Minetest logo

IRC log for #minetest-dev, 2013-11-26

| Channels | #minetest-dev index | Today | | Google Search | Plaintext

All times shown according to UTC.

Time Nick Message
00:00 celeron55 pitriss: what version of iceweasel is that?
00:00 celeron55 looks old as hell 8)
00:00 pitriss and maybe also some really thin border around those dropdowns could be nice (do bottom border under menu and then border around left, bottom and right side of dropdown) It will split menu from content:)
00:00 pitriss celeron55: old.. 3.6, because I don't like those new versions:)
00:01 pitriss I want to have iceweasel and not chrome clone:D
00:01 celeron55 i think firefox still is pretty good
00:02 pitriss Not ui copied from chrome/chromium
00:02 pitriss I don't like tabs above addressbar etc..
00:03 celeron55 oops, now i broke it completely
00:05 Taoki joined #minetest-dev
00:05 celeron55 pitriss: how about now?
00:06 pitriss Now it is bit better, mods dropdown is singleline now but needs to be positioned to the left.. I will post screenshot
00:08 pitriss http://chytrex.wjake.com/crap/mt-site1.png
00:09 celeron55 now?
00:10 SmugLeaf joined #minetest-dev
00:11 celeron55 pitriss: please be quick because i'm going to sleep just about now
00:11 pitriss same
00:11 VanessaE ew.
00:11 pitriss celeron55: ah sorry:)
00:12 VanessaE is it supposed to be so plain?
00:12 VanessaE (the hovered links look fine to me, btw)
00:12 VanessaE (ff 25)
00:13 celeron55 pitriss: now?
00:13 pitriss celeron55: if it is not broken in other browsers, let it as is.. It is usable now even in my prehistoric browser:)
00:13 celeron55 VanessaE: plain?
00:13 pitriss great
00:14 Ritchie joined #minetest-dev
00:14 VanessaE celeron55: aside from the formatting being correct for me, it looks the same as in pitriss's last screenshot.
00:14 VanessaE e.g. the black text/white background, etc.
00:14 VanessaE I guess I didn't get the memo about the site being redone :)
00:15 celeron55 none of the site is being redone; it's just the "skin" of the site
00:15 celeron55 i can toggle between this and the previous one from dokuwiki's admin panel 8)
00:15 celeron55 but yes, it's somewhat plain
00:15 VanessaE well that said, this "skin" is perhaps a little "cleaner" than the other one
00:15 celeron55 by that definition
00:16 VanessaE but no screenshots!!! D:
00:16 VanessaE :)
00:16 celeron55 give me a good screenshot
00:16 VanessaE anything I could come up with will not be suitable for the main website :)
00:16 VanessaE (because mods)
00:17 celeron55 mods is the point of minetest
00:17 VanessaE true..
00:17 celeron55 i'll just write what mods it uses under the screenshot and it's perfectly fine
00:17 VanessaE I'll see what I can come up with later, catch you when you wake up
00:18 celeron55 one thing i might do is tune up the contrast of the text; it's not really my style to have so little contrast
00:18 VanessaE oh G*d yes
00:18 celeron55 (i'm obviously basing this on a random template copied from the 'net which had that)
00:18 pitriss celeron55: yep contrast should be higher, yes
00:20 pitriss Idea: how about to make those dropowns in light grey ?
00:22 VanessaE they do need a little bit of difference from the main background, yeah
00:22 VanessaE "Simplistic Modding API that..."  <--- maybe "Sophisticated" now?  It's hardly simplistic :)
00:24 celeron55 now it's less grey
00:25 VanessaE better.
00:31 pitriss i played with it a bit in stylish: http://chytrex.wjake.com/crap/mt-site-idea.png //how about that border under menu?
00:31 VanessaE tht's not bad.
00:33 pitriss thanks:) It is far from perfect now, it needs more work but it was only example:)
00:44 celeron55 the template is here if someone wants to play around with it: https://github.com/celeron55​/minetest.net_dokutemplate2
00:45 celeron55 making a random local dokuwiki install for developing is quite easy
00:47 pitriss celeron55: ok.
01:00 Exio4 wow kde3
03:22 OldCoder joined #minetest-dev
05:17 thexyz celeron55: it's better than what we had
05:18 thexyz by the way, anyone contacted the debian maintainer about the release?
05:53 RealBadAngel joined #minetest-dev
06:02 celeron55 probably not
06:03 RealBadAngel hi
06:12 werwerwer joined #minetest-dev
06:13 Panda joined #minetest-dev
06:35 Akien joined #minetest-dev
06:42 djdduty joined #minetest-dev
07:06 darkrose joined #minetest-dev
07:06 darkrose joined #minetest-dev
07:49 Jordach joined #minetest-dev
08:31 BlockMen joined #minetest-dev
08:32 BlockMen celeron55, i know my template wasnt perfect, but sry, this is 0% representative for a game website
08:32 BlockMen IMO it looks like a fancier version of your blog
08:43 proller joined #minetest-dev
08:44 thexyz BlockMen: what would be representative for a gane website?
08:44 darkrose joined #minetest-dev
08:45 BlockMen thexyz, IMO some textures, but at least some color
08:45 thexyz text with a background texture which doesn't even tile?
