*** pitillo has joined #crux-devel | 05:54 | |
tnut | Morning | 08:30 |
---|---|---|
*** tnut has quit IRC | 08:44 | |
Romster | evening then quits :D | 10:04 |
Romster | you don't remember yesterday where you broke stuff with libffi frinnst :D | 10:05 |
Romster | i've held off on libffi still | 10:06 |
Romster | works been to busy and i haven't looked at glslang yet | 10:06 |
frinnst | sure I do. I sent out a big fat warning too | 10:14 |
crash_ | yeah i noticed that warning :) | 12:01 |
*** pedja has joined #crux-devel | 12:19 | |
*** chinarulezzz has joined #crux-devel | 12:47 | |
jaeger | same | 17:46 |
*** chinarulezzz has quit IRC | 18:13 | |
ryuo | jaeger: thoughts on the proposed DB format? it seems the easiest way to extend the existing one. | 19:04 |
ryuo | though i'd need to use a new name for it. | 19:05 |
ryuo | something like dbe or something. | 19:05 |
ryuo | data base extended | 19:05 |
pedja | dbng :) | 19:05 |
ryuo | lol | 19:05 |
ryuo | i was thinking of making this a hybrid of the debian approach and crux approach/ | 19:06 |
ryuo | debian's approach allows a lot of flexibility so that's probably why it's still in use 20 years later. | 19:06 |
pedja | debian packaging looks weird to me, tbh :) | 19:07 |
pedja | never actually *used* IT, thou, so | 19:08 |
pedja | plan to add metadata, too? | 19:09 |
pedja | I know json is the new coolness, but how about yaml? looks more readable/easier to use to me | 19:11 |
jaeger | I don't have a strong feeling about it. IT would be easy to use | 19:14 |
pedja | afaik, when AWS CloudFormation added yaml support, there was much rejoicing :) | 19:21 |
pedja | jaeger, plan to buy some new toys on Black Friday? | 19:25 |
pedja | here, most of the BF-like deals are kind of crap, so people will go abroad :) | 19:27 |
pedja | and local customs are well prepared to take their share... | 19:28 |
pedja | sorry, OT | 19:29 |
ryuo | jaeger: ok. i'll give it a try at some point. | 19:30 |
ryuo | pedja: well, i'm trying to keep from adding a ton of dependencies to pkgutils. | 19:30 |
ryuo | the DB is mostly automatically generated. | 19:31 |
pedja | from the contents of Pkgfile or some new .metadata file? | 19:32 |
ryuo | pedja: neither. all the metadata is extracted from the package's file name and contents. | 19:34 |
pedja | ah, OK | 19:34 |
ryuo | i'm proposing to expand what the db can hold for now. | 19:34 |
ryuo | and worry about adding to it later. | 19:34 |
ryuo | named fields makes it possible to extend it indefinitely. | 19:35 |
*** chinarulezzz has joined #crux-devel | 19:35 | |
ryuo | currently they're purely positional so expanding is pretty much impossible. | 19:35 |
pedja | MVP first, in modern talk :) | 19:35 |
ryuo | i was considering adding some extra fields to it that are free, mostly install times and such. | 19:36 |
ryuo | that can be provied by calls to the system clock. | 19:36 |
ryuo | indeed. i have a pretty good idea of what i'm going to do now. | 19:37 |
ryuo | i'm planning to stop using printf when i can afford to do so. the overhead of using it in performance critical areas is high. | 19:38 |
pedja | can reverse deps be inferred that way? from the generated metadata? | 19:45 |
pedja | I know there is revdep, obviously | 19:47 |
pedja | but that always looked to me as a brute-force approach, out of necessity. | 19:49 |
ryuo | pedja: not really? revdep is only for finding missing libraries. | 20:40 |
jaeger | pedja: no plans to. I just got a new desk and usually black friday sucks for PC stuff | 21:10 |
*** chinarulezzz has quit IRC | 21:33 | |
*** pedja has quit IRC | 22:49 | |
*** chinarul2zzz has joined #crux-devel | 23:30 |
Generated by irclog2html.py 2.14.0 by Marius Gedminas - find it at mg.pov.lt!