Perl 6 - the future is here, just unevenly distributed

IRC log for #sdl, 2015-03-13

| Channels | #sdl index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
02:39 kaare joined #sdl
06:57 dod joined #sdl
07:03 dod joined #sdl
07:25 FROGGS joined #sdl
08:08 basiliscos joined #sdl
10:30 Mithaldu holy balls you guys, not mentioning in SDLx::App that the move handler primarily dispatches partial movies is *amazingly* cruel
10:30 Mithaldu here i was wondering why the damn game was being skippy at high FPS
10:31 Mithaldu FROGGS: ^^^^ >:|
10:34 FROGGS ?
10:34 Mithaldu what's unclear about this?
10:34 FROGGS your attitude
10:35 Mithaldu no i think *that* is fairly clear
10:35 Mithaldu my attitude is fairly damn grumbly because SDLx::App makes people waste time due to being incompletely documented
10:36 Mithaldu what's unclear is what actual question is hiding behind your "?"
10:38 FROGGS I am just puzzled, that's all
10:38 Mithaldu about what?
10:38 FROGGS <FROGGS> your attitude
10:38 Mithaldu please help me help you by telling me what puzzles you
10:38 Mithaldu that's not an explanation
10:38 Mithaldu why is it puzzling that i'm grumbly?
10:38 FROGGS if you have a problem please open an issue or discuss it here nicely instead or shouting at me
10:39 Mithaldu i'm not shouting
10:39 FROGGS well, felt like
10:39 Mithaldu also, why open issues if you don't plan to make more releases anyhow?
10:39 FROGGS I do make releases, what makes you think otherwise?
10:40 Mithaldu your comments a year or so back about working on sdl for perl6 now, also this: May 14, 2014
10:40 FROGGS I just dont put several hours a week into SDL development
10:41 FROGGS I think I said that if i'm going to work on SDL2 it would be for Perl 6 or some such
10:41 Mithaldu i could grep my logs, but at one point you straight-up said you weren't gonna touch SDL anymore :P
10:41 FROGGS but yeah, SDL is not my one and only lovechild anymore
10:41 Mithaldu i mean, great to hear that that's not the case
10:41 FROGGS I will touch it if I have to :o)
10:41 Mithaldu PR incoming in that case
10:41 FROGGS Mithaldu++
10:41 FROGGS and sorry for my wrong guessing :S
10:42 Mithaldu happens, just please throw more words at me next time so we can work out what's happening :D
10:43 Mithaldu FROGGS: the stuff in pods is not auto-generated, is it?
10:44 FROGGS no, isnt
10:44 Mithaldu cool
10:44 FROGGS https://github.com/PerlGameDev/SDL/blob/master/lib/pods/SDLx/App.pod
10:44 Mithaldu yeah, already editing that
10:44 Mithaldu it's just one extra line
10:44 FROGGS only the README gets generated AFAIK
10:51 Mithaldu FROGGS: https://github.com/PerlGameDev/SDL/pull/273/files
10:53 Mithaldu FROGGS: i'd also argue the delta time passed to the show handler is wrong, since it is calculated before the move handlers are processed. any thoughts on that?
10:53 Mithaldu actually, ticketing that
10:54 Mithaldu since it clearly contradicts the docs, but i'm unsure which is the right way
10:56 FROGGS hmm, good question
10:57 SDL joined #sdl
10:57 SDL [SDL] FROGGS pushed 2 new commits to master: http://git.io/pS7Z
10:57 SDL SDL/master 4aaa0a9 Christian Walde: clarify argments to handler subs in SDLx::App...
10:57 SDL SDL/master cccc117 Tobias Leich: Merge pull request #273 from wchristian/patch-1...
10:57 SDL left #sdl
10:57 FROGGS I'll do a release tonight
10:58 Mithaldu https://github.com/PerlGameDev/SDL/issues/274
10:59 FROGGS Mithaldu: you've got a commit bit now btw
10:59 Mithaldu FROGGS: i mainly found this by trying to figure out what the hell the lower red bits on the timestamp graph are: https://dl.dropboxusercontent.com/u/10190786/Screenshot%202015-03-13%2011.58.42.png
10:59 Mithaldu \o/
10:59 Mithaldu thanks
10:59 FROGGS you're welcome
11:00 Mithaldu i mean, i'm still not entirely clear on what exactly it is
11:00 Mithaldu SDL seems to cycle a bunch before actually doing the move bit
11:01 Mithaldu might be it trying to cap the framerate
11:02 Mithaldu oh, right, yeah
11:02 Mithaldu if i uncap the framerate that disappears, but then everything becomes choppy as hell due to the stepness of the move handler
11:03 Mithaldu on that note: visual frame timing debuggers are amazing
11:04 Mithaldu FROGGS: also cheers on the new release :)
11:10 travis-ci joined #sdl
11:10 travis-ci PerlGameDev/SDL#26 (master - cccc117 : Tobias Leich): The build passed.
