Older blog entries for tausq (starting at number 23)

heard from a few people about my previous diary entry, which made it into the salon article. it's good to know people still care despite all the ranting that goes on sometimes (some of which i've mentioned in previous postings here).


got my newpkgs script working again finally. stupid typos..... i need to figure out a better way to get that info. right now it is very hackish.... it's one of those projects i keep on telling myself i'd get to, but never do. *sigh*

spent some time working with joey and culus yesterday to speed up dpkg-preconfigure. found out that forking is terribly expensive, despite some things i have previously read. we ended up speeding it up about 3x with inprocess unpacking of the config/template stuff, but the program that is used to do it is about 36k after various optimizations. g++ really sucks when it comes to dealing with templates. culus has an idea that we should just make a non-templatized string class and link all our stuff against it. maybe i'll give that a shot this weekend.

Today is a sad day for Debian. It is odd how you can feel you know someone as a friend, and yet have never met that person in the flesh. To our friend, who will remain unnamed for a few more days... you shall always be remembered as a part of what some people (quite aptly, perhaps) call the most dysfunctional family there can be. We may quarrel, but we all share a passion for the cause. We shall all miss you.

Long weekends are awesome, although I tend to just end up sitting home and not doing much. Saw _The Patriot_ the other day. It was all right, but at some points it seems to drag on and on. _Independence Day_ was definitely a better Jul4 movie.

Anthony helped me set up xcin so now I can practise typing in Chinese. Figure it's a good way to (re-)learn mandarin. I am really interested in learning all sorts of new languages, but without using it on a day to day basis it's so easy to forget.

gtk capt continues to go well. I checked out stormpkg yesterday. They have some interesting ideas which ltd and I hope to incorporate into capt. My biggest complaint for stormpkg is it's huge memory usage (15-18M). It's great though to see that there are finally some nice packaging tool frontends available for Debian. More experienced users will probably stick to the command-line tools, but the UI's definitely have their place.

Related to this, I finally managed to complete my build-dependency patches for apt. Now if Culus would just release it.....

the gtk capt work is going well. i think we can probably release it in a couple more weeks. there are still a lot of rough edges, but most of them should be easy to fix.... hm..

my employer makes people sign these waivers to work on opensource projects. i finally did one today. really wonder how legally enforcable their restrictions are. it really irks me that a company can regulate what i do on my free time though.

seems like i might be heading back to cornell in september. that should be fun.

Starting a overhaul of capt's UI subsystem with ltd's help. Basically we want to modularize it into a class so that capt can be more extensible. All things considered I think it's going well. There's still a lot of work to be done, but amazingly enough I got it to compile again after breaking it so much this morning.

hrm, been a while. Mostly have been busy with stuff at work. On a positive note it looks like lully is finally back (knock on wood). We have the RAID up (25G) and mirror sync is in progress. Seems like that'll finish in another 12-15 hours or so. Gawd debian is really huge :-) For those interested, this is syncing from sourceforge at 200-300kB/s. Not sure if lully's bandwidth qualifies it for a tier-1 Debian mirror, but it's certainly not bad.

Trying to fix the db.debian.org security problem joey reported. Stripping tags seem to be the easiest, but it doesn't solve all problems. Unfortunately, "solving all problems" in the context of supporting all browsers (i.e. no cookies, no javascript) is rather difficult. (At least when your site uses multiple scripts....) The only thing I can think of right now is do redirects on the server side -- i.e. have one script run the other script and pass query strings to it via stdin. Sounds kludgy but should be perfectly safe. Unfortunately I'll be gone for a couple of days starting tomorrow, so I guess i'll just implement the tag-stripping for now.

Oh, and Shaleh, if you are looking for ugly Makefiles, look no further than boot-floppies! :/

ok, more rant time :)

Someone was complaining on the Debian list that he was caught by surprised by the first test cycle, which starts today. I am deeply bothered that people have time to read and respond to flamewars that are several tens of messages long, and can't be bothered to read about release cycles that were posted both to mailing lists and in Joey's weekly newsletter. What does this mean, really?

Still no word from TLS yet. grrr.....

Looks like RedHat decided to redirect the www.cygnus.com site to www.redhat.com. Shame on them.... :-(

Just for fun, I decided to set up an oracle database on one of my computers. Ended up taking most of the day though, because the Oracle installer is so broken on Linux...

Case in point: I went around looking for the sqlplus binary, and all there is in the bin directory was a sqlplus directory. The installer insists that sqlplus is already installed and refuses to let me reinstall it. So I redid the install from the beginning, and when I watched it in another window, I saw it create the sqlplus executable, and then overwrite it with a sqlplus directory. What kind of junk is this?! I've done several Oracle installs on other platforms (albeit an older version) and the installers were not nearly as crappy.

welp, it looks like the root cause of the modconf bug is actually not in modconf itself.... i was wondering why things worked in my testrig (chroot jail) but not when tested in actual potato boot-floppies. I guess if I had set up my test environment more carefully I would have found this sooner.... ah well. Now the bugs are filed, and things that I could fix are fixed, this time I'm pretty confident we've licked the bug for good.

Haven't been in the best of moods likely, I seem to be ticking people off with some of my more terse emails. <sigh>

14 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!