08:46 thexyz dunno
08:46 thexyz what we had looked awful
08:46 BlockMen and this looks naked
08:46 * VanessaE mumbles something about the Emperor's New Clothes...
08:46 thexyz perhaps
08:47 BlockMen and you can use background texture and 'just' color for text area
08:47 thexyz to me it looks simple but that's a good thing compared to that web1.0-ish texture mess
08:47 thexyz also fuck my internet connection
08:48 thexyz but that's not what we had
08:48 thexyz we had texture hell
08:48 thexyz which didn't even tile
08:48 BlockMen thexyz, "i know my template wasnt perfect"
08:48 BlockMen clear enough?!
08:49 thexyz for how much time
08:49 thexyz it wasn't good and it wasn't improving
08:51 proller freebsd port updated ! /usr/ports/games/minetest
08:53 Jordach BlockMen, here's an idea, go browse some well known game websites, and look at the themes
08:56 BlockMen Jordach, ok i visited 6 websites now and none looks that plain and colorless
08:57 Jordach BlockMen, i would have said the same
08:57 thexyz what about non tiling textures?
09:01 BlockMen idk what the perfect theme looks like
09:01 BlockMen i just said my opinion on the new
09:02 thexyz alright
09:07 thexyz to you it may look plain or colorless but to me it looks light and easy to navigate and read
09:17 pitriss thexyz: I also like the new web template but.. I modified it a bit in stylish to add thin black borders to left and right side of content, and line below menu. I can post screenshot with css changes I made.. it look less blank now..
09:24 smoke_fumus joined #minetest-dev
09:48 Akien joined #minetest-dev
09:50 pitriss https://gist.github.com/Pi​triss/5f0ef9a9d7d028fb6373 //my changes to the website template. Untested on real dokuwiki, I only applied my local stylish code to original one.
10:35 bas080 joined #minetest-dev
10:36 us^0gb joined #minetest-dev
11:00 KingsleyT joined #minetest-dev
11:03 ecube joined #minetest-dev
11:15 bas080 joined #minetest-dev
11:32 yno joined #minetest-dev
11:39 celeron55 oh apparently i made something very controversial
11:39 celeron55 job well done *pats myself*
11:41 * us^0gb giggles
11:43 Neological joined #minetest-dev
11:52 iqualfragile joined #minetest-dev
12:36 iqualfragile joined #minetest-dev
12:39 Zeitgeist_ joined #minetest-dev
12:39 Zeitgeist_ joined #minetest-dev
12:44 damiel joined #minetest-dev
12:48 Zeitgeis1 joined #minetest-dev
12:49 Zeitgeist_ joined #minetest-dev
13:17 Jordach joined #minetest-dev
13:18 Akien joined #minetest-dev
13:20 Akien joined #minetest-dev
13:25 hmmmm joined #minetest-dev
13:26 EvergreenTree joined #minetest-dev
13:29 Akien joined #minetest-dev
13:33 thexyz yes because everyone agreed previous design was not-so-good
13:37 thexyz design of the quote is quite meh though, the quotation mark isn't even visible
13:46 NakedFury joined #minetest-dev
13:54 ImQ009 joined #minetest-dev
14:20 ecube joined #minetest-dev
14:25 zat joined #minetest-dev
14:56 proller joined #minetest-dev
14:57 Gethiox joined #minetest-dev
15:03 celeron55 the site is of course open for other designs too, and that design is open for improvements
15:10 Akien_ joined #minetest-dev
15:11 BlockMen joined #minetest-dev
15:12 BlockMen my interpretation of clean design for a game website: https://dl.dropboxusercontent.com/s​/v9gni23wlbys0dq/hp_t3_test_3_3.png
15:13 pitriss celeron55: i posted my idea here: https://gist.github.com/Pi​triss/5f0ef9a9d7d028fb6373 (In case you didn't read backlog) (not tested on real dokuwiki)
15:16 celeron55 BlockMen: and how would the non-front pages look like?
15:16 BlockMen same but with a smaler image like 1/3 or something
15:39 thexyz pitriss: can you post some pics?
15:40 pitriss thexyz: screenshot link is in that gist description:) http://chytrex.wjake.com/crap/mt-site-idea2.png I just added some lines only..:)
15:41 thexyz oh sorry
15:41 pitriss np
15:41 thexyz but honestly I prefer the version without any lines
15:42 pitriss line is also around dropdown if triggered..
15:42 thexyz celeron55: fix dropdowns
15:42 pitriss thexyz: I'ts ok, it was only idea..
15:42 thexyz celeron55: and blockquote
15:43 celeron55 BlockMen: i'd prefer no image, but more than that i'd prefer seeing how it actually would look
15:45 BlockMen celeron55, its just a mockup since that is much less work. if you are really interested in i will make a template next day(s)
15:45 celeron55 baah, not that interested 8)
15:46 celeron55 thexyz: what's there to fix about dropdowns
15:46 BlockMen ok, then not
15:48 thexyz celeron55: it's not really clear a link is a dropdown actually, though it seems it's not really a problem
15:48 thexyz and while we're talking about designs here's what I mocked up for MMDB http://i.imgur.com/YhTNaox.png
15:48 celeron55 but overally your designs (that too) put too much focus to the layout so that the actual content looks almost nonexistent
15:48 celeron55 and insignificant
15:49 celeron55 while i would like to put content first
15:50 celeron55 i like mmdb quite a lot as-is, except that there are things that don't have any styling at all which is bad
15:51 thexyz and I find the current design awful so that's why I built what I did
15:53 BlockMen thexyz, why you think its awful? the is much white, not tile textures, strange color...
