| Time |
S |
Nick |
Message |
| 00:00 |
|
prologic |
on which platform? |
| 00:00 |
|
prologic |
whicbh binary did you use? |
| 00:01 |
|
prologic |
oh dear |
| 00:01 |
|
prologic |
I think you may have an arm build |
| 00:01 |
|
prologic |
maybe |
| 00:01 |
|
prologic |
my script is broken :( |
| 00:08 |
|
prologic |
pdurbin okay I uploaded some new binaries => https://github.com/prologic/eris/releases/tag/v1.6.1 |
| 00:09 |
|
prologic |
*fingers crossed* or I'll eat my hat :) |
| 00:10 |
|
prologic |
$ ./eris-Linux-x86_64 -v |
| 00:10 |
|
prologic |
eris v1.6.0-dev (HEAD) |
| 00:10 |
|
prologic |
seems to work for me |
| 00:13 |
|
pdurbin |
prologic: works now. Thanks! |
| 00:16 |
|
prologic |
cool |
| 00:37 |
|
prologic |
pdurbin so just testing it out or? |
| 01:51 |
|
prologic |
bear how about this: [17:51:45] prologic has userhost prologic 49960de5880e8c687434170f6476605b8fe4aeb9a28632c7995cf3ba831d9763 and real name “James Mills” |
| 02:20 |
|
pdurbin |
prologic: you and bear seem to be having so much fun I suppose I have a touch of FOMO :) |
| 02:20 |
|
prologic |
ahaha |
| 02:20 |
|
prologic |
:D |
| 02:27 |
|
bear |
cool - new binary, i'll have to try them out on my linux host |
| 02:27 |
|
bear |
yea, I like that userhost hash |
| 02:30 |
|
pdurbin |
now I'm getting: FATA[0000] error loading tls cert/key pair: open cert.pem: no such file or directory |
| 02:30 |
|
pdurbin |
prologic: can you please update the readme with what I'm supposed to do? |
| 02:30 |
|
prologic |
ahh |
| 02:31 |
|
prologic |
for now just comment out the tls listener in the ircd.yml config |
| 02:31 |
|
prologic |
file me an issue so I can create a PR against it |
| 02:32 |
|
prologic |
but if you want TLS |
| 02:32 |
|
prologic |
do this |
| 02:32 |
|
prologic |
$ go get github.com/prologic/mksslcert |
| 02:32 |
|
prologic |
mksslcert |
| 02:32 |
|
prologic |
and it'll generaet some self-signed certs in $PWD |
| 02:47 |
|
prologic |
pdurbin https://github.com/prologic/eris/blob/master/README.md#quick-start |
| 02:53 |
|
pdurbin |
wow, this thing really works |
| 02:54 |
|
pdurbin |
I'm running an IRC network! |
| 02:56 |
|
prologic |
lol |
| 02:56 |
|
prologic |
what did you expect :P |
| 02:57 |
|
pdurbin |
more problems, I guess |
| 02:57 |
|
prologic |
yeeof little faith :) |
| 02:57 |
|
pdurbin |
Do you provide an init script? |
| 02:59 |
|
prologic |
I could |
| 02:59 |
|
prologic |
but for which system? :) |
| 02:59 |
|
prologic |
feel free to contrib one |
| 02:59 |
|
prologic |
I tend to run all my things via Docker these days tbh |
| 02:59 |
|
pdurbin |
I'm running it on Ubuntu 14.04. |
| 03:00 |
|
prologic |
hmm |
| 03:00 |
|
prologic |
does that still use upstart or systemd? |
| 03:00 |
|
prologic |
hard to keep up :/ |
| 03:02 |
|
pdurbin |
I keep philbot up with supervisord. I was running searchbot (RIP) in a tmux session. I'm not picky. What do you and bear recommend? |
| 03:05 |
|
prologic |
depends on your infra |
| 03:05 |
|
prologic |
really |
| 03:05 |
|
bear |
ignore upstart if you can, the world is moving to systemd |
| 03:05 |
|
prologic |
like I said I just use docker :) |
| 03:06 |
|
bear |
yea, docker is saner |
| 03:06 |
|
prologic |
docker stack deploy -c docker-compose.yml eris |
| 03:06 |
|
prologic |
done :) |
| 03:08 |
|
pdurbin |
Hmm. I don't know if Docker runs on Ubuntu 14.04. |
| 03:09 |
|
prologic |
it does/can :) |
| 03:09 |
|
prologic |
But write up a systemd unit and be done with it :) |
| 03:09 |
|
prologic |
file a PR and add it to contrib/eris.unit |
| 03:09 |
|
prologic |
should be dead simple |
| 03:10 |
|
pdurbin |
meh, for now I just put eris in the background :) |
| 03:10 |
|
prologic |
haha |
