**** BEGIN LOGGING AT Sun Jan 08 02:59:57 2012
Jan 08 04:28:45 <dtzWill> Brybry: https://github.com/PuffTheMagic/wTerm/pull/17 :D
Jan 08 04:34:58 <Brybry> I'll definitely be taking a look at that tomorrow :D
Jan 08 04:37:19 <dtzWill> \o/
Jan 08 04:37:22 <dtzWill> alright well i'm off
Jan 08 05:02:48 <PuffTheMagic> dtzWill, not sure if u have talked to me in the past few hours but i cant get on irccloud
Jan 08 05:02:53 <PuffTheMagic> which has all my backlogs
Jan 08 06:59:18 <dtzWill> PuffTheMagic: https://pastee.org/mf95u
Jan 08 06:59:34 <dtzWill> oh you already merged it
Jan 08 06:59:44 <dtzWill> hooray
Jan 08 16:06:31 <PuffTheMagic> great
Jan 08 16:06:51 <PuffTheMagic>  6 new bugs
Jan 08 16:06:55 <PuffTheMagic> i had the list down to 3 ;)
Jan 08 16:07:14 <PuffTheMagic> dtzWill, i know the code that causes the top line disappearance in wterm
Jan 08 16:07:19 <PuffTheMagic> i just dont know what its a bug
Jan 08 16:07:23 <PuffTheMagic> cause the code seems right
Jan 08 16:11:20 <dtzWill> PuffTheMagic: oh, cool.  where'd you track it down to?
Jan 08 16:11:48 <PuffTheMagic> i commented on your issue
Jan 08 16:11:56 <dtzWill> :D
Jan 08 16:18:20 <PuffTheMagic> that particular escape sequence is non vt100/102
Jan 08 16:18:24 <PuffTheMagic> its more like xterm/rxvt
Jan 08 16:18:28 <PuffTheMagic> use this site http://rtfm.etla.org/xterm/ctlseq.html
Jan 08 16:22:09 <PuffTheMagic> its actually a vt200 command
Jan 08 16:24:26 <dtzWill> oh ty for the link
Jan 08 16:26:11 <PuffTheMagic> dtzWill, your changes from yesterday are wicked fast
Jan 08 16:27:14 <dtzWill> :D good to hear! is that manifesting in any particular way?
Jan 08 16:27:52 <PuffTheMagic> i have my own color script
Jan 08 16:27:57 <PuffTheMagic> that i test with
Jan 08 16:28:03 <PuffTheMagic> and its faster
Jan 08 16:28:06 <PuffTheMagic> dmesg is way faster
Jan 08 16:28:19 <dtzWill> (and given the time spent on that, it's a pleasant result hehe.  can't always rely on that time-to-success ratio to pull out for ya)
Jan 08 16:28:30 <dtzWill> yeah afaict we "can't" render much faster than it is now
Jan 08 16:28:46 <dtzWill> optimizations along the line of some variant of dirty-region tracking would of course help
Jan 08 16:29:16 <dtzWill> but instead just made drawing the full screen super cheap and figured if somehow that wasn't enough we could do some kind of dirty tracking later
Jan 08 16:30:15 <dtzWill> but ty for the report, i was unable to show how much faster it really is in my testing
Jan 08 16:30:45 <PuffTheMagic> dtzWill, so since you are playing with font stuff i want to run something past you
Jan 08 16:31:27 <PuffTheMagic> the orig author had fonts for bold and italics
Jan 08 16:31:34 <PuffTheMagic> not vt100 doesnt have italics
Jan 08 16:31:46 <PuffTheMagic> but for bold maybe we want to do more than just brighter text
Jan 08 16:31:47 <PuffTheMagic> idk
Jan 08 16:31:49 <PuffTheMagic> just a though
Jan 08 16:32:00 <PuffTheMagic> that would probably slow down rendering
Jan 08 16:32:03 <dtzWill> heh yeah i was wondering about that, the existing code *should* work fine with those fonts
Jan 08 16:32:28 <dtzWill> nopers, i render all 4 fonts (lazily) into a glyph cache texture, i'd expect zero slow-down
Jan 08 16:32:42 <dtzWill> *into the same texture
Jan 08 16:33:08 <dtzWill> so the GPU would do zero work, and i already compute offset into tetxure based on which font we're using....