15:53 celeron55 or, well... really the only thing that i consider properly made is the top navigation bar
15:53 BlockMen *there
15:54 thexyz BlockMen: just look at https://forum.minetest.net/mmdb/mod/worldedit/
15:54 thexyz that random space on the left of "WorldEdit"
15:54 thexyz that randomly placed Download button
15:55 thexyz that gallery ;_;
15:55 thexyz reviews
15:55 celeron55 i think comments and reviews should shown as one list, if they have to co-exist in the first place
15:55 BlockMen well, ofc its not perfect. but that is IMO nothing atm
15:55 thexyz celeron55: that's one of my points, mmdb should only have mods and reviews
15:56 BlockMen and i think there should be one "master" design that should be used on forum, mmdb and so on
15:56 thexyz everything else goes elsewhere: comments (forums), bug reports (github, forum)
15:56 celeron55 i think layouts of different websites should show just hints of a common theme and otherwise have their own personality
15:57 thexyz BlockMen: yeah go on and apply minetest.net's design to the forum
15:57 celeron55 it's boring if everything uses the same layout
15:57 celeron55 and the same layout doesn't fit all
15:57 thexyz any opinions on the new design, not the old one?
15:58 BlockMen at least similar, not a blank and white mainpage, a dirt backgrounded forum and a green mmdb
15:58 BlockMen that is even worse than everything the same
15:58 celeron55 thexyz: your screenshot doesn't fit on my screen and is
15:58 celeron55 -and is
15:59 thexyz celeron55: what's your width?
15:59 celeron55 but not considering that, it's contrast is very random
15:59 celeron55 the top bar is completely black and then the borders all around are so dim it's almost impossible to see them
15:59 BlockMen thexyz, umm..you kept the "random" gap at mod title
15:59 thexyz BlockMen: what?
15:59 thexyz it's centered
16:00 thexyz celeron55: I'm more talking about the layout, colors can be easily changed
16:00 celeron55 it could be a better start than the current one
16:00 BlockMen thexyz, ic. but it doesnt look like ;)
16:00 thexyz BlockMen: then fix your vision or something?
16:00 celeron55 thexyz: how about putting the reviews on the right column instead of under the columns?
16:00 thexyz no offense and shit
16:00 celeron55 having text and fields at the full width looks just bad and is bad to use too
16:00 thexyz celeron55: put where? in place of description?
16:01 celeron55 under it, without going under the image and metadata thing
16:01 thexyz celeron55: I don't see the problem here, the full width is limited to 1000px
16:01 thexyz celeron55: and then there's a lot of space on the left and it'll look bad
16:02 BlockMen thexyz, nice you stay that objective -.-'
16:02 Calinou joined #minetest-dev
16:02 celeron55 well then, how about moving the "write review" box under the reviews so that the reviews don't get buried down at the bottom
16:02 thexyz BlockMen: what? you can't see it's centered and it's my problem? or what?
16:03 thexyz I don't really understand, you didn't even tell me what the problem is
16:03 thexyz celeron55: what if there are too many reviews and no one will get to the "write a review" section
16:03 celeron55 well if there are so many reviews, why would anyone need to write a review anymore
16:04 celeron55 it's not like you need 1000 of them
16:04 BlockMen YOU asked for opinions, i told you mine. that no reason to get personal
16:04 thexyz BlockMen: okay, I'm sorry
16:04 celeron55 thexyz: how about making "write a review" a button that will expand out the form?
16:05 thexyz celeron55: and leave it at its place?
16:05 celeron55 yes
16:05 thexyz maybe
16:05 thexyz should it be expanded?
16:05 celeron55 the problem for me is that i don't want to be looking at empty boxes where i'm not going to want to write anything
16:06 thexyz http://i.imgur.com/Xc1ZRxx.png
16:07 celeron55 that's good
16:07 celeron55 now do something to the huge black hole at the top
16:09 PilzAdam joined #minetest-dev
16:09 thexyz what? what's the problem here?
16:10 celeron55 it has too much contrast
16:10 thexyz oh, i thought you were talking about too much wasted space
16:11 * sfan5 was thinking that too
16:13 thexyz I haven't even touched the colors yet
16:13 thexyz http://i.imgur.com/tIrs50V.png
16:13 thexyz and the css in general
16:13 sfan5 what about a green tint for the top bar?
16:15 celeron55 how about if you change that top bar to what mmdb has currently?
16:15 celeron55 with a smaller logo though; the current one is too huge
16:15 BlockMen that green looks not good
16:16 thexyz I don't like that it's sticky
16:19 thexyz we need the official Minetest Green color
16:20 BlockMen official green?
