IRCaBot 2.1.0
GPLv3 © acetone, 2021-2022
#saltr
/2024/05/23
RN Hey everyone! I made cookies! Please help yourself: i2peek-a-boo.i2p/imadecookies.png NO, NOT BROWSER COOKIES!
RN dr|z3d, you inspired my cookies!
zzz put up my take on graph refresh, please review/test git.idk.i2p/i2p-hackers/i2p.i2p/-/merge_requests/199
zzz nice detail on those cookies RN
dr|z3d aww, cute, RN.
zzz *** test test test ***
dr|z3d bizzy zzz iz bizzy.
zzz *** if you see this let me know ***
zzz didn't take much from yours other than inspiration, but was still helpful
dr|z3d great. the main thing is you've got rid of that nasty meta refresh.
dr|z3d if you want to go the full distance, you might also want to use graphSingle.js as well.
dr|z3d (for inspiration, if nothing else)
zzz don't think I want refresh there?
dr|z3d that's up to you, but it's not just for refresh.
dr|z3d instead of reloading the entire page when you tweak the parameters, it handles that with ajax so you get a smooth update.
dr|z3d there's also some work in progress code for single graphs to add a show/hide legend toggle independently of the global config.
zzz then how does the POST get thru to save the config?
dr|z3d mostly works, not quite right yet when you change the graph params, however.
dr|z3d that's the thing, you don't want to save the config, you want it as a transient value that doesn't mess with the globals.
zzz so now you have no way to change the defaults
dr|z3d this is for the single graph only.
dr|z3d globals are configured from the main page.
zzz oh ok, makes more sense
dr|z3d you don't want a legend generally, but maybe when you view a single graph you might want to pull up the legend without jumping back to main and toggling the setting there. that's the thought process.
RN thanks zzz. The frosting was a real challenge to get right. I ate the ones that didn't look right
dr|z3d *if you don't want
dr|z3d I wasn't sure if you baked them or stable diffused them, RN :)
RN and that is exactly what I was going for dr. ;)
dr|z3d the other things graphs.js does in + is set the height of the graphs based on the first displayed graph, zzz.
dr|z3d just in case there's variation in height.
dr|z3d or rather, it sets the height of the containers.
zzz not really seeing any sizing issues here
dr|z3d and it also hides the settings section and adds an on-click toggle.
dr|z3d the sizing issue appears with legends enabled when there are several restarts logged. then the heights can get out of sync.
zzz ok will keep an eye out for that
dr|z3d speaking of legends, the number of logged events could probably be capped, haven't got around to fixing that yet.
dr|z3d things start to look a bit ridiculous when you've got a 48h period with 30 restarts logged.
zzz I took them out for vmCommSystem for that reason
dr|z3d the lines on the graph could persist, but maybe we could cap the number of events printed to the legend to something sane. 3? 5?
zzz sure
Unbur im still having a lot of connection problems, irc is the worst by far on the bright side i2pd works just fine. how are we on the attack?
T3s|4_ he's left, but guessing Unbur is seeing many of IRC connection failures, like these: `irc: connection timeout (message 001 not received)`. Same here, despite having many tunnels and all 'green stars'
zzz dr|z3d, fyi your fix for not loading http blocklist every time was a fail; you're calling compplieRegexPattern twice, once before and once after the last-mod check
dr|z3d zzz: thx, will look into it.
dr|z3d I see it, good catch. Does this mean you're investigating a canon implementation? :)
zzz no, just following up from before
dr|z3d ok, thanks, fixed. only needs to be run once as you say, in the if clause.
dr|z3d if that's all you've got I'm doing pretty well *grin*
zzz *not a comprehensive code review
dr|z3d appreciate you looking at it nonetheless.
dr|z3d something I noticed with the tunnel filter relating to that you may also want to take a look at..
dr|z3d tunnel filter doesn't like empty deny lists. will crap out all tunnels where configured if the deny list is empty.
zzz gitlab ticket please if you're sure it's a canon issue
dr|z3d you can test by creating a deny rule pointing at an empty file, or echo > current_deny_list.txt and watch the fireworks.
zzz and you've done that on canon, or alternatively, did a diff of your code vs canon to verify identical?
dr|z3d haven't changed anything in tunnel filter, so pretty sure you have the same issue.
zzz then stick it all in a ticket please
dr|z3d having a "can't login" moment on git.idk
dr|z3d "the change you requested was rejected".
dr|z3d the chnage being "sign in".
dr|z3d and then it just disables the sign in button. classy.
zzz using https?
dr|z3d I wasn't. let me try that.
zzz if you get in, pls include the tunnel death exception/stack trace
dr|z3d I'll see if I can find it, but chances are it's gone.
dr|z3d yeah, looks like it's disappeared from the logs, sorry.
zzz then point at the offending line, and was it an NPE or what type of exception
dr|z3d *get in
zzz I'm looking at that change and it isn't apparent to me where the problem is, that's why I'm asking for the deets in the ticket, including what type of exception was thrown and from where
dr|z3d this line: throw new IOException("Invalid entry in Tunnel Filter access list (bad b32)", bad32);
dr|z3d that was the error I was seeing. wouldn't accept connections, restart caused server tunnels to fail.
zzz thx
dr|z3d sorry I couldn't provide more info. :|
zzz the reason I need specifics on symptom, not just a link to your "fix" is that it's all over the place and buggy
zzz the if line you added at the top is not necessary (duplicative of the while test) and causes you to lose the first line of the file
dr|z3d I was trying to test for an empty file and returning early if one is found.
zzz yeah but you read the first line and didn't do anything with it
dr|z3d so this doesn't do what I think it should? if ((b32 = reader.readLine()) == null) {return;}
zzz but what if it wasn't null? you still read the line
dr|z3d if it's not null, then hand it over to the while loop.
zzz but the while loop reads the next line
zzz there is no 'hand over'