Alert list

[MrDoh]

I took a quick look inside the client code, and the filters.conf file. I'm curious, would it be better (quicker, use less memory, processor etc.) to go the apparant route that seq and put zone names inside the .conf, and only loading that part of the alert list (upon zone in or map load) to check every client specificed update time. I guess currently, it only loads the file once (unless you tell it to reload it) and then checks the entire list read in from the file against the entire zonelist (which in itself can be quite long itself.) Is that how it actally works? (i.e., did I read the code correctly?? hehe)

I don't know how large a file is too large to have to go through on a 250ms (default example) basis. I looked through the latest seq filters.conf, and it's 20k. I also noticed that there was several mobs that were not in the file that I would want to add, so that would add to the file size as well.

If this isn't an issue (problem), or won't be in the future, let me know so I can get it out of my head!

Mr. Doh

[Cryonic]

SEQ's method of filtering for alerts is to look for a <zonename>.conf file and if failing to find that, to fall back to the default filter file. The default filter is used fully and is not broken up by SEQ based on comments in the file.

[MrDoh]

Thank you for the correction!

I've never actually used SEQ, so I mistook comments for parsing information.

Mr. Doh

[cavemanbob]

I suppose either method could be done without too much trouble, looking for individual zone files would likely be easier though. I doubt there would much of a performance gain though as the alerts list is only scanned when a new mob is added, but it would be easier to edit.

[lane]

Along this line of chat... I was trying to add some alerts and found that I needed to add an "_" between most of the words (I can't remember not having to add, but still say most since I can't confirm right now). This is also the way they display in the list.

I don't know the format of the mobs while in memory, so I don't know if this is normal, but the SEQ filters were always normal naming "MobFirst MobSecond" with a space.

If we could get the way you hold the names the same (i.e. no underscores) we could potentially use the seq filter(s).

-Lane

[cavemanbob]

Yeah all that needs to be done is to pass the test through a function that replaces spaces with underscores. I'll look at the seq filter/alert code and see how difficult it'd be to migrate to something closer to the way it works at some point.