953752 - Some system messages contain emoticons where they should not.

Status: RESOLVED DUPLICATE of bug 954572

(Instantbird Graveyard :: Conversation, defect)

Description

[Florian Quèze [:florian]]

*** Original post on bio 309 at 2010-01-20 16:57:00 UTC ***

A typical example is when we have in the #instantbird IRC chatroom:
 DGMurdockIII left the room (Quit: get satisfied! ? :: www.unitedservers.de ??? (Gamers.IRC) ??? gamersirc.net ::).

The last character of the message coming from the user client is used with the ")" added by libpurple to create a :) emoticon.
Matching an emoticon in this case should not be possible.
For this specific example, I think we could add a span node around the quit message to prevent the patch, but I don't know if we can generalize this solution to all other system messages generated by libpurple.

Or maybe we should not match emoticons at all in system messages?

Comment 1

[aleth [:aleth]]

Attachment: Patch

*** Original post on bio 309 as attmnt 1125 at 2012-01-19 00:13:00 UTC ***

Another instance (at least with the emoticons I am using) is that the :/ in unlinkified http://... is converted to a smiley, which looks ugly and happens a lot.

This rules out the first proposed solution. 

The drawback of this patch is that it also stops the :) in the #instantbird channel topic from being parsed ;)

Comment 2

[Florian Quèze [:florian]]

*** Original post on bio 309 at 2012-01-19 00:17:35 UTC ***

How come you have unlinkified urls?

The solution I would suggest (which you may not like) is to remove :/ from the emoticon theme.

Comment 3

[aleth [:aleth]]

*** Original post on bio 309 at 2012-01-19 00:27:44 UTC ***

URLs in quit messages are not linkified for example.

I suspect :/ is not that uncommon (it's certainly not unique to this theme). More generally URLs should not be emotified even if they are in quit messages.

If you can suggest a more specific way of doing this, that would be great. A problem is smileNode can't tell what it's parsing is a URL if it's not linkified.

An alternative would be to linkify the URLs... but then that would make system messages stand out more, which is probably as bad as a smiley or two?

Comment 4

[Patrick Cloke [:clokep]]

*** Original post on bio 309 at 2012-01-19 00:29:29 UTC ***

(In reply to comment #3)
> URLs in quit messages are not linkified for example.
> [...]
> An alternative would be to linkify the URLs... but then that would make system
> messages stand out more, which is probably as bad as a smiley or two?

I disagree, I think they should be linkified, I actually can't think of a single use case you'd want things to NOT be linkified. Is there one? If not, we can just ignore that flag.

Comment 5

[aleth [:aleth]]

*** Original post on bio 309 at 2012-01-19 00:53:02 UTC ***

> I disagree, I think they should be linkified, I actually can't think of a
> single use case you'd want things to NOT be linkified. Is there one? If not, we
> can just ignore that flag.

All I can say is that it was the first thing I tried out (by removing http://lxr.instantbird.org/instantbird/source/chat/content/convbrowser.xml#359) and it looked incredibly noisy, as every join message and many quit messages contain something that is then linkified, but that you are pretty much never going to want to click on.

Comment 6

[aleth [:aleth]]

*** Original post on bio 309 at 2012-01-19 19:23:06 UTC ***

(In reply to comment #4 & comment #5)

It strikes me that there is an obvious possible solution here: linkify everything while changing the link color in the message style for system messages to a shade of grey rather than blue, so links in system messages don't jump out at you. 

This is getting off-topic, but: if one were to do this, the IP in IRC join messages should probably still not be linkified (currently if you make the change mentioned in comment #5 they get interpreted as email addresses I think).

Comment 7

[aleth [:aleth]]

Comment on attachment 8352868 [details] [diff] [review]
Patch

*** Original change on bio 309 attmnt 1125 at 2012-01-19 19:25:50 UTC was without comment, so any subsequent comment numbers will be shifted ***

Comment 8

[aleth [:aleth]]

*** Original post on bio 309 at 2012-07-15 08:51:39 UTC ***

This is a subset of bug 954572 (bio 1139), so resolving.