16:20 sfan5 I'd say the green we have in the logo
16:20 BlockMen that green looks better https://dl.dropboxusercontent.com/s/nj​geuf3e9pxxvgl/hp_t3_test_3_3-green.png
16:21 sfan5 that's #73d216
16:21 sfan5 (in the logo)
16:21 BlockMen oh, thats easier, gd idea. i prefer #65B41D
16:21 EvergreenTree joined #minetest-dev
16:22 thexyz http://i.imgur.com/X5DWyvz.png
16:23 thexyz just upgraded the foundation as well
16:24 thexyz both colors you provided look awful when used in top bar
16:29 BlockMen IMO it looks fresher http://i.imgur.com/SU0IwX2.png :)
16:30 thexyz eh
16:30 celeron55 how about making the bar simply white and some heavy hover effects (eg. that background) for the links in there
16:30 thexyz my eyes certainly don't like it
16:32 thexyz celeron55: i think it'll look too plain and therefore hard to intuitively distinguish from the actual content and this is bad because it's the main navigation item
16:32 sfan5 what about the water blue from the logo (#729fcf)
16:33 thexyz what about it?
16:33 sfan5 for the top bar
16:33 celeron55 thexyz: try making it somewhat darker and less saturated then; maybe switch the colors there
16:33 sfan5 it's not too bright, but not too dark
16:33 thexyz not it should be green
16:34 celeron55 (i mean, in BlockMen's image)
16:34 thexyz what?
16:34 thexyz what's so bad about the current color?
16:36 thexyz why do we want to base it on the logo color?
16:37 thexyz where it clearly wasn't designed to suit topbar
16:37 sfan5 why do we need to have everything in green?
16:38 BlockMen right, the download button color doesnt look bad either
16:38 thexyz dunno it just looks nice
16:39 sfan5 please try #729fcf once
16:40 PilzAdam let the users choose the color
16:41 thexyz sfan5: of course i've tried it
16:41 thexyz PilzAdam: nice joke
16:41 sfan5 and it doesn't look nice?
16:41 BlockMen #008CBA looks better :P
16:41 thexyz yes, i feel like it doesn't because everything is blue already
16:42 thexyz anyway arguing here is not productive
16:42 thexyz i should find iqualfragile and ask him what pages should i do instead
16:43 sfan5 BlockMen: that is a bit dark
16:45 BlockMen sfan, #009DD1 ?
16:46 sfan5 that's better
17:05 sapier joined #minetest-dev
17:08 OWNSyouAll joined #minetest-dev
17:08 ImQ009 joined #minetest-dev
17:10 ImQ009 joined #minetest-dev
17:16 sapier https://github.com/minetest/minetest/pull/977 squashed and (hopfully) ready for merge did you have any additional issues PilzAdam?
17:16 Akien joined #minetest-dev
17:19 PilzAdam sapier, remove debug output
17:20 sapier did I forget something again?
17:20 PilzAdam engine.debug("Initializing Asynchronous environment")
17:20 PilzAdam woah, just crashed again
17:20 sapier argh
17:20 PilzAdam ERROR: Async ENGINE step: /home/adam/Minetest/minetest​/builtin/async_event.lua:14: assertion failed!
17:20 PilzAdam lets gdb
17:21 sapier that's the one thing I wasn't sure if it can happen at all
17:21 sapier why do we have engine.debug if it shouldn't be used?
17:21 proller sapier, why marshal ?
17:22 PilzAdam sapier, https://gist.github.com/PilzAdam/7662338
17:22 sapier because you need to transfer the code to another INDEPENDENT environment
17:22 PilzAdam sapier, its useful to print() while debuging a mod, but it shouldnt be in code that is to be released
17:22 sapier it's not a print but engine.debug
17:23 PilzAdam print() is an alias for minetest/engine.debug()
17:23 sapier why? I expected this to not do anything in release mode
17:23 PilzAdam because mods need it, and mods are developed on release builds
17:24 PilzAdam is that backtrace useful?
17:24 sapier I don't care enough to discuss any further about the ininializing message ;-)
17:25 sapier I'm not quite sure looks like one of those random crashes supposed to be fixed by curl init
17:25 PilzAdam just use mintest.log("info", ...)
17:26 sapier are you sure minetest.log is available in mainmenu ?
17:27 PilzAdam engine.log then
17:27 PilzAdam https://github.com/minetest/minetest/​blob/master/doc/menu_lua_api.txt#L133
17:27 sapier I'll try ... yet I'm more concerned about that crash
17:28 PilzAdam I dont like that you remove so many tabs from empty lines in main.cpp, this probably breaks other pull requests
17:28 sapier I guess I need to find out how to make eclipse not do that automaticaly
17:29 sapier but I won't turn it of for spaces at end of line ... I hate those ;-)
17:30 sapier can you provide the other threads backtraces for that crash too?
17:31 PilzAdam https://gist.github.com/PilzAdam/7662510
17:31 sapier and maybe in AsyncEngine::doAsyncJob the lenght of serialized data passed from lua
17:32 sapier I have a feeling that something with converting marshalled function to string may be involved ... but I could be completely wrong with this too
17:32 PilzAdam this? https://gist.github.com/PilzAdam/7662540
17:33 sapier <optimized out> .... thats why I hate -O1
17:34 sapier but let me check your second log first
17:34 PilzAdam building with -O0
17:35 PilzAdam wow, that builds quite fast
17:35 sapier 3 threads in modstore details ... all reading settings ... suspicious
17:37 PilzAdam I just reproduced it with -O0 and AsyncEngine::doAsyncJob runs in no thread
17:38 sapier ok plz post backtrace the more are available the better chances to find it
17:38 sapier what machine do you have? :-) guess it's way faster then mine
17:38 PilzAdam https://gist.github.com/PilzAdam/7662628
17:39 sapier at least within this trace it's obvious to be a memory corruption :-(
17:39 PilzAdam intel core i7 8 x 2.93 GHz
17:40 sapier way faster true
17:40 PilzAdam memory corruption?