11:10 travis-ci Change view : https://github.com/PerlGameDev/SDL/compare/ba434d518bd1...cccc117197eb
11:10 travis-ci Build details : http://travis-ci.org/PerlGameDev/SDL/builds/54226350
11:10 travis-ci left #sdl
11:18 Mithaldu FROGGS: i'm amused how a number of test releases have large amounts of commits that do not exist in master
11:22 Mithaldu FROGGS: think it would be a decent idea to dig through the experimental and debian perl branches for stuff to pull in?
11:25 Mithaldu yeah, damn, there is good stuff in here
11:26 Mithaldu plus, none of y'all is likely to even try and rebase any of this :v
11:53 Mithaldu FROGGS: i'm gonna be implementing the last variant of this here at some point: http://gafferongames.com/game-physics/fix-your-timestep/
11:53 Mithaldu FROGGS: do you think it'd be better to update the existing loop to that, or add an alternative loop?
12:03 FROGGS depends if that will likely brak anything or not
12:03 FROGGS with experimental branch you mean SDL's branch or something else?
12:04 Mithaldu yeah, the one that forked off in 2012 and was never curated and is a giant mess, now ignored by the recent releases :P
12:04 FROGGS btw, the changes of the dev releases will be pulled into master when they prove to be stable
12:04 FROGGS :/
12:04 Mithaldu haha
12:04 Mithaldu you *can't* just pull them in
12:04 FROGGS I'll check this evening
12:04 Mithaldu they're amazingly divergent at this point
12:04 Mithaldu just give me a week to curate that stuff :)
12:05 FROGGS okay
12:05 Mithaldu i can make the history at least sanely readable
12:05 Mithaldu also, for the timestep function
12:05 Mithaldu breaking is not hard-defined at that point
12:05 FROGGS true
12:06 Mithaldu i guess i'll just make a second loop and see how well it works
12:06 Mithaldu though it would be nice to have access to some code that actually uses SDL in which i can test how it affects other people
12:06 Mithaldu i kind of have the feeling there's no user of perl sdl who does as strenuous stuff as i am
12:07 basiliscos joined #sdl
12:17 FROGGS hmmm
12:18 FROGGS FrozenBubble does not use SDLx::App most likely
12:29 FROGGS and the other programs (games) that are worth anything are saving sue and bubblebreaker
13:01 Mithaldu FROGGS: links?
13:10 FROGGS https://github.com/garu/Games-SavingSue
13:10 FROGGS https://github.com/FROGGS/Games-Solitaire
13:10 FROGGS https://github.com/FROGGS/p6-Games-BubbleBreaker
13:11 Mithaldu nice, thanks :)
13:16 Mithaldu FROGGS: also i SDL puts the delay at the wrong place in the function
13:17 Mithaldu even when i'm operating below performance pressure margin does my framerate fluctuate a lot
13:43 Mithaldu FROGGS: there's also this weird phase at startup where the frame delay slowly scales down despite a lack of performance pressure: https://dl.dropboxusercontent.com/u/10190786/Screenshot%202015-03-13%2014.42.15.png
13:52 basiliscos joined #sdl
14:26 Mithaldu FROGGS: the timings graph makes sense now \o/ https://dl.dropboxusercontent.com/u/10190786/fps.png
14:26 Mithaldu also i understand the hitching at the start now
14:27 Mithaldu the game was just running faster because not all of the bots had spawned, which with the fucked timing loop caused the physics to be super weird
14:30 Mithaldu also the fps locking is super not accurate and at the wrong place in the loop
14:31 Mithaldu hm, or maybe i'm still measuring that wrong
14:44 FROGGS bbl
15:18 FROGGS joined #sdl
15:22 dod joined #sdl
15:26 Mithaldu FROGGS: can you explain what the hell this does? https://metacpan.org/source/FROGGS/SDL-2.544/lib/SDLx/Controller.pm#L14-15
15:26 Mithaldu err
15:26 Mithaldu what the hell this means?
15:27 Mithaldu is it because SDL::Surface is not a class at all, but just an XS pointer or something?
15:33 FROGGS yes, the intend was to let other packages inherit from SDL::Surface, which comes from XS
15:34 FROGGS though, I guess there are better ways to achieve that
15:34 FROGGS kthakore did that
15:37 Mithaldu ok, i think i kinda get it
15:37 Mithaldu since SDL::Surface is a struct it is like a class and carries data around in itself
15:37 Mithaldu but if you subclass it you can't add any attributes
15:37 Mithaldu and if you make a class and access it in there with "handles" then you have a permanent time loss on all calls
15:38 Mithaldu man, this shit's a little fucked :(
15:38 FROGGS aye
15:38 FROGGS P5 OO with XS does not rock :o)
15:40 Mithaldu FROGGS: i'm tempted to say "fuck the performance loss"
15:40 Mithaldu since uh
15:40 FROGGS I'd say: dont touch it
15:40 Mithaldu it only applies to stuff like surface width, height, which you can cache
15:40 Mithaldu or get/set pixel
15:41 Mithaldu which, when you're using it, you've already written all performance off
15:42 Mithaldu also, i'm not really thinking of changing SDLx and friends, but rewriting them wholesale
15:42 Mithaldu good god
15:42 Mithaldu SDLx::Surface already *does* handles-like stuff
15:43 Mithaldu so the performance loss is already there :v
17:51 dod joined #sdl
18:29 Mithaldu FROGGS: thanks to this weird habit of cherry-picking commits from a different branch and then *merging* these two branches together, it's really friggin hard to make heads or tails from the experimental branch
18:35 Mithaldu also some commits seem to be cherry-picked twice?
