Time Nick Message 00:04 Hijiri could someone review #4685 00:04 ShadowBot https://github.com/minetest/minetest/issues/4685 -- Add control information to player interacts by raymoo 01:40 Zeno` paramat: #4766 01:40 ShadowBot https://github.com/minetest/minetest/issues/4766 -- Fixed tooltips not resizing with \n by lisacvuk 01:40 Zeno` oh... he's not here 01:40 Zeno` est31, hmmmm, what are your thoughts on my comment 06:07 hmmmm Zeno, I do agree that having a numerically comparable version is a good idea 06:09 Zeno` est31, the bug fix for that newline thing (in irrlicht source, not ours) is kind of a mess 06:10 Zeno` it's in a revision IIRC and not a MINOR version 06:10 est31 I still don't understand 06:10 est31 the PR 06:10 Zeno` i disagree with the PR because it will break 06:10 est31 see my comment 06:11 Zeno` just as my PR/commit broke 06:11 Zeno` yeah I don't understand that either 06:11 est31 why did it break? 06:11 Zeno` my hypothesis at one point was that some distros were patching irrlicht 06:12 Zeno` I don't know (I still don't know) because the irrlicht documentation and official source suggested it should not have broken (yet) 06:12 Zeno` but on some distros it has 06:12 est31 ?? 06:12 est31 you mean distros have changed it back? 06:12 est31 what exactly is the issue 06:12 Zeno` that is my hypothesis 06:12 est31 I have read that 1.8.2 _fixed_ the problem 06:12 Zeno` yeah it did 06:12 est31 so the ifdef is for irrlicht before 1.8.2 06:13 est31 why does it have to be changed to 1.8.4? 06:13 Zeno` but I am now on 1.8.4 (maybe 5) and the old behaviour is back 06:13 Zeno` I dunno\ 06:13 est31 wtf 06:13 est31 so the PR is wrong in any case then 06:13 Zeno` it's confusing me as well which is why I haven't written my own PR 06:13 est31 because it should check for 1.8.2 06:13 est31 and special case it 06:13 Zeno` I'll dig up my research notes 06:14 est31 1.8.5 doesn't exist yet 06:14 Zeno` but it was about a month ago when I last looked at it 06:14 Zeno` there needs to be a proper "number" comparison and /also/ an exception if things don't become clearer 06:15 est31 but its not the number failing here, right? 06:15 Zeno` for such a simple thing it's caused me a lot of confusion 06:15 Zeno` no, it's that the behaviour changed yet again 06:15 est31 this is stunning 06:16 est31 no mention of this in changes.txt 06:16 Zeno` my original commit updated to the latest behaviour observed and documented in 1.8.2 (?) and then 1.8.3 changed it all again 06:16 Zeno` yeah it's in the irrlicht change logs somewhere 06:16 Zeno` I hunted it down 06:16 est31 https://sourceforge.net/p/irrlicht/code/HEAD/tree/branches/releases/1.8/changes.txt 06:16 Zeno` it's stupid because it happened in a revision and not even a minor version 06:16 est31 - IGUIStaticText::getTextHeight returns now the correct height for texts with newlines even WordWrap is not set. 06:17 Zeno` yeah 06:17 Zeno` what version is that in 06:17 est31 1.8.2 06:17 est31 like your commit says 06:17 est31 https://github.com/minetest/minetest/commit/88acda02567d586820867059718cac551dc7a58b 06:17 Zeno` ok, and then 1.8.3 on fedora has changed it back to the old behaviour 06:17 est31 wtf 06:17 Zeno` let me just check what version I'm actually using 06:18 est31 1.8.3 almost doesn't differ from 1.8.2 06:18 Zeno` I'm on 1.8.4 06:18 est31 the only difference is that it includes a patch from me to fix compilation with gcc 5 06:18 Zeno` and the fedora supplied lib now behaves like it did before 1.8.2 06:19 Zeno` so I got very confused and kinda didn't know what to do about it 06:19 Zeno` pretty sure I mentioned it in here as well 06:19 est31 so, the ifdef should check for 1.8.2 and 1.8.3 and otherwise go to the else part? 06:19 Zeno` I haven't investigated other distros though 06:20 Zeno` but according to the irrlicht docs nothing should need changing at all (does it?) 06:20 est31 the patches fedora uses are public, I'm digging them up 06:20 Zeno` k 06:22 est31 