17:41 sapier malloc/free new/delete not used as it should be
17:41 sapier or stack objects ... or something else
17:42 sapier the interesting thing is that most time job queueing is involved
17:43 sapier argh
17:44 PilzAdam https://gist.github.com/PilzAdam/7662696
17:44 sapier m_JobQueueMutex.Init();
17:44 sapier m_ResultQueueMutex.Init();
17:44 PilzAdam I should write a script to produce these ;-)
17:44 sapier can you put those two lines in AsyncEngine constructor
17:45 sapier I't great to have a mutex to protect data ... unless you forget to initialize it ;-)
17:46 PilzAdam l_async_events.cpp:68?
17:46 sapier yes
17:46 sapier I guess I should start checking return values for lock and unlock functions
17:47 PilzAdam seems that fixed it
17:48 sapier good ... it's a simple yet effective method to avoid locking if you don't even initialize it ... yet it doesn't protect anything this way
17:52 sapier did anyone else test it?
18:07 BlockMen left #minetest-dev
18:16 Calinou joined #minetest-dev
18:18 celeron55 oh, this refreshed my mind about a thing https://forum.minetest.net/vie​wtopic.php?pid=119741#p119741
18:19 celeron55 was that fix committed to upstream?
18:19 celeron55 it should have been, but i don't remember at all
18:20 celeron55 looks like it wasn't
18:20 sapier I don't think so proller merged all his pull request to a single one
18:21 celeron55 well fuck proller
18:21 sapier but of course that's not an excuse for not merging the good ones ... just "out of sight out of mind"
18:21 celeron55 can someone find the commit
18:22 sapier I'll have look which of his closed pull request is most likely matching this issue
18:23 PilzAdam sapier, the "Make game better" one ;-)
18:23 sapier not helpfull ;-P
18:24 celeron55 it looks to me that he didn't make a pull request for it
18:24 sapier he did he just closed it
18:24 celeron55 well i don't find it in his closed pull requests
18:25 sapier #895
18:25 celeron55 nope
18:26 celeron55 already looked at that
18:26 PilzAdam its propably part of one of the pull requests, he never devided features into several ones
18:26 celeron55 proller should seriously get his shit in order
18:27 sapier ok this is at least the one I had in mind but it could be part of his other pulls too ... they haven't been exactly single feature pulls
18:27 celeron55 minetest suffers because of him; there's broken code by proller in upstream and fixed by proller can't be applied
18:27 celeron55 fixes*
18:27 Exio4 broken? are you talking about weather?
18:27 celeron55 yes
18:28 sapier is there any way to sort closed pulls by user on github?
18:28 PilzAdam sapier, just filter for one user and then click on "closed"
18:28 celeron55 there are three options: 1) proller stops making it impossible to merge his fixes, 2) somebody else continues them, 3) proller's original broken things are taken out
18:29 sapier great ... sometimes I wonder why order of clicking is relevant
18:29 Zeitgeist_ joined #minetest-dev
18:29 celeron55 i actually just made the same thing; clicked closed before proller
18:30 celeron55 but then tried the other way
18:30 PilzAdam thats how you shouldnt design an interface
18:30 celeron55 it's like they didn't know if they wanted to make tabs or filters
18:31 sapier #980 seems to be a performance fix too
18:32 celeron55 the fix modifies stuff around server.cpp:112
18:32 sapier ok I have a look at "make game better"
18:32 thexyz PLOT TWIST: this is actually #895, and getting fed up with this stuff not being merged for months and "lost" is understandable
18:33 sapier there it is
18:33 celeron55 thexyz: it sucks for both sides and the solution is not to make it harder for the other side
18:34 thexyz indeed
18:34 thexyz but people are different, some are fine with their stuff being "useless", some eventually get fed up with it
18:34 thexyz and give up
18:34 sapier I understand proller too (in some way) but adding new features while old ones aren't ready for merge isn't helpfull too
18:34 celeron55 okay it looks like the fix is in that pull request
18:35 thexyz or improve in other ways
18:36 celeron55 my stand on this is that we don't accept things based on how much time people put on them, but based on how much benefit vs. trouble we see coming from the end result
18:37 thexyz that's not what I'm talking about, and what I'm talking about is that because of poor communication people give up and quit
18:37 celeron55 it's related
18:38 thexyz as you wish
18:39 sapier I can't see any poor communication with proller
18:39 sapier I guess I'm not the only one who suggested to split his pull requests to avoid having multiple (sometimes questionable) features in a single one
18:39 thexyz sapier: and then what's the reason of #895 not being merged?
18:40 thexyz if not poor communication
18:40 sapier I guess the liquid update things?
18:41 sapier and if there are more then 100 timers those are ignored too
18:41 celeron55 proller doesn't ask anything beforehand from anyone but instead comes with things he thinks he has completed and then wants to push them without even showing much signs that he understands what the modifications do
18:41 celeron55 or, well, this is how i perceive it
18:41 thexyz so instead of telling him what's wrong you've decided to just ignore his pull requests
18:41 sapier thexyz we did tell him multiple times
18:42 sapier but he just kept on adding new things without fixing suggestions
18:42 celeron55 i wonder how many times proller has been told that he has too many features in one pull request or commit...
