incorrect emotification

Isn't that how it works in other messengers? In most of the messengers I've used, no matter where the :/ is located in the message it is converted to an emoticon since it's hard to tell where people do and don't want it to do so...

I'm not sure how the algorithm could know that you weren't saying something like:

that's the topic:\

or perhaps you wanted to send that face with foo after it to your friend :\foo

Lots of people I know, don't put spaces around their emoticons, so I would think the current behaviour is correct. What do others think?

I think it would be nicer to require spaces around smilies... that just seems like it would make more sense.

Just my 2¢.

I have no idea what other clients do, I've been using Adium exclusively since .53 or so. :) But I agree with greatcaffeine that it makes more sense to require space delimiting. What's the alternative? If someone needs to IM me a string the includes ":/" how can it be prevented from becoming emotified? Does Adium have some sort of <code> tags? Yes, I realize I could set emoticons to "none", and that's what I normally do. That seems overkill for an occasional IM. Furthermore, the current method interferes with URL linkification, which is how I discovered this. I think space delimiting can stand on it's own merits, but since you may not be convinced of that I'll add that (IMHO) URL linkification is more important than providing emotification anywhere that a sender might possibly want it.

Adium has the best emoticon determination algorithm of any client I've seen; you've found an edge case.

If you type c:\windows you'll find that the :\ is not turned into an emoticon. However, c:\ is is. The reason is that an emoticon at the end of the message is always turned into its graphical form... because people send messages like "Hello:)" and intend an emoticon fairly often. Similarly, ":(That makes me sad" is a reasonable thing to send.. but if you sent - (void)processString:(NSString *)inString you wouldn't see the graphic, as that :( is inside the message without spaces on either side.

Could you provide the specific message which interferes with linkification?

I don't doubt that Adium is the best at most everything.

I also wouldn't be surprised if this ticket is about something that few people would notice or care about. I suspect most of my tickets are like that. :)

I didn't do extensive enough testing to see that you are correct about how well Adium handles the examples you provided. Sorry about that, I do try to be thorough.

After looking at my logs for the chat in which I received the non-linked url and the IRC chat in which this was discussed, I see that my linkification claim was in error. I discovered the topic of this ticket while trying to figure out why http://x1024.net:/blog/ didn't get linkified, so my sleep-deprived mind conflated the two. My apologies.

I just discovered that "O=C1C2=C(N=CN2C)N(C(=O)N1C)C" gets emotified. Specifically, the "=O". But maybe Adium intentionally puts that in the middle of my caffeine molecule - it seems strangely appropriate. :) Or maybe it's just something more mundane like different handling for text inside of parentheses.

Matthew, the length of this ticket is confusing me a little. ;)
What exactly is wrong with emotification in Adium?

This ticket was closed automatically by the system. It was previously set to a Pending status and hasn't been updated within 14 days.

I'm not sure when it was fixed, but as of r23902 the "=O" in "O=C1C2=C(N=CN2C)N(C(=O)N1C)C" no longer gets emotified.

I'm not sure when it regressed, but as of r25087 the "=O" in "O=C1C2=C(N=CN2C)N(C(=O)N1C)C" no gets emotified again.

"Regression"...

I don't think that's an edge case we can appropriately address. Text is sent far more often than caffeine molecular representations in Adium, and the message

I see you(:))!

should emotify the :).

Can anybody propose (in words or pseudocode) an algorithm which could properly determine when an emoticon surrounded by parenthesis should be emotified?

Replying to evands:

I don't think that's an edge case we can appropriately address.

Personally, I'd prefer to eliminate false positives at the expense of false negatives, but I realize that most other people like graphical emoticons more than I do.

Can anybody propose (in words or pseudocode) an algorithm which could properly determine when an emoticon surrounded by parenthesis should be emotified?

I can't.

Whichever way this goes, it is still problematic that the emotification seems inconsistent. I don't think I've mentioned it before, but I didn't notice it until after I'd forgotten about this ticket. Sometimes my caffeine molecule gets emotified in my chat window, and sometimes it does not. I've not been able to identify any correlations.

Patches or suggestions welcome. I don't have a solution to this problem.