Older blog entries for LSchiere (starting at number 14)

gaim
faceprint and i decided that his xml blist work and my person work overlap enough it doesn't make sense to be working on 2 differnet patches, so he set up a cvs server (neither of us have access to the one on sf.net), and we've merged our work. several steps backward for a long-er term net gain. currently having trouble getting people to show up in the right group, not sure why yet. also, i put the edit tree back in, but it doesn't do anything but sit there right now. it shows groups and the persons in them.

Gaim:
After a day at work, spent the afternoon with the family, then worked on a patch KingAnt was having trouble with, between us, most of the effort on his side, I think we've squashed the group renaming bug. Problem was that when you renamed a group, the change wasn't uploaded to the server. Kingant squashed that easily, but doing so introduced some other issues, basically, it came down to calling functions in the right order. Given how much I've been working with the buddy list code on the core side of things as it is, I picked it up for him... :-)
so hopefully JSeymour can commit that tomorrow, and one more thing can drop off the TODO list. Hopefully I'll make some significant progress on the edit code for the blist with person support also. of course this patch adds in an additional function to have edit support into, but that's life for you.

bytesplit: obscene and/or profane ranting is never acceptable. In grade school I was taught that "two wrongs do not make a right," you would do well to take note of that. You might also take note of the fact that people have only "instigated" your behavior after you have over-reacted, or simply reacted inappopriately to comments others ment to be helpfull, or at very least, innocent. By and large, I've watched this community attempt to help you become a productive member, to channel your energy in a constructive manner, but you continue to lash out. Perrier gave a well-reasoned, calm, logical statement of his opinion, supported by concrete assertions. Your reply is generalized, vague, and inflamatory. I believe you are mistaking the optimism and hope for change of some members for friendship, and believe that the diaries of far more people express disgust and dislike of your behavior than express even the mildest sort of support or tollerance.

bytesplit: ChipX86 stalk you? talk about an off the wall, and thoughtless assertion. who's been on advogato longer? chipx86. who was in the various chat rooms you visited him on opn longer? again, chipx86. its clear to anyone familiar with this situation who's been stalking who, so stop lying through your teeth. As for the maturity level, again, ChipX86 comes out clearly as the more mature. You are the one constantly ranting about everyone who you percieve to have offended your delecate ego, while ChipX86 on the other hand has even tried to put the past behind him and help you. Any sensible person would have done the same instead of dredging up the past with each new interaction.

djcb there's a reason your leak fix hasn't been applied. only 3 people make commits to gaim: our maintainer, rob flynn, is very inactive, about the only thing he reliably does is control when releases are; Sean Egan, our lead developer, is down with a broken computer; and Jim Seymour, who's been given commit access solely to facilitate his work on the jabber plugin. Eric and Mid still have commit access but are retired and chipx86 has commit access, but his job is keeping our web page in good shape, not dealing with the code base. Once Sean gets his computer fixed, your patch will be looked at and most likely applied. if its not, you'll be told why.

Gaim update: so it's now possible to write to a person instead of a specific buddy, a big improvement, only issue is that it might cause some confusion if you have two buddies with the same screenname on different services who aren't the same person, but have it set to merge on screenname anyway. If you merge on same alias in same group, which I suspect will be the most common setting, then that issue shouldn't be a problem. Probly need to get rid of the send_as menu, i don't think its doing anything at this point, but i'll leave it till i know ofr sure. now i'm working on re-enabling editing of the blist. this is tricky, cause i don't want to use a 3 level tree, i think that's ughly and more error prone.

bytesplit: joining a community is an implicent agreement to abide by the rules of that community, and the rules for certification, as other have noted, explicietly state that the certification is to be primarily based on the person's contributions to the code base. It gives some recognition to web designers, documentation writters, and the other people who make successfull projects possible, but the primary qualification for high certification levels is the amount and quality of the code a given person produces. Your personal dislike of chipx86 and perception that he is unfriendly is irrelevent on those grounds. Similarly, your own certification of journeyer seems like a perfect example of the certification issues that raph was talking about earlier. As for open source projects, specifically the variety of overlaping and repetative programs out there, if you'd ever been around for a editor flame war, you'd very quickly notice that about the only things that people DO agree on is that a program should be relatively stable. Some people like emacs, others vi, others one of the myriad of other editors out there, and no one can agree on exactly what a quality editor should have; and alot of people do agree that one of the prerequisites of a quality editor is that it doesn't take alot of memory, so even trying to meet all the other features that people want in editors wouldn't satisfy because the resultant size would be overly massive. As to programs like Office2000, you see projects like gnumeric, abiword, koffice, openoffice, and others working to meet that need. and surprisingly, you DON'T see alot of the overlap that you complain about in editors, just a minimal amount of overlap where people can't decide if they want a single monolithic application (openoffice), a gtk app, or a qt app. Personally, i think its a great thing that the open source world lets you customize your enviroment to best match your own thought patterns, it wouldn't be good if we tryied to force everyone to match a single window manager, single editor, single gui toolkit, so on and so forth. people are just too different. With games, you have projects like wine and winex working to bring your windows games to the unix desktop. instead of complaining, you might try contributing to one of those projects, or you might join one of the projects that tries to code a cool game specifically for unix. Almost all of us have jobs or school work outside of open source programming, meaning we're coding in our free time, time taken from other aspects of our lives. When you start bashing someone for failing to meet your expectations when they aren't being paid and are coding at all to meet some personal need or for fun, developers start thinking about quitting: the work stops being fun, and meeting the minimal need instead of perfecting the app starts to seem like the way to go.

so. write_person() is TOTALLY broken, it sends to the wrong buddy, to no buddy, and the debug output is freaky, and i'm ont sure what's up. not sure where all these people with null and "" names are coming from, esp. when they work in every other function.... *sigh* if this project were trying to teach me humility, it couldn't be doing much better. i've pretty much realized i don't have to change conversation.c except to replace calls to serve_send_im with write_person, but it would help if i could figure out why write_person isn't working. on top of that, a friend who was helping me look at it, finds segfaults that i don't get left and right. why i manage to avoid all the segfault conditions i'm not sure, but at least segfaults are easier to find and fix than screwwy results.

I added in person support to the core last night, after a little bit of trouble with find_buddy, i managed to get things in a nice transitional state. next big step is to make a struct conversation be assoicated with a person/person_show (not sure which is best yet. *grumbles* conversation.c is NOT core/ui split) so that i can remove the gc from it, and use write_person so that you'll automatically send to the person with the highest priority. once i've got that done, i can finish stripping out alot of the gaim_connections from slist.c (replaces buddy.c), and start to add the edit_buddy code back in. adding that back in will take alot of thought. turns out i can't drag and drop a GtkTree the way i can a GtkCtree, so i'm not sure right now what i'm going to do. i do like the idea of edit_mode better than edit_tab, but if i have to use 2 different trees.... we'll see.

UGH. so there is an idiot on my subnet at work. Someone in one of the other labs gave 2 machines the same ip address, and I've been getting arp reports about it since yesterday. Sent an email out to get it fixed, but as of 9:30est, hasn't been yet. Pain in the neck. Unfortunately, I have to run arp because scientists have the unfortunate tendency to do things like change their ip number randomly with some frequency, but as support people, we are SUPPOSED to be watching for that and to fix it asap. *sigh* and THIS is why i managed to convince the higherups to move most of the work stations to a dhcp solution. Oh, and for some reason i can't seem to get the i and em tags to close here.