Jan 08 16:33:23 <dtzWill> so just setting bBold and bItalic accordingly should do the trick
Jan 08 16:33:33 <PuffTheMagic> well if u want to play with that mr font man go ahead
Jan 08 16:33:57 <dtzWill> well, i'm not sure when would be logically appropriate to make use of those fonts
Jan 08 16:35:45 <dtzWill> aka if you have ideas when bold might be put to good use i'd be happy to look into it, but not sure i have any myself
Jan 08 16:36:43 <PuffTheMagic> bold would be use when the bold/bright graphics property is enabled
Jan 08 16:37:47 <dtzWill> btw on issue #23 (htop/CS_ECH bug) when you say that's the problem what do you mean--does it _not_ erase the requested character(s)? or does it do so ..incorrectly?
Jan 08 16:39:21 <dtzWill> PuffTheMagic: oh, okay.  and that's already tracked too it seems just needs to be hooked up to the rendering? (half thinking out-loud...)
Jan 08 16:40:37 <dtzWill> does 'italic' have a place? can't think of a time my text has been italic
Jan 08 16:40:48 <PuffTheMagic> i dont think so
Jan 08 16:41:03 <PuffTheMagic> dtzWill, the other thing that sorta needs to happen is unicode/utf-8 support
Jan 08 16:41:11 <PuffTheMagic> the buffer lines are only chars
Jan 08 16:41:16 <PuffTheMagic> which sorta makes this hard
Jan 08 16:41:26 <dtzWill> makes it very hard :/
Jan 08 16:41:31 <PuffTheMagic> i had to hack the fonts to get the box drawing chars work
Jan 08 16:41:47 <PuffTheMagic> but if we could support unicode then i could just point to the right glyphs
Jan 08 16:41:51 <PuffTheMagic> and it would work on every font
Jan 08 16:41:53 <PuffTheMagic> with out hacking
Jan 08 16:42:28 <dtzWill> what do you mean re:box drawing/hack the fonts? (is there a commit i should look at?)
Jan 08 16:43:09 <PuffTheMagic> midnight commander
Jan 08 16:43:16 <PuffTheMagic> know how it has nice lines for the boxes
Jan 08 16:43:34 <PuffTheMagic> terminals support a handful of alternate character sets
Jan 08 16:43:45 <PuffTheMagic> and one of them is designed for drawing lines and boxes
Jan 08 16:49:28 <PuffTheMagic> dtzWill, http://paste.pocoo.org/show/531837/
Jan 08 16:50:31 <dtzWill> is that different than https://gist.github.com/1577147 ? (other than the 1000x loop, lol)
Jan 08 16:50:55 <PuffTheMagic> oh prob not
Jan 08 16:51:18 <PuffTheMagic> it was mad late when i saw your script
Jan 08 16:51:36 <PuffTheMagic> and i was sorta drained from the wake earlier
Jan 08 16:52:13 <dtzWill> oh np at all, i'm just thought i'd ask before i tried to diff them or something :)
Jan 08 16:53:04 <PuffTheMagic> lol
Jan 08 16:53:35 <dtzWill> yay bold for bright works
Jan 08 16:53:41 <PuffTheMagic> nice
Jan 08 16:53:48 <dtzWill> it's presently both bold *and* slightly brighter (as it is now)
Jan 08 16:54:06 <PuffTheMagic> though as i think about it, its probably not so portable
Jan 08 16:54:15 <dtzWill> what do you mean?
Jan 08 16:54:16 <PuffTheMagic> if we are going to allow people to specify their own fonts downt he road
Jan 08 16:54:40 <PuffTheMagic> like i have a handful of mono fonts in preware that i installed for Tide
Jan 08 16:54:49 <dtzWill> we can just do what other terminals do and let people choose a font for normal text, font for bold, etc
Jan 08 16:54:59 <PuffTheMagic> ahh
Jan 08 16:55:02 <PuffTheMagic> good call
Jan 08 16:55:33 <PuffTheMagic> well Konsole doesnt
Jan 08 16:55:42 <dtzWill> oh :(
Jan 08 16:55:45 <PuffTheMagic> doesnt have a place to specify 2 fonts
Jan 08 16:55:52 <PuffTheMagic> but it does have a checkbox for bold as bright
Jan 08 16:55:59 <dtzWill> well urxvt does, via .Xdefaults --there's a property for 'font' as well as 'boldFont'
Jan 08 16:56:12 <dtzWill> also confirmed bold has no performance hit which is good
Jan 08 16:56:26 <PuffTheMagic> i would say that we could make this a checkbox option like konsole
Jan 08 16:56:26 <dtzWill> well afaict anyway, but like i said originally it shouldn't
Jan 08 16:56:37 <PuffTheMagic> and then if we can find the corresponding bold font we can use it
Jan 08 16:56:46 <PuffTheMagic> else we fallback to just bright
Jan 08 16:56:51 <dtzWill> (i actually didn't realize that we didn't use bold/italic fonts, so supported them anyway xD)
Jan 08 16:56:55 <dtzWill> ah
Jan 08 16:56:59 <dtzWill> sure thing.