http://pkgs.fedoraproject.org/cgit/rpms/irrlicht.git/tree/irrlicht18-optflags.patch 06:22 est31 http://pkgs.fedoraproject.org/cgit/rpms/irrlicht.git/tree/irrlicht-1.8-glext.patch 06:22 est31 http://pkgs.fedoraproject.org/cgit/rpms/irrlicht.git/tree/irrlicht18-libaesgm.patch 06:22 est31 http://pkgs.fedoraproject.org/cgit/rpms/irrlicht.git/tree/irrlicht18-fastatof-improvements-typefixes.patch 06:23 est31 http://pkgs.fedoraproject.org/cgit/rpms/irrlicht.git/tree/irrlicht18-irrXML-shared-library.patch 06:23 est31 http://pkgs.fedoraproject.org/cgit/rpms/irrlicht.git/tree/irrlicht-1.8-fix-locale-decimal-points.patch 06:23 est31 http://pkgs.fedoraproject.org/cgit/rpms/irrlicht.git/tree/irrlicht-1.8.1-mesa10.patch 06:23 est31 I think these were all patches they use 06:23 Zeno` I wondering now if the issue is somewhere else in the pipeline. I can't really see how now. I'm reinstalling the mod I use to test things to trace again 06:24 est31 http://pkgs.fedoraproject.org/cgit/rpms/irrlicht.git/tree/irrlicht.spec#n12 06:24 Zeno` I do know that I was shocked when I saw the old behaviour occurring occuring about a month ago 06:24 Zeno` and I also do remember making sure that the '\n' were not somehow being stripped etc etc etc 06:25 Zeno` what's && USE_FREETYPE == 1 for? 06:25 Zeno` could freetype be affecting things?! 06:26 Zeno` (in the PR I mean) 06:27 Zeno` maybe that's why fedora has changed... not because of irrlicht but because I ran my latest tests not using freetype 06:27 Zeno` that would explain my source of confusion but then the PR would/should only be adding && USE_FREETYPE == 1 and not messing with the version numbers 06:30 Zeno` building with just that change 06:32 Zeno` will have to do it twice. Once with and without freetype 06:33 Zeno` nah, see this is stupid 06:33 Zeno` now it works even without the PR 06:34 Zeno` I have no idea what's going on now 06:34 Zeno` on 1.8.2 it worked, last month on 1.8.3 it didn't work, now it's working again 06:40 Zeno` now I;m even more confused than before :( 06:40 Zeno` maybe 1.8.3 had a regression? 06:40 Zeno` that isn't documented? 06:40 Zeno` est31, ^ 06:41 est31 1.8.3 didn't change anything compared to 1.8.2 06:41 est31 almost nothing 06:41 est31 the only change is a build fix for gcc 5.2 06:41 Zeno` well I have nfi then :( 06:41 est31 #3130 06:41 ShadowBot https://github.com/minetest/minetest/issues/3130 -- segfault on startup for irrlicht 1.8.2 on arch linux 06:41 est31 the fix for this bug is the only difference 06:42 Zeno` well nothing makes sense :( 06:42 Zeno` will wait and see what lisacvuk says about their irrlicht version 06:43 est31 lisa seems to have 1.83 06:43 est31 1.8.3* 06:43 est31 but thats just confusing 06:43 est31 my guess is that the header and the binary don't match 06:43 est31 that it still links to irrlicht 1.8.1 or something, uing the 1.8.3 headers 06:43 Zeno` I *do* know that about a month ago the old behaviour had reappeared. I can find the exact date actually because I asked someone to email me to remind me to fix it 06:44 Zeno` 17-Sep-2016 06:44 Zeno` on that date on 1.8.3 the old behaviour was happening again (I tested) 06:44 Zeno` now, everything is fine again 06:44 Zeno` it's bizzarre 06:46 Zeno` it was when Amaz brought it up. I tested because I could not believe it. Tested it, was stunned because things were not as they should be, and got Amaz to email me 06:47 Zeno` it was regardinding #4142 06:47 ShadowBot https://github.com/minetest/minetest/issues/4142 -- Item description bug 06:47 Zeno` oh yuck 06:47 Zeno` I even fixed then phantom bug 06:48 Zeno` so my fix is the culprit: https://github.com/minetest/minetest/commit/88acda02567d586820867059718cac551dc7a58b 06:48 Zeno` no wait 06:48 Zeno` nah, that's right 06:49 Zeno` it's since then that the anomolies have cropped up 06:49 Zeno` amaz just used that PR as a reference to the bug that had resurfaced 06:51 Zeno` I bet that && USE_FREETYPE == 1 is a clue 06:51 Zeno` it should not