| 03:10 |
|
prologic |
you lazy sod :) |
| 03:10 |
|
pdurbin |
the kids are still awake. you'll have to cut me some slack :) |
| 03:11 |
|
pdurbin |
no, cut me some irc! |
| 03:11 |
|
prologic |
I also recommend Palaver as a Mobile (iOS) client btw |
| 03:11 |
|
prologic |
haha |
| 03:29 |
|
pdurbin |
I used IRC999 back when I had an iPhone 4. |
| 03:29 |
|
prologic |
never heard of it :) |
| 03:30 |
|
prologic |
bear https://github.com/prologic/eris/pull/24 |
| 03:30 |
|
prologic |
:D |
| 03:30 |
|
pdurbin |
That was the only iPhone I had. Work paid for it. I'm back on Android. |
| 03:33 |
|
pdurbin |
I love how fault tolerant IRC is. My client (weechat) is perfectly happy to keep attempting to reconnect when my new IRC server is down because I'm messing with it. |
| 03:35 |
|
pdurbin |
prologic: nice to see that there's some Cadmus code now! Are you planning on throwing a release over the wall? With binaries. :) |
| 03:37 |
|
prologic |
pdurbin I am |
| 03:38 |
|
prologic |
soon |
| 03:38 |
|
prologic |
just finishing up porting the old codebase over |
| 03:38 |
|
pdurbin |
It writes to BoltDB? |
| 03:51 |
|
prologic |
nope to files for now |
| 03:51 |
|
prologic |
compatible with irclog2html |
| 04:07 |
|
prologic |
bear ip masks work well |
| 04:08 |
|
prologic |
thanks! |
| 04:08 |
|
prologic |
I can't conirm entirely but I think this implementation works |
| 04:08 |
|
prologic |
generating a rainbow table on sha256 is quite costly from what I've read |
| 04:11 |
|
bear |
i'll install it either later tonight or tomorrow morning and check it out - thanks! |
| 04:17 |
|
prologic |
no worries |
| 04:18 |
|
prologic |
I may as well release a new version then |
| 04:18 |
|
prologic |
at least it makes distributing easier |
| 13:24 |
|
pdurbin |
prologic: ok, in the code I saw "cadmus.db" and an import for https://github.com/asdine/storm |
| 16:23 |
|
|
prologic joined #sourcefu |
| 16:55 |
|
pdurbin |
prologic: what if logging were built in to eris? Is that crazy talk? |
| 16:57 |
|
pdurbin |
bear: I'm curious about your opinion too. |
| 17:12 |
|
pdurbin |
Rather than running Cadmus, I'm considering hacking on philbot so that it can log multiple IRC networks. I just opened https://github.com/pdurbin/philbot/issues/3 |
| 18:13 |
|
prologic |
pdurbin yes cadmus is built on the same codebase (more or less) as Soter |
| 18:14 |
|
prologic |
it uses a bolt db (using storm as an orm layer) to persist things like which channels it should join |
| 18:14 |
|
prologic |
it could also store logs there too I suppose but not for the first version -- just going to port irclogger as-is (with a few minor improvements) so that the output of cadmus is compatible with irclog2html |
| 18:14 |
|
prologic |
then I can think about writing a new storage backend and api and web ui |
| 18:15 |
|
pdurbin |
prologic: I feel like you're from the future. Like I said, I'm thinking about hacking on philbot, which is in Perl, and adding another column to its MySQL database. :) |
| 18:15 |
|
prologic |
fine :) |
| 18:15 |
|
prologic |
there is no advantage hwoever in using a db over the file system |
| 18:15 |
|
prologic |
but as you wish :) |
| 18:15 |
|
prologic |
also cadmus awill support multiple networks once I'm done |
| 18:16 |
|
prologic |
I've already told some of the biggest users of irclogger that itll get replaced with cadmus soon |
| 18:16 |
|
prologic |
and they're main feature request? |
| 18:16 |
|
prologic |
stats on who are the chattiest users :P |
| 18:16 |
|
prologic |
haha |
| 18:17 |
|
pdurbin |
We all know you're the chattiest. Which I like. :) |
| 18:17 |
|
prologic |
lol |
| 18:17 |
|
prologic |
sadly if I produce data it will show that is not true at all :) |
