Age | Commit message (Collapse) | Author |
|
- Prevent correction of delayed messages
- Prevent correction of messages by someone else in a MUC (and in a
private tab)
- Messages with unauthorized corrections (above) or wrong message id
will be displayed as normal messages
TODO: restrict the corrections to the same fullJID (only in direct
"normal" conversations, because we can know in private an muc tabs, via
the User object)
|
|
|
|
(if the room was not joined)
|
|
(like in weechat)
|
|
|
|
(add a defaultdict to keep the folded state in each group)
|
|
|
|
M-d and M-c become M-D and M-C
|
|
when no message has been sent yet
|
|
|
|
|
|
|
|
This makes it easy to review all the highlights after the separator was
placed, using M-h, M-n, M-n, M-n…
We just add a counter of highlights which is incremented each time there’s
an hl, and set to zero when we reset the separator. We use that counter to
set hl_pos when we scroll to the separator.
|
|
|
|
|
|
|
|
|
|
With a few specific behaviours: When manually opening a conversation with a
bare jid, we open a normal conversation that follows the XEP (locked and
unlocked accordingly). If the user manually opens a conversation with a
fulljid (by selecting a specific resource in the roster, or by specifying a
fulljid to the /message command), we open a special tab that doesn’t follow
the XEP (it is always locked to the same resource, and cannot be unlocked).
When a message is received, unless a special tab has been manually opened by
the other with that specific resource, we always send the messages to a uniq
normal tab, unlocking or locking it according to the XEP.
This means that only one tab can be opened with a given contact, unless the
user specifically chooses to open a special tab for a specific resource.
fixes #2159
|
|
- because it is an heavy operation, and it is useless, since the lines
only change when the width changes.
|
|
|
|
|
|
- The code is more understandable
- The number of iterations may have slightly increased
- Less things are done inside the lock, so the overall experience should
be smoother
|
|
|
|
(while staying backwards-compatible)
|
|
|
|
Do not traceback when we receive a chatstate from someone not in our
roster.
|
|
fixes #2195
|
|
- remove the decline command that is not in the trunk (and mediated
declines are supported nowhere anyway)
- change a bit xhtml-im support
- change the bookmarks management a bit
- Add a verification to avoid crashing when poezio will be launched the
next time
- Fix the (unrelated) bug when setting a jid affiliation
|
|
fix refresh bugs, display the scrolling status, add an info bar, add the
global info win
|
|
|
|
|
|
TODO: change the plugin API to take advantage of this
This change is backwards-compatible (as in “should not break anything”)
|
|
|
|
(thanks, jappix)
|
|
Python recursion sucks, and namedtuple.repr() uses that, so I rewrote
__str__ and __repr__ in order to make it the iterative way instead.
|
|
(the corrected message is way more likely to be recent)
|
|
(the name show is not intuitive, so I used status and status_message)
- The status is send when connecting, and is the same used when joining
rooms
- save_status is true by default, and will make poezio save the status
whenever it changes
- status and status_message are empty by default
|
|
|
|
|
|
|
|
This is now already done in sleekxmpp, so I added a “lang” configuration
option. TODO: use the locale of the system.
|
|
|
|
|
|
Fix /accept for transports.
I think there are more people who use transports than people who have
\2f into a legitimate JID. So I believe this is ok.
|
|
|
|
#2189 wasn’t crashing, but well.
Also fix a crash with the separator and /correct
|
|
Show headline messages in the info buffer, and show error messages (or
other errors) with a special style.
|
|
(should fix the “leak” on /correct, and make it faster)
|
|
|
|
|