Time  Nick        Message
14:25 Taoki       celeron55_: http://minetest.net/forum/viewtopic.php?pid=55888 More people are starting to ask about using shaders to add post-processing like in those images. Any chance you might look into adding a second render pass during the next days? If that's added I can try to find some freebie GLSL filters and add them to the shader file, once there is one for that
14:26 Taoki       days / weeks / etc. It's ok if you can't but that would be a really nice thing
15:15 celeron55_  Taoki: no
15:17 Taoki       ok
15:18 celeron55_  the main thing in those images is lighting and shadows
15:19 celeron55_  you're not going to get anything like that with post-processing
15:22 Taoki       Well bloom is a post-process, since it's not applied to speficic textures
15:22 Taoki       As well as light rays
15:23 celeron55   you can't have meaningful bloom or light rays without first having some kind of a thing that emits "brighter than the screen can show" light in such a way that it's detectable in a post processing step
15:24 Taoki       ah, right
15:25 Taoki       For bloom you can, since you can just over-brighten and blur bright areas of the image at each step
15:25 celeron55   that would be horrible
15:25 Taoki       With that render pass
15:25 Taoki       Why? I use it in photoshop on images and it looks nice
15:25 celeron55   it needs more than that to not be only a huge headache
15:27 celeron55   well i guess you can do something
15:27 celeron55   at least i wouldn't use it...
16:06 thexyz      celeron55: https://github.com/celeron55/minetest/pull/344
16:07 celeron55   lol, just as if i didn't have anything to do tonight already 8)
16:07 thexyz      k
16:07 celeron55   i'll look at it
16:08 celeron55   where is the font from?
16:08 celeron55   i don't see any license additions for it
16:09 thexyz      it's dejavu sans mono
16:09 thexyz      same as the one used in console
16:12 VanessaE    http://dejavu-fonts.org/wiki/License
16:13 VanessaE    if that's of any use.
16:13 celeron55   currently reading that
16:13 celeron55   seems quite useless; this is mostly just ambiguous to someone who hasn't dealt with legal texts for 10 years
16:14 VanessaE    well, debian distributes deja-vu sans, so the license is probably compatible with minetest.
16:15 VanessaE    (considering how license-obsessive they are :-) )
16:16 celeron55   i think a mention of the font being dejavu sans and a link to http://dejavu-fonts.org/wiki/License should be added to README.txt
16:17 celeron55   (or, well, whatever the proper name for it is)
16:17 VanessaE    meh.  that License file ^^^ is the same one they include with the debian package
16:17 VanessaE    that's about useless.
16:17 VanessaE    (well in part it's the same)
16:18 thexyz      celeron55: the last readme.txt section is about DejaVu Sans Mono
16:18 thexyz      though it's missing this link
16:18 VanessaE    celeron55: it looks like your idea should work.  link to the license from README.
16:18 celeron55   hmm, i see
16:19 celeron55   i guess the existing text can suffice
16:19 VanessaE    as I read it:  make any change you want and distribute it as you see fit, as long as you include the copyright notice and don't use the words "bitstream" or "vera" in the names.
16:23 Calinou     "don't use the words "bitstream" or "vera" in the names" < huh, why
16:24 Calinou     i know a lot of games where fonts are called bitstream/vera
16:24 VanessaE    because bitstream vera is a font of its own
16:24 VanessaE    deja-vu sans is a modified version of it, and the Bitstream team claims trademark on the name.
16:24 VanessaE    "All Rights Reserved. Bitstream Vera is a trademark of Bitstream, Inc."
16:25 VanessaE    so telling downstream packagers to just use a different name was probably the easiest way to avoid a conflict or confusion without restricting derivative works.
16:26 celeron55   oh dog
16:26 celeron55   more utf-16
16:26 celeron55   i don't understand why somebody even invented utf-16
16:26 celeron55   it should burn deep in hell
16:26 VanessaE    the rest of this reads almost like a modified version of the warranty disclaimer included with most GNU software.
16:27 celeron55   it has all the cons of utf-32, AND it has all the cons of utf-8
16:27 VanessaE    heh
16:27 celeron55   and no pros whatsoever 8D
16:28 celeron55   i guess it's fine as long as we have all protocols and files in utf-8
16:29 VanessaE    well, aside from being just plain overkill for this project, is there any functional negative for using a UTF-16-based resource?
16:29 VanessaE    (and aside from it obviously being a pain in the ass in general ;-) )
16:29 celeron55   utf-16 is bad because it's not compatible with ascii and is not 8-bit, yet still it has the extended characters (which are familiar from utf-8) because 16 bits will not fit all of unicode (not even close)
16:30 celeron55   noobs like it because it will usually work for many languages even if you use it like a fixed 2-byte character set
16:30 celeron55   but it's not
16:30 VanessaE    "not compatible with ASCII" and they thought it was a good idea?
16:31 celeron55   if you want to use a character set like that, utf-32 is the only one that will do the job
16:31 VanessaE    wtf did they want to work with, EBCDIC?
