ## earlier ones, and that those ids stay constant across sessions
+ ## (in the absence of some other client going in and fucking
+ ## everything up). For example, for mboxes I use the file offset of
+ ## the start of the message. If a source does NOT have that
+ ## capability, e.g. IMAP, then you have to do a little more work to
+ ## simulate it.
+ ##
+ ## To write a new source, subclass this class, and should implement:
+ ##
+ ## - start_offset
+ ## - end_offset
+ ## - load_header offset
+ ## - load_message offset
+ ## - raw_header offset
+ ## - raw_full_message offset-
+ ## - next (or each, if you prefer)
+ ##
+ ## You can throw SourceErrors from any of those, but we don't catch
+ ## anything else, so make sure you catch *all* errors and reraise
+ ## them as SourceErrors, and set broken_msg to something if the
+ ## source needs to be rescanned.
+ ##
+ ## Also, be sure to make the source thread-safe, since it WILL be
+ ## pummeled from multiple threads at once.
+ ## dirty? described whether cur_offset has changed, which means the
+ ## source info needs to be re-saved to sources.yaml.
+ ##
+ ## broken? means no message can be loaded, e.g. IMAP server is
+ ## down, mbox file is corrupt and needs to be rescanned, etc.
bool_reader :usual, :archived, :dirty
attr_reader :uri, :cur_offset, :broken_msg
attr_accessor :id