18:42 celeron55 then he fixes it by putting even more of them
18:42 celeron55 good luck making friends that way
18:43 sapier I'd be glad if proller came back and continues to contribute but by now his work is often more like prototypes that need cleanup
18:43 sapier ok I admit my first pulls sometimes seem to be same true
18:44 celeron55 sapier has a wibe of proller to how he does things but it's not "problem grade" bad
18:44 thexyz sapier: I've just checked the log for "895" and it seems he mentioned it several times and all those times this was 1) ignored 2) not ignored but answered with a pointless comment like "why did you raise max clients setting"
18:44 sapier so what about 895 take the fix from server.cpp and merge separatly or use whole thing?
18:44 celeron55 vibe*
18:45 thexyz sapier: I don't understand how do you want to separate this pull request and why
18:45 thexyz can you please explain?
18:45 celeron55 we're not talking about specifically this pull request
18:45 sapier server.cpp line 101 -120
18:45 celeron55 and people's reactions to a single pull request are affected by the overall impression from everything done before
18:45 thexyz alright, let's talk about other pull requests then if so you insist
18:46 thexyz sapier: do you want to only pick this? for what reason?
18:46 sapier thexyz I never told everything proller did was bad ... I'm almost sure most of it was good
18:46 sapier because I'm not sure what sideefects the other changes might cause
18:48 sapier but if you're sure they don't do anything bad I'm fine too
18:49 sapier but at least with timer things I have a bad feeling
18:49 thexyz then wait for him and ask
18:49 thexyz if you're not sure about something
18:49 thexyz or ask in the github issue
18:49 thexyz instead, for some reason when someone's unsure about something he or she keeps quiet about it and things move nowhere
18:50 sapier that doesn't change the fact this pull request is 10 features in one
18:50 thexyz not really
18:50 celeron55 proller has been known to try to push his finite liquids in ways that make original liquids work badly, and then people have had to fix stuff that he broke; it's no wonder at all that people are prejudiced
18:50 thexyz and what? that doesn't mean you should silently ignore his pulls
18:50 thexyz everyone fucks up
18:51 celeron55 well, people do this as a hobby and they don't want to do boring things like thoroughly testing other people's code
18:51 celeron55 there needs to be trust
18:51 sapier thexyz I don't think his pulls got silently ignored ... if this looks this way to otheres it's my/our fault of course
18:52 sapier but it's always two to make a problem ... proller didn't do anything to reduce conflict too
18:53 sapier btw the timer things look like a hack but most likely don't cause any harm
18:55 sapier but no idea what the liquid things do I just don't know that code enough to even guess
18:55 celeron55 well for one thing unintuitive things like that should always be commented so that people can know what is happening
18:55 celeron55 that's missing completely from those commits
18:57 celeron55 sprinkling random pieces of timers and dynamically changing things in the codebase is bad and any sane person would want to just not have it
18:57 celeron55 a better idea could be making a system for managing those things in a more controllable way
18:58 thexyz and even better idea would be to post this to the github issue instead of saying stuff here where the pull request author most likely won't even see this and no one is going to pass that to him obviously
18:58 EvergreenTree joined #minetest-dev
18:59 celeron55 it's okay to make heuristics, dynamically changing hacks and other random code if it's reasonably modularized into one place (like sky.cpp or the viewing range tuner by me), but it's not so much when it's a few lines here and there
19:00 sapier thexyz do I understand correct that you suggest moving pull request discussion (partly/total) to github?
19:00 celeron55 thexyz: that's an another problem we have; people discuss things in IRC and obviously they don't want to rewrite what they said to issue comments, and also the irc log is too long for linking from the issue
19:00 celeron55 thexyz: if you have some magic wand for handling that, use it
19:01 thexyz sapier: no, I mean, we should discuss them in a way that those discussions are accessible for all people; people shouldn't dig up the logs to find that you discussed the issue while he or she was away
19:01 thexyz celeron55: yes, it's called copypasting
19:01 thexyz just copy and paste the related lines into the comment on this github issue and that's good enough
19:01 sapier at least at this point thexyz is right at least important issues need to be added to the pull requests
19:02 sapier but I guess this won't help in this special case
19:03 thexyz eh
19:03 thexyz well whatever
19:03 celeron55 but what makes this all unacceptable for me is that any coder should know that sprinkling random code around a project like what those commits do isn't acceptable
19:03 celeron55 it's stupid to have to even say to someone that it's not ok
19:03 sapier proller won't fix it anyway ... but we should at least learn from things that have been done wrong in past
19:04 celeron55 always when one does something, he should think of what would happen if everyone would do things exactly like that infinitely into the future
19:04 celeron55 if the end result wouldn't be acceptable, then there needs to be a good reasoning why that won't actually happen
19:05 celeron55 this case could have been immediately merged if proller would have explained why those lines are completely enough for handling the problem, and saying something about the dynamics they cause
19:06 celeron55 well there's something in the strict meaning of that word, but it's that non-understandable proller gibberish
19:09 celeron55 if someone wants to see how to do things, look at how kahrl does things
19:09 celeron55 it's fucking art
19:19 sapier but back to original issue are those lines in server.cpp enough to fix the most pressing issue?