Jan 08 16:57:17 <dtzWill> i'll push this in a sec anyway, i like bold on bright now that you mention it
Jan 08 16:57:35 <PuffTheMagic> ya im sure it will look good
Jan 08 16:57:50 <PuffTheMagic> im gonna file a bug for the wide char support
Jan 08 16:58:03 <PuffTheMagic> cause that is probably a top priority for a number of features
Jan 08 16:58:24 <PuffTheMagic> i just dont if we should use wchar_t or c++ wstring
Jan 08 16:59:35 <dtzWill> i know almost nothing about that kind of support (see? I can't even give it a good name) unfortunately
Jan 08 17:00:09 <PuffTheMagic> i tried porting everything to wchar_t twice now
Jan 08 17:00:13 <PuffTheMagic> and it broke parsing
Jan 08 17:00:23 <PuffTheMagic> so I will have to take a different approach
Jan 08 17:07:31 <dtzWill> d'oh, and yeah not having any experience means it's all the fun of learning string manip all over again only much harder :P
Jan 08 17:09:11 <dtzWill> (also the rendering stuff will break since it assumes things about printable chars)
Jan 08 17:09:29 <dtzWill> if that ends up being an issue i can fix that as needed down the line
Jan 08 17:12:34 <PuffTheMagic> well are u no longer using the TTF functions?
Jan 08 17:12:40 <PuffTheMagic> cause there are UNICODE variants of them
Jan 08 17:12:58 <dtzWill> oh it's not a nightmare, just non-trivial
Jan 08 17:13:40 <dtzWill> for example i store a font pre-rendered into a texture as a single row--because i happen to know i can do so with all 223 printable characters without exceeding the texture size limit in that dimension
Jan 08 17:14:20 <dtzWill> when you introduce unicode (how many unique glyphs does that give us?) we'd probably have to do something more clever
Jan 08 17:14:52 <PuffTheMagic> well im not thinking full blown utf8 support
Jan 08 17:14:59 <PuffTheMagic> but more for special line drawing chars
Jan 08 17:15:03 <PuffTheMagic> and for keyboard layouts
Jan 08 17:15:09 <PuffTheMagic> so it would still be sorta a fixed set
Jan 08 17:15:21 <dtzWill> oh, okay.
Jan 08 17:15:50 <dtzWill> well it's solvable either way just not quite a one-liner and i was offering to maintain my code with respect to future wterm improvements :)
Jan 08 17:23:27 <PuffTheMagic> well if custom fonts are going to be an option down the road I need to be able to select the line drawing chars from upper level unicode ranges
Jan 08 17:23:48 <PuffTheMagic> and if we want international keyboards we also need support for wide chars
Jan 08 17:24:04 <PuffTheMagic> and if we want actual utf-8 term we need it everywhere
Jan 08 17:24:16 <dtzWill> to me custom fonts mean "change the font to $FOO", i'm not sure i understand the connection to line drawing characters
Jan 08 17:24:44 <dtzWill> does one typically select a custom font and then map certain characters within that font for use as the various line symbols?
Jan 08 17:24:48 <PuffTheMagic> well traditionally with vt terminals they could only support 127 chars at a time
Jan 08 17:24:56 <PuffTheMagic> so there are special escape codes to load custom roms
Jan 08 17:25:10 <PuffTheMagic> which loaded different char sets into the standard 127 font slots
Jan 08 17:25:32 <dtzWill> oh.  interesting.
Jan 08 17:25:33 <PuffTheMagic> now to support this with our current font, i used fontforge
Jan 08 17:25:45 <PuffTheMagic> and copied the line drawing chars into the extended ascii range
Jan 08 17:25:51 <dtzWill> lol! oh wow
Jan 08 17:25:54 <PuffTheMagic> and when the right escape code comes in
Jan 08 17:26:00 <PuffTheMagic> i subtract 128 from the char/int
Jan 08 17:26:07 <PuffTheMagic> and thats how i get the special chars
Jan 08 17:26:32 <PuffTheMagic> but it would be much easier to just point to the actual unicode char
Jan 08 17:26:40 <dtzWill> ahh okay. yeah.