18:36 FROGGS that sounds highly unlikely
18:37 Mithaldu FROGGS:
18:37 Mithaldu SHA-1: c13693c3e9ba61d522571688f88e59925664ac9c
18:37 Mithaldu * platform specific notes by Alex
18:37 Mithaldu SHA-1: abd03a3043d9a87e5cbff006fb9da36774feccd8
18:37 Mithaldu * Reverted to 2.537_02
18:37 Mithaldu SHA-1: 13c6aebccb590fc7082d7c0a752fed3007804511
18:37 Mithaldu * platform specific notes by Alex
18:37 FROGGS ahhh
18:37 Mithaldu just to make sure i get *really* frustrated with this :P
18:38 FROGGS :P
18:57 Mithaldu FROGGS: so why exactly did that atomic bomb revert happen?
18:58 FROGGS I have no idea
18:58 Mithaldu awesome :D
18:58 FROGGS :P
18:59 Mithaldu there's weird stuff in there like, apparently at some point SDL::delay was swtched to Time::HiRes
18:59 Mithaldu and that was reverted back to delay
19:00 Mithaldu and of course the original commit message for introducing Time::HiRes explains nothing
19:00 Mithaldu so it might still be good?
19:00 FROGGS I'd say so
19:00 Mithaldu basically what i'm trying to say is:
19:01 Mithaldu merging experimental to master is never gonna happen
19:01 Mithaldu ever
19:01 FROGGS I remember vaguely that there was a regression that was hard to bisect
19:01 FROGGS maybe the result was to revert a lot
19:01 Mithaldu there's no way you'd be able to untangle that mess
19:04 FROGGS put your stuff in a feature branch then
19:05 Mithaldu the thing is i was trying to figure out what from the feature stuff can actually be used
19:05 Mithaldu *experimental
19:05 Mithaldu that's usually done by rebasing
19:06 Mithaldu but that giant revert pretty much broke rebasing past a certain point
19:07 Mithaldu i can't even find an issue that mentions revert or reverting
19:10 Mithaldu argh, the revert wasn't even a full revert
19:11 Mithaldu it's like that one single commit was made with the sole intent of creating maximum trolling
19:30 Mithaldu SDL::delay seems to work better than Time::HiRes, so i've no idea why anyone would replace it
19:50 Mithaldu FROGGS: and the worst thing are deprecations without explanation for why they're deprecated
20:01 FROGGS SDL::delay was problematic on BSD IIRC
20:05 Mithaldu interesting
20:05 Mithaldu bit of a shamer blaizer did that change in one big mega commit
20:09 FROGGS well, most contributers are not git/svn professionals with 10+ years of experience...
20:09 FROGGS so, these things happen
20:10 Mithaldu just consider that i'm going back to 2011 to find the original commits for changes ;)
20:10 Mithaldu also, i don't mind a fucked history
20:11 Mithaldu but commits that do a lot of disparate changes AND don't bother to explain why, you don't need to be a pro to see how that's wrong
20:11 FROGGS Mithaldu: yes, but everybody starts as a n00b
20:11 FROGGS and we were certainly n00bs back then
20:12 Mithaldu right :)
20:14 FROGGS so please just care about what you want to achieve and do that in a feature branch :o)
20:15 Mithaldu i do care about not throwing away useful work ;)
20:15 Mithaldu kthakore's last commit to x::Controller is interesting
20:16 Mithaldu he removed the sdl pump call
20:30 Mithaldu FROGGS: the current incarnation of x::Controller is literally from 17.09.2011
20:31 FROGGS is that good or bad?
20:32 Mithaldu well, i do assume that the screenful of commits i am seeing inbetween does something useful :)
20:32 FROGGS ahh :o)
20:33 Mithaldu i just noticed you never even touched it
20:53 Mithaldu FROGGS: 5315653386274aaa75ada98d519078d18ba67266
20:53 Mithaldu wouldn't that function need an empty return at the end?
20:54 FROGGS Mithaldu: why? it retuns $_stash{ $ref } which you can assign to
20:55 FROGGS unless I dont understand how :lvalue is supposed to work
20:55 Mithaldu what does this message mean then?
20:55 Mithaldu * lvalue subs dont return stuff
21:01 FROGGS probably a thinko
21:01 FROGGS it was an attempt to fix a weird issue spotted in a cpan testers report IIRC
21:02 Mithaldu aight, not very important
21:02 Mithaldu i was just curious what was going on there :)
22:27 kaare joined #sdl

| Channels | #sdl index | Today | | Search | Google Search | Plain-Text | summary