19:19 celeron55 the receive loop added there fixes the most horrible lagging under CPU load
19:20 celeron55 and has already been reviewed well
19:20 celeron55 the other stuff nobody really knows about
19:21 sapier ok I'm gonna create a pull request with only this change
19:28 celeron55 you could just push it
19:30 PilzAdam what disklike most about prollers way of handling things is how he ignore most of my suggestions / criticism about his pull request
19:30 PilzAdam +I
19:49 Gethiox joined #minetest-dev
19:49 proller joined #minetest-dev
19:55 sapier any idea what value to use for dedicated_server_step ?
19:57 PilzAdam 0.05 or 0.1
19:58 Gethiox2 joined #minetest-dev
19:59 sapier what convention do we use on counters? machine default type or specifying with u16/u32 ? later one is less performant
20:03 celeron55 yes, the code tends to use them quite randomly
20:03 celeron55 i these days usually use size_t for general unsigned needs
20:07 Akien joined #minetest-dev
20:09 sapier ok then I'm gonna use size_t too ... I just didn't want to use u16 as this is even slow on 32bit
20:11 djdduty joined #minetest-dev
20:11 djdduty joined #minetest-dev
20:15 sapier ok those changes aren't  as simple as they seemd to be
20:15 VanessaE regarding proller's pull requests, and noting that I have *not* used his fork myself, why stop at just one feature?  As I recall from helping him test server performance recently, there are several beneficial changes that he's made.  what is the *real* reason for not just merging the whole branch?  I mean aside from "well we just don't understand it".  Or to put it differently, let's suppose someone were to sit down one day
20:15 VanessaE , look at the code, and decide "you know what, this is shit and I'm gonna rewrite it".  That person does so, ends up with 10x better code in every respect than what came before, but it comes with the caveat that half the codebase got replaced.
20:15 VanessaE what happens then?
20:16 sapier VanessaE because his branch doesn't only contain performance fixes but things like wingsuite trees in caves increased cloud level too ... just to name some
20:17 VanessaE gotcha
20:18 sapier I wouldn't wait a second to merge a "performance fixes" branch containing them ... but prolller usually has a "performance fix here and feature x+y and change some default there" pull request
20:18 VanessaE right
20:18 VanessaE but here's the thing
20:18 VanessaE he's absolutely right
20:19 VanessaE his pull requests have, to some degree, been ignored.
20:19 VanessaE the question is what do you expect him to do?
20:19 VanessaE stop developing entirely while waiting the 1-2 months it seems to take?
20:19 sapier I do not accept this vanessae he hasn't been ignored I told him weeks ago to split those things to get the non controversial things merged
20:19 VanessaE (not you, specifically)
20:19 celeron55 communicate.
20:19 celeron55 in some other way than repeatedly linking the pull request here
20:20 celeron55 i have said this so many times i really don't want to repat this
20:20 thexyz in what way then?
20:20 sapier but he told me it's to much work for him ... so what to do start working after him to get a clean pull?
20:20 VanessaE celeron55: the problem is no one's ever here to answer him
20:20 thexyz "hey guys please look" "hey i made this pull" "please check this out"
20:20 thexyz how else should we communicate?
20:20 celeron55 if people aren't interested in it, tell people why they should be interested in it, goddamnit
20:20 VanessaE celeron55: who
20:20 VanessaE oops
20:20 VanessaE celeron55: who's to tell if there's no one here to respond?
20:21 thexyz maybe he meant the dev should hilight all other people or something
20:21 celeron55 that's a stupid question
20:21 VanessaE I've left huge diatribes in this channel on issues I'm dealing with, and despite there being a log that we all know about, sometimes I get zero response.
20:21 celeron55 there's always somebody here reading the backlog
20:21 VanessaE mind you it doesn't bother me too much
20:21 celeron55 if the backlog is boring or it's an issue nobody thinks they can solve, then people ignore it
20:21 VanessaE but I bet you it bothers the everloving shit out of him
20:22 celeron55 but in that case it wouldn't go forward by not ignoring it either
20:24 thexyz https://gist.github.com/xyzz/da52765655e45ae525e9
20:24 proller i better fix something or write something new instead of describing every line of pull to every dev and exio
20:24 VanessaE 22:27 <   celeron55> | if nobody is interested, [...] 3) code more to have a more impressive bunch of stuff
20:25 VanessaE and therein lies the problem.
20:25 thexyz as you can clearly see he didn't even try to communicate
20:25 VanessaE (no offense, c55....)
20:25 proller i always fix all reasonable objections in all pulls
20:25 celeron55 VanessaE: you're free to criticize me
20:25 VanessaE celeron55: I'm just saying that you literally told him to do exactly what he's done...
20:25 celeron55 VanessaE: 3 isn't a primary option, as you can guess
20:25 sapier NO proller you're working in a team if team doesn't understand what your doing your work isn't worth anything
20:26 bbqroast joined #minetest-dev
20:26 thexyz the "team" doesn't even try or so it seems
20:26 thexyz to communicate, that is
20:26 sapier I know explaining obvious things to others is annoying but there's no other way
20:26 VanessaE celeron55: of course, I know that.  but it seems to me that options 1 and 2 have been exhausted in this case.