Jan 08 17:26:41 <PuffTheMagic> in a switch statement or something
Jan 08 17:27:17 <PuffTheMagic> and that would work with fairly complete unicode font
Jan 08 17:27:21 <PuffTheMagic> like dejavu
Jan 08 17:27:24 <PuffTheMagic> and bitstream
Jan 08 17:35:44 <dtzWill> PuffTheMagic: where'd you put things like setting TERM=xterm? /etc/profile.d ?
Jan 08 17:35:57 <PuffTheMagic> ya
Jan 08 17:36:04 <PuffTheMagic> i have a file in there called term
Jan 08 17:36:07 <PuffTheMagic> which sets TERM
Jan 08 17:36:29 <PuffTheMagic> although I am thinking of changing it
Jan 08 17:36:38 <PuffTheMagic> so that it sets a random term each time i log in
Jan 08 17:36:43 <PuffTheMagic> that way i set more setups
Jan 08 17:37:14 <dtzWill> xD
Jan 08 17:37:22 <PuffTheMagic> s/set/test/
Jan 08 17:37:31 <dtzWill> well ty, did the same.  got damn annoying to set it every time
Jan 08 17:37:47 <PuffTheMagic> sure does
Jan 08 17:37:55 <dtzWill> esp since whatever bash i'm using doesn't support control-r
Jan 08 17:38:10 <PuffTheMagic> what does that do?
Jan 08 17:38:14 <PuffTheMagic> irssi,mc htop i my main test apps
Jan 08 17:38:23 <PuffTheMagic> though i need to start testing with ssh
Jan 08 17:38:41 <PuffTheMagic> cause that seems to break things
Jan 08 17:38:48 <dtzWill> control-r searches backwards in your history
Jan 08 17:39:02 <PuffTheMagic> ahh
Jan 08 17:39:10 <PuffTheMagic> i've dont that once or twice by accident it hink
Jan 08 17:39:22 <dtzWill> very useful if you frequently use a set of say 10 commands (i often do while dev/testing) and you end up doing things like "control-r, m" should pull up last make command, etc
Jan 08 17:39:31 <dtzWill> but just generally useful as well
Jan 08 17:39:40 <dtzWill> anyway hooray for TERM
Jan 08 17:39:46 <PuffTheMagic> lol
Jan 08 17:39:50 <dtzWill> need to get ncmpcpp onto my touchpad
Jan 08 17:39:57 <PuffTheMagic> whats that
Jan 08 17:40:01 <dtzWill> there's nmpc in optware but
Jan 08 17:40:14 <PuffTheMagic> ahh mpd
Jan 08 17:40:16 <dtzWill> command line mpd client
Jan 08 17:40:16 <PuffTheMagic> nice
Jan 08 17:40:17 <dtzWill> yeah
Jan 08 17:40:24 <dtzWill> very purdy
Jan 08 17:40:35 <PuffTheMagic> i've been thinking of making a xmm2/mpd client
Jan 08 17:41:01 <PuffTheMagic> i want to make way too many things
Jan 08 17:41:35 <PuffTheMagic> is ncmpcpp that much better than ncmp?
Jan 08 17:41:53 <dtzWill> i haven't used ncmp extensively so don't remember offhand
Jan 08 17:42:34 <dtzWill> huh looks like it might not be, in terms of things i actually use. cool.
Jan 08 17:51:20 <dwc-> I use ^R for finding various longer things from history.  otherwise it's !ma !scp etc
Jan 08 17:56:16 <PuffTheMagic> dtzWill, you are on a commit roll, thank you very much these performance improvements, i've been so focused on base compliance stuff and other features idk if I would have gotten around to these
Jan 08 18:27:05 <dtzWill> PuffTheMagic: very welcome--makes me sad we didn't team up on a terminal long time ago xD :)
Jan 08 18:27:55 <dtzWill> ah HA
Jan 08 18:27:55 <PuffTheMagic> well everyone had different ideas of how they wanted the term implemented
Jan 08 18:28:01 <dtzWill> i found you you damn bug
Jan 08 18:28:02 <dtzWill> *ahem*
Jan 08 18:28:15 <PuffTheMagic> some were dead set on making the old terminal work and didnt want to work on a new one
Jan 08 18:28:19 <PuffTheMagic> others found xterm
Jan 08 18:28:20 <PuffTheMagic> ;)
Jan 08 18:28:21 <dtzWill> well narrowed it down anyway hopefully fix isn't long off
Jan 08 18:28:30 <dtzWill> yeah, but that was always just a POC that we had an x11 server
Jan 08 18:28:34 <PuffTheMagic> the erase chars bug?