behave differently with or without freetype 06:51 Zeno` it doesn't for me 06:55 Zeno` omg 06:56 * Zeno` starts kicking himself 06:57 Zeno` i forgot to rebuild after re-enabling freetype 06:57 Zeno` it *does* make a difference 06:58 Zeno` the correct (but not future-proof) code should be #if IRRLICHT_VERSION_MAJOR <= 1 && IRRLICHT_VERSION_MINOR <= 8 && IRRLICHT_VERSION_REVISION < 2 && USE_FREETYPE == 1 07:01 Zeno` || 07:01 Zeno` that's stupid 07:04 Zeno` updated comment in PR 07:04 Zeno` brb 07:04 Zeno` hi CWz 07:04 Zeno` brb 07:06 Zeno` no wonder I got confused. I had no idea (at all!) that freetype and no freetype would make a difference 07:07 est31 Zeno`: which does bind stricter && or || ? 07:07 est31 have you added parens to confirm? 07:07 Zeno` I'll add parens now 07:07 Zeno` but without parens it works with and without freetype 07:07 Zeno` I'll add parens for clarification anyway 07:08 CWz Zeno`, the lizard people too over the US 07:08 hmmmm Zeno`: this is stupid. make a single version code to compare against 07:08 Zeno` hmmmm, I know 07:09 Zeno` see my earlier comment 07:09 Zeno` but that isn't the issue 07:09 hmmmm the freetype making a difference is the issue? 07:09 Zeno` the issue was/is the freetype vs. non-free type 07:09 Zeno` yes 07:09 hmmmm hrmm 07:09 Zeno` which is why I was got so f'ing confused 07:10 Zeno` est31, parens make no difference. I'll add them for clarity anyway 07:10 est31 thanks 07:10 Zeno` hmmmm, wtf freetype has to do with anything... nfi 07:10 Zeno` but it does... i've just tested 4 times heh 07:11 hmmmm might be some kind of outdated code in cguittffont 07:15 Zeno` perhaps 07:15 Zeno` it's not something I even thought to check :( 07:16 Zeno` I expected for some reason that the behaviour would be the same with or without freetype 07:16 Zeno` lesson learned 07:17 Zeno` just recompiled twice more and the (now updated) line in my latest comment in #4766 is correct 07:17 ShadowBot https://github.com/minetest/minetest/issues/4766 -- Fixed tooltips not resizing with \n by lisacvuk 07:19 Zeno` est31, also && has higher precedence than || (obviously since it's been checked) but I just checked the standard as well for thoroughness :D 07:22 Zeno` hmmmm, did you want to check the bundled cguifont? 07:22 hmmmm not really, if it's fixed, it's fixed 07:23 Zeno` well, it'll fix it yeah so *shrug* 07:23 Zeno` After this I'm going to convert to the "single version code" method as well 07:24 Zeno` otherwise in several years when we all might be gone further confusion may arise 07:24 Zeno` 'cause minetest will be around forever! 07:29 Zeno` what a silly thing 07:29 * Zeno` chalks it down as a learning experience 07:30 Zeno` it means that we should test all text related changes twice though (yuck) 07:33 Zeno` Any objections on merging #4765 ? 07:33 ShadowBot https://github.com/minetest/minetest/issues/4765 -- Fix incorrect distance computation for visible blocks by Rogier-5 07:54 hmmmm i think the const variable is more idiomatic as static const, but that's really nitpicking 07:54 hmmmm approved 08:30 Zeno` I agree that static const is better (it's what I would prefer), but... yeah, *meh* 08:31 Zeno` I doubt it even matters with compilers post 2000 08:31 Zeno` but it's more style for me 08:56 Zeno` I think I might change the "Request" label to "Request/Suggestion" 08:56 Zeno` any thoughts on that? 08:59 nrzkt good idea 11:10 Zeno` ok will merge #4766 11:10 ShadowBot https://github.com/minetest/minetest/issues/4766 -- Fixed tooltips not resizing with \n by lisacvuk 20:37 sofar "Foghrye4, unless you publish your nude photographies here." 20:37 sofar LOL, I had completely missed that part of the thread :D 20:48 TranslucBugs thread? 20:58 sofar the particle enhancements he posted 22:40 TranslucBugs new fog is so much better and immersive 22:51 red-001 #4642 22:51 ShadowBot https://github.com/minetest/minetest/issues/4642 -- Allow the join/leave message to be overridden by mods. by red-001