20:27 thexyz how the hell is he supposed to know what to improve if you don't communicate
20:27 proller and my pull do not broke old liquids, except one place where need to make decision (drop float flag in old liquids OR leave 8 levels for old liquids)
20:29 sapier thexyz you keep telling noone told him anything just plain repeating doesn't make this more correct than first time
20:29 proller and for my current work in dynamic branch need 20-30 pulls is split by feature
20:29 celeron55 well i definitely now said what is generally wrong with #895
20:29 thexyz sapier: some people still seem to not understand what I meant
20:29 thexyz i.e. this ^^
20:30 thexyz "i said this so everyone now knows that"
20:30 Calinou joined #minetest-dev
20:31 proller and i have no time to rewrite 50% engine for fix every small problem
20:31 sapier ok now what's up to 895: 1) multiple features, 2) some missing explanations why to do things this way, 3) missing default parameter dedicated_server_step ... I guess this isn't everything
20:31 VanessaE translation:  "stop dawdling and start merging so I can keep writing new stuff instead of repeatedly rewriting or rebasing my existing stuff"
20:32 proller all pull is about limit step to 0.1*x seconds. now step can take up to 60s
20:32 sapier and yes I know prollers pulls are a huge block of work I guess getting them in shape for merge is about half a month to a complete month
20:33 celeron55 that's why proller shouldn't aim for achieving so much
20:33 sapier proller I'm about to extract that feature right now could you please verify I didn't miss something?
20:33 celeron55 doing less but better would lead to much better results
20:33 proller sapier, where missing?
20:33 NakedFury joined #minetest-dev
20:34 proller i want to make game playable, and server running without problems and restarts every hour
20:34 sapier I don't know that was my question:
20:34 sapier https://gist.github.com/anonymous/7665717
20:35 sapier VanessaE we can't merge things if they aren't ready for merge just because proller keeps on writing new things ;-)
20:35 celeron55 proller: that's completely fine; but make it in as small steps as possible and don't overflow our slow but trustworthy voluntary review process
20:35 VanessaE sapier: I understand that.
20:36 proller celeron55, or do nothing?
20:36 sapier we all want to do our own work too and I guess It's only fair we don't only review and merge prollers changes in our freetime ;-)
20:37 sapier reviewing is even more time consuming than writing we will never keep up if proller doesn't stop writing from time to time ;-)
20:37 celeron55 proller: i can't change the facts
20:37 proller sapier, your gist diff is strange, how you get it?
20:38 sapier git diff ... but I'm open for suggestions to get a better one
20:38 proller in src/defaultsettings.cpp and minetest.conf.example  - not my changes
20:38 proller git diff upstream/master
20:38 sapier I know but it's required to add this setting if you use it
20:39 proller its already in master
20:40 sapier oh ... ok good to know
20:41 sapier https://gist.github.com/sapier/7665838 better?
20:42 proller better for what?
20:43 sapier better meaning "is this everything required to add this single feature?"
20:43 proller for what?
20:43 proller for remove lag - no
20:44 proller need whole pull
20:44 sapier "make a perfect game" isn't a single feature too
20:44 VanessaE proller: perhaps you should separate off the lag-fixing code into a new branch and pull-request just that
20:45 sapier and if you 100% guarantee this pull will remove any lag that will ever occurr I will merge it at once ... I just don't believe this is true
20:45 proller VanessaE, its already was here, but nothing happen
20:46 sapier proller no there never was a pull request containing performance fixes only (at least not since I am allowed to merge)
20:46 proller it remove all lags, which i found per 3 days of debugging
20:46 sapier I'm sorry to tell proller amount of time spent creating isn't a reason for merging something
20:47 proller 895 is all about lag
20:47 proller merging or not - its not my problem
20:48 sapier it also seems to fix liquid things which may be reason for lag too
20:48 proller you can always cherrypick anything from my branch
20:48 sapier I give up proller I really try to find a way to help but if you don't move a single mm I can't help
20:49 sapier I'm not willing to be your personal review and merge slave ...
20:50 sapier and obviously noone else is willing to be
20:51 bas080 joined #minetest-dev
20:53 sapier Anyone to review https://gist.github.com/sapier/7665838 and give permission to merge?
21:00 celeron55 uh what
21:00 celeron55 int received isn't initialized at all
21:01 sapier1 joined #minetest-dev
21:02 sapier1 https://gist.github.com/sapier/7666179
21:02 celeron55 fine to me
21:03 sapier1 the one problem with this is as it's a part of 895 merging it will make merging other things more difficult
21:52 Calinou_ joined #minetest-dev
22:20 OldCoder joined #minetest-dev
22:23 Zeitgeist_ joined #minetest-dev
22:24 sapier left #minetest-dev
22:25 Zeitgeist_ joined #minetest-dev
22:35 Zeitgeist_ joined #minetest-dev
22:35 fairiestoy joined #minetest-dev
22:42 OWNSyouAll joined #minetest-dev
23:22 iqualfragile joined #minetest-dev
23:31 Taoki joined #minetest-dev
23:47 RealBadAngel joined #minetest-dev
23:59 Gethiox joined #minetest-dev

| Channels | #minetest-dev index | Today | | Google Search | Plaintext