Jan 08 18:28:44 <dtzWill> which is cool if you wanna do X stuff, but less so if you just want a damn prompt
Jan 08 18:28:46 <dtzWill> PuffTheMagic: yeppers
Jan 08 18:31:26 <PuffTheMagic> dtzWill, was it simple math error
Jan 08 18:31:37 <PuffTheMagic> or was it a mis interpretation of what the command should do
Jan 08 18:32:06 <dtzWill> it's a bug in databuffer.cpp, at least i'm pretty sure.  fixing/testing atm, if i'm right you'll see the commit soon enough xD
Jan 08 18:32:30 <PuffTheMagic> well i wont be able to merge till later
Jan 08 18:32:37 <PuffTheMagic> going to wake/viewing part 2
Jan 08 18:33:49 <dtzWill> oh i mean there's no hurry to merge anything, i just meant you'll be able to see it if you were so inclined
Jan 08 18:35:27 <dtzWill> oh i see
Jan 08 18:39:08 <dtzWill> okay problem is that erasing characters sets their buffer implementation to 0 (NULL, not '0')
Jan 08 18:39:36 <dtzWill> and various pieces of the code (sdlterminal, sdlcore? and sdlglfont) assume the data is a string
Jan 08 18:40:09 <dtzWill> i'm thinking we should just erase to blanks instead (instead of restructuring the rest to handle byte-arrays instead of strings (passing len instead of assuming null-terminated))
Jan 08 18:40:22 <dtzWill> but it's not the bug i was hoping for (silly miscalculation, as you suggested)
Jan 08 18:40:48 <Brybry> dtzWill, you could also just drop a .profile file in /var/home/root with export TERM=xterm
Jan 08 18:44:41 <dtzWill> htop works :D
Jan 08 19:15:44 <Brybry> screen + TERM=xterm do not play nicely =(
Jan 08 19:17:54 <Brybry> actually, I think screen might be borked entirely in the latest code
Jan 08 19:20:36 <Brybry> I wonder how far back that goes
Jan 08 19:38:06 <rwhitby> dtzWill: are you watching http://forums.webosnation.com/webos-internals/224974-visualboyadvance-new-post.html ?
Jan 08 19:44:58 * rwhitby bbl
Jan 08 21:55:35 <PuffTheMagic> dtzWill, nulls?
Jan 08 21:55:36 <PuffTheMagic> really
Jan 08 21:55:39 <PuffTheMagic> :(
Jan 08 22:31:12 <dtzWill> rwhitby: no, i hadn't hit that yet and ty for the ping
Jan 08 22:35:11 <dtzWill> Brybry: seeing strangeness here too
Jan 08 22:39:03 <PuffTheMagic> Brybry, ssh plus wterm is also not so nice
Jan 08 22:47:12 <Brybry> I enjoy hitting capslock on my bluetooth keyboard and watching the light come on my usb keyboard :3
Jan 08 22:48:29 <Brybry> screen works fine with vt100, I had just forgotten to reset
Jan 08 22:49:17 <Brybry> just TERM=xterm has that weirdness
Jan 08 22:50:00 <Brybry> is there a reason we send 0x007F for backspace instead of ^H/0x0008?
Jan 08 22:51:14 <Brybry> oh, hmm, 0x007F is supposed to be delete
Jan 08 22:51:15 <Brybry> interesting
Jan 08 22:51:47 <Brybry> SDL officially makes no sense
Jan 08 23:03:48 <PuffTheMagic> Brybry,
Jan 08 23:03:56 <PuffTheMagic> yes there is
Jan 08 23:04:06 <PuffTheMagic> 08 is the ascii key for backspace
Jan 08 23:04:11 <PuffTheMagic> but linux consoles send 7f
Jan 08 23:04:42 <PuffTheMagic> in most shells both will delete a char as expected
Jan 08 23:05:02 <PuffTheMagic> but in many apps 08 will insert ^H instead of deleting the char
Jan 08 23:05:15 <PuffTheMagic> but this is often a keybinding in most terms that people can set
Jan 08 23:07:23 <PuffTheMagic> Brybry, and each app is hit or miss with what is "currently" the best TERM to use
Jan 08 23:07:42 <PuffTheMagic> if an app uses termcaps or terminfo
Jan 08 23:08:02 <PuffTheMagic> they usually default to using newer stuff
Jan 08 23:08:09 <PuffTheMagic> other apps just hardcode stuff
Jan 08 23:08:24 <PuffTheMagic> there is a bunch of basic vt100 stuff that is still not perfect
Jan 08 23:08:50 <PuffTheMagic> and there is also a few vt220/xterm thing that arent done
**** ENDING LOGGING AT Mon Jan 09 02:59:57 2012