16:31 celeron55   it is not compatible with ascii, but it is fixed width
16:32 celeron55   i guess utf-16 was made because windows uses 16-bit characters
16:32 celeron55   or something like that...
16:32 VanessaE    meh.
16:33 celeron55   and irrlicht uses wchar_t, which can be 16 or 32 bit depending on the system 8D
16:33 celeron55   more namely, on windows it's 16, on *nix it's 32
16:33 celeron55   completely useless nonsense
16:34 VanessaE    "So jurisdictionally it's..."  "--a g*ddamn nightmare."
16:34 celeron55   with some luck, the things available in non-extended utf-16 characters will work for us
16:35 celeron55   the rest, whatever
16:35 celeron55   i won't even try
16:35 VanessaE    yeah.. There's really no need for anything outside of the basic Latin character set anyway in this game.
16:36 VanessaE    it's not like you're writing a word processor something.
16:36 celeron55   nope
16:36 celeron55   the chat and some mod need to support many languages
16:36 celeron55   mods*
16:37 VanessaE    hrm.
16:37 celeron55   the currently used latin-1 (or something like that) doesn't even support all european languages
16:37 celeron55   then there is russian and all that
16:37 celeron55   greece, whatever
16:37 celeron55   supporting chinese and japanese is a good idea too
16:38 VanessaE    *looks at basic multilingual plane*
16:38 celeron55   but those do probably require some rarer characters from the higher parts of unicode, not sure
16:38 VanessaE    all of this shit looks like the same stuff UTF-8 already has..
16:39 VanessaE    (ignoring the multi-byte char encodings for a sec)
16:39 celeron55   utf-8 has all of unicode
16:39 celeron55   it just works for ascii too, and has no endianess problems and... ehm... is just the sane thing to use
16:40 celeron55   anyway, whatever; i hope this thexyz's thing works 8)
16:40 VanessaE    heh
16:41 VanessaE    famous last words? :D
16:41 thexyz      i've disabled it by default though
16:42 celeron55   it probably should automatically enable itself if freetype is available
16:42 celeron55   hmm... that would mean that the option is enabled by default but it just falls back to not using it if it's not available
16:42 celeron55   it's how gettext is used
16:43 celeron55   it always kicks in on linux but on windows it's not on anyone's way
16:43 celeron55   i guess nobody has heard any problems caused by that behavior of gettext in minetest?
16:44 celeron55   i haven't, at least...
16:44 VanessaE    the only text entry issue I've ever seen, you already know about (copy&paste).
16:44 VanessaE    everything else seems to work pretty much as one would expect.
16:45 VanessaE    of course that's just the user-facing stuff, I've no idea what goes on under the hood.
16:45 celeron55   by the way, the minetest network protocol does use 16-bit strings in chat messages...
16:45 celeron55   should make that utf-8
16:45 celeron55   VanessaE: that is up to irrlicht
16:46 VanessaE    right.  minor confusion there, don't mind me.
16:47 celeron55   // As linux is successfully switched to UTF-8 completely at about year 2005
16:47 celeron55   // Windows still uses obsolete codepage based locales because you
16:47 celeron55   // cannot recompile closed-source applications
16:47 * celeron55 smiles
16:47 VanessaE    heh
16:48 thexyz      celeron55: does minetest use utf-8 for formspec text?
16:49 celeron55   it should, but i wouldn't be surprised if it doesn't work
16:49 celeron55   the protocol uses 8-bit strings for them
16:53 VanessaE    bbl, time to head out.
16:55 celeron55   thexyz: src/guiChatConsole.cpp
16:55 celeron55   why does that need only the font filename while src/main.cpp uses a full path?
16:56 thexyz      oops
16:57 thexyz      oh, and by the way, where should the font.ttf be stored?
16:59 celeron55   probably client/fonts/font.ttf (client/ was added to contain shaders in client/shaders/)
16:59 celeron55   or, well
17:00 PilzAdam    to follow this textures/ should be moved to client/
17:00 celeron55   they kind of could be all flat in /shaders and /fonts (like /textures is), but i think that will create too many flat directories in the long run
17:01 celeron55   on the other hand they were intentionally made all flat (by me and kahrl) in the directory rearrangement that was done in the early days of 0.4...
17:01 doserj      don't you want mods to be able to supply fonts?
17:02 celeron55   that doesn't matter
17:02 celeron55   there always has to be a client-side default font
17:02 doserj      oh, sure
17:05 celeron55   i think we'll keep it all flat
17:06 celeron55   there won't be that many different things and it is fun to make people immediately see what the thing consists of 8)
17:07 celeron55   so this being the way, font.ttf should be /fonts/font.ttf, and /client/shaders/ should be moved to /shaders/
17:16 PilzAdam    whats about this: https://github.com/celeron55/minetest/pull/345
17:39 thexyz      celeron55: https://github.com/minetest/minetest/commit/efc3bb513c8c0411b2251ae652ffe7cf624db30f
21:27 celeron55   thexyz: i don't think that works
21:28 celeron55   oh...
21:28 celeron55   no, yes it should 8)