| 18:17 |
|
pdurbin |
prologic: Kenneth Reitz of Python fame just stated talking about Go in #positivepython: https://botbot.me/freenode/positivepython/msg/93936289/ |
| 18:18 |
|
prologic |
nice |
| 18:18 |
|
prologic |
IHMO (and others) Go is the better Python |
| 18:19 |
|
pdurbin |
prologic: don't get me wrong. I'm interested in the future from which you come. I'd like be able to say "just run it in Docker rather than using an init script" at all that. I'm just not there yet. Too busy writing Java in my day job, I guess. |
| 18:21 |
|
pdurbin |
sounds like he's struggling a bit: "extraordinarily difficult", "I feel like a total green programmer", "something that would take me literally six minutes in python is taken me a week so far" |
| 18:21 |
|
prologic |
oh my |
| 18:22 |
|
pdurbin |
and* all that |
| 18:22 |
|
prologic |
I always believe the no. one thing to do when learning a new language |
| 18:22 |
|
prologic |
is drop everything you know about all pevious languages |
| 18:22 |
|
prologic |
they almost never translate very well |
| 18:22 |
|
pdurbin |
makes sense |
| 18:22 |
|
prologic |
you'll have a better time |
| 18:26 |
|
pdurbin |
I guess I need to decide if I want to keep logging IRC channels into MySQL or not. I suppose I *could* dump all the data out into a series of files that compatible with irclog2html (never heard of it) and Cadmus. |
| 18:26 |
|
pdurbin |
looks like it's written in Python: https://mg.pov.lt/irclog2html/ |
| 18:29 |
|
pdurbin |
I'm a big fan of keeping knowledge in plain text as recommended by The Pragmatic Programmer but at the moment it feel tidier to keep all my IRC logs in a single table in a database. |
| 18:31 |
|
pdurbin |
has a web interface for searching logs, which I use quite a bit. And if I really can't find something, I just grep the whole table as a tsv file, which I produce every night with cron. |
| 18:34 |
|
prologic |
yeah |
| 18:34 |
|
prologic |
I realized the other day when looking through irclogger's archives |
| 18:34 |
|
prologic |
that I have over a decade worth of ir clogs rom various soruces |
| 18:34 |
|
prologic |
of folks that I've been logging for |
| 18:34 |
|
pdurbin |
And you like having all those files? |
| 18:35 |
|
prologic |
doesn't bother me |
| 18:35 |
|
prologic |
I'd hate to lose the data though :) |
| 18:37 |
|
pdurbin |
That's why I back up my IRC log database to a tsv file every night and rsync it over to a friend's server. :) |
| 18:38 |
|
prologic |
yeah this dataset if replicated via rsync too :) |
| 18:38 |
|
prologic |
but... |
| 18:38 |
|
prologic |
cadmus once I'm done with the initiail version |
| 18:38 |
|
prologic |
will become distributed (that's the plan) |
| 18:38 |
|
prologic |
so it will self-replicate its own data |
| 18:39 |
|
pdurbin |
nice |
| 18:39 |
|
prologic |
and try to maintain connections to the network from N different servers where N is the cluster size |
| 18:39 |
|
prologic |
either 3 or 5 |
| 18:39 |
|
prologic |
why? |
| 18:39 |
|
prologic |
it means its much harder to suffer from a netsplit this way |
| 18:52 |
|
pdurbin |
sounds quite fancy |
| 18:54 |
|
prologic |
Not really 🤣 |
| 18:54 |
|
prologic |
Consensus 😀 |
| 18:55 |
|
pdurbin |
I have all kinds of thoughts on this IRC logging experiment I've been running for almost six years. Before I start running my own IRC server and figure out how to log it, I should probably write up what I'm thinking. My free time is very limited and I want to make the most of it. |
| 18:57 |
|
pdurbin |
I can certainly imagine running Eris and Cadmus on the side just to play a bit, especially to learn Docker better. I don't know that it's the best use of my time to switch away from philbot, which has been rock solid and provides a useful service to me. |
