Older blog entries for jimw (starting at number 29)

walters is my hero today for fixing an annoying bug in rsync that results in server processes just sitting there after the client has dropped its connection. if only the upstream maintainer (cough mbp cough) had incorporated it earlier.
rkarl: you mean this?

all caught up. now there's just 4836 notes. just in time for me to leave town for a few days and let the new ones pile up.

time for a nap.

the obsession with coding standards that seems to grip the php developer community absolutely boggles my mind. then again, i also really like perl. but i still don't completely comprehend the '..' operator in a scalar context.

oh, i'm 222 notes behind. there's 5156 total. the ten pages with the most notes make up 12% of that total.

the connectivity between the two php.net machines hosted at pair got fixed (duplex issues), so i'll be able to get back to the news.php.net stuff in the new year.

i've been chipping away at the user notes in the php manual. i was doing okay for a while in keeping up with newly submitted notes, but stopped for a week or two and before i knew it, several hundred had piled up for review. (unfortunately, it appears that zak and i are the only ones regularly keeping on eye on these things.) right now, there's 5797 notes. there were well over 6000 when i started.

i also flipped the switch on the non-subscriber self-moderation for the rest of the php mailing lists. hopefully this will cut down on the spam to the lists without resulting in more list admin work. one spam has already been blocked from php-general. whee!

i was thinking earlier today about how annoying it is that none of the (non-gui) newsreaders i've tried have a decent interface for handling groups from multiple news servers. (none of the gui-based ones i've seen do much better, but i'm really not interested in a gui newsreader to begin with.)

then i remembered that i have the guts of an nntp server written in perl that i was already planning to overhaul to use an sql backend for the overview storage, and realized with some additional elbow grease, it could probably be turned into an nntp proxy, too. it probably wouldn't be a bad idea to leave the door open for using mbox files as the news spool, too.

just what i needed, another project.

curious about why it takes forever to generate the chunked-html version of the php docs (in docbook, using openjade), i did a little profiling of openjade generating the phpdoc howto (a much smaller document than the 1600+ page php documentation).

10% of the time is calls to the assign method of the opensp string template class, 8% to operator= of the ptr template class, and 5% to the string template class's grow method. ugh.

nearly 8% of the execution time is going to reference counting.

i get the feeling my time would be better spent figuring out how to move the phpdoc processing to an xslt processor instead of poking around with openjade.

deekayen: people can get books for free at a library. they still buy lots of them. people can pick up free weekly and daily newspapers all over the world. they still buy lots of them. free counterparts exist for a great many things (and cheaper counterparts exist for nearly everything).

nothing too exciting to report. spit out version 1.0 of colobus when nobody was looking. (it's an nntp server that runs off ezmlm message archives used by news.php.net and nntp.perl.org.) some day i'll get motivated to use a sql backend for the overview database instead of berkeleydb, and build a better web front-end.

for fun, i created rudimentary clones of metafilter and blogger, called conventia and gir. conventia was inspired by discussions on the php-mefi list, and gir by the blogger xml-rpc api (in fact, it doesn't have much of a web interface except to do things not exposed by the xml-rpc api).

buried inside gir is yet another xml-rpc library for php, which currently just implements the server-side. it takes a rather different approach from most xml-rpc libraries. some day i'll finish up the client side, document it properly, and release it as its own project.

so many half-baked projects, so little motivation.

20 older entries...

New Advogato Features

New HTML Parser: The long-awaited libxml2 based HTML parser code is live. It needs further work but already handles most markup better than the original parser.

Keep up with the latest Advogato features by reading the Advogato status blog.

If you're a C programmer with some spare time, take a look at the mod_virgule project page and help us with one of the tasks on the ToDo list!