I've been waiting for MySQL to crash and it finally did! But sadly, I can't find its core file anywhere.... core file... where are you? I have to send you to monty to have the db fixed!
fucking with mysql all night:
-- compiled another version, with more/different debugging stuff
-- ready to catch another core! (better work this time)
-- took snapshot of database, rsync to my home machine now (and for the next 4 hours... 2.5 gig)
-- turned off regular logging, turned on binlogging... on separate disk
-- ready now for replication
This go on dev list, since I'm not sure if it'd get anywhere on lj_biz...
LiveJournal seems a bit ... bloated? Maybe? Approaching that state? This might just be because I'm a minimalist who drools over useless embedded products all day, but it might be true if someone out there agrees.
So, is it possible to make a "livejournal lite" mode, or something similar? Where most of the options someone just looking for a vanilla journal doesn't need would disappear?
Dunno. I'm looking at blog, seeing what goes into it, looking at livejournal, and getting momentarily overwhelmed. It's also the argument I've heard five times in the last ten minutes by some blog people.
Just a passing thought...
Oh, and I'll have a writeup of the proposal process tomorrow. That goes on lj_biz :)
news.gnome.org is down, or something. so the syndication for it was making the script time out and crash. Anyone know how to make XML::RSS not die when it's passed no data?
The email one receives when another user comments in their journal or replies to a comment that user made, it currently contains the URL to view the entire discussion. I think it would be nice if it included (along side, or instead of the other URL) a URL to the reply form as well. Saves on clicks, page hits, and db queries to get from point A to point B that way.
I am creating a LiveJournal library for PHP.
Which is better to do in the case of an error (i.e. non 200 reply from the server, or success == "FAIL"): call die()... or store the error and make the user of the library check to see if an error occurred after each function call, therefore allowing them to handle errors in thier own way?