| 18:59 |
|
prologic |
Tell me what’s the one feature that I would need to implement to convince you to switch away? |
| 19:00 |
|
* prologic |
curio |
| 19:09 |
|
pdurbin |
prologic: for now I'm only interested in an IRC logging bot that logs to a MySQL database with a table like this one, for backward compatibility with my existing data: https://github.com/pdurbin/philbot/blob/f096d72c07780603fb49306c1346b0c39b27a177/ilbot.sql#L25 |
| 19:28 |
|
prologic |
pdurbin: well first let me finish the file storage backend and then I’ll write a SQL backend that is compatible with your philbot |
| 20:11 |
|
bear |
ircd should never log channels - it's a privacy thing IMO |
| 20:11 |
|
bear |
if someone (or a channel) wants logs then they can either use a client that logs or a bot |
| 20:11 |
|
bear |
and yes, i'm very firmly in the flatfile camp for everything - including things that could go in a database |
| 20:12 |
|
prologic |
👍 |
| 20:13 |
|
bear |
in the 30+ years i've been doing professional work - i've never not been able to convert flat files |
| 20:14 |
|
prologic |
🤣 |
| 20:14 |
|
prologic |
Same |
| 21:56 |
|
pdurbin |
prologic: please take your time writing that SQL backend :) |
| 21:57 |
|
pdurbin |
bear: everything? Oh, right, I met you in #indiewebcamp where people aren't so excited about databases. :) |
| 21:58 |
|
prologic |
sure |
| 21:58 |
|
prologic |
you know I won't :) |
| 21:58 |
|
bear |
don't get me wrong - databases are great when you need indexes and searches |
| 21:58 |
|
prologic |
but a file system is also a freak'n dtabase :) |
| 21:58 |
|
bear |
heck, i'll even use a database to store markers for the log flat files |
| 21:59 |
|
bear |
it's not that it's a database that causes me to pause, but it's the maintenance tax for files that will be around forever |
| 22:00 |
|
prologic |
yeah |
| 22:00 |
|
prologic |
managing MySQL/PostgresSQL/whatever is a lot more work |
| 22:00 |
|
pdurbin |
I just remember a question Ted Ts'o got from an audience. "How can I make this filesystem work well with millions of tiny files?" "Have you considered using a database instead?" Something like that. :) |
| 22:01 |
|
prologic |
well |
| 22:01 |
|
prologic |
the file system can hdnele this actually |
| 22:01 |
|
prologic |
I once wrote a GIS backend in Python that was backed by millions of tiny tiles |
| 22:01 |
|
prologic |
but I also had indexes to lookup things quickly |
| 22:01 |
|
prologic |
which were also files |
| 22:02 |
|
bear |
newer filesystems have less trouble with 10000's of files |
| 22:02 |
|
prologic |
it was NFS that didn't like it :) |
| 22:02 |
|
bear |
yes, the older ones would slow to a crawl with 1000 dirs of 1000 files each |
| 22:02 |
|
prologic |
trying to backup/copy/etc large numbers of files with NFS is well um a recipe for disaster |
| 22:02 |
|
bear |
we ran into that problem at Seesmic with our video files - had to partition them a lot |
| 22:02 |
|
prologic |
many NFS server/client(s) just b0rk at that |
| 22:03 |
|
bear |
s/trying to backup/copy/etc large numbers of files with// <-- fixed that for ya ;) |
| 22:03 |
|
prologic |
:D |
| 22:03 |
|
prologic |
laying with a laptop is not good for typing corrrectly :) |
| 22:03 |
|
prologic |
also bear (different topic) |
| 22:03 |
|
prologic |
https://github.com/prologic/eris/issues/31 |
| 22:04 |
|
* bear |
looks |
| 22:04 |
|
prologic |
seems the whole sha256 of the entire hostmask was not well recieved :) |
| 22:04 |
|
prologic |
received* |
| 22:04 |
|
prologic |
seems we can improve on this somewhat |
| 22:05 |
|
bear |
:) |
