IRC logs for #opendune on OFTC at 2009-12-15
⏴ go to previous day
01:23:18 <DorpsGek> SVN: glx (r759) -Add: named more GUI functions
01:25:13 <glx> it seems f__B4A2_0039_000B_EC51 is some kind of GUI_MainLoop used by different windows
01:25:51 <glx> but I'm sure so I didn't named it
01:50:15 *** TrueBrain has joined #openDune
01:50:15 *** larich.oftc.net sets mode: +o TrueBrain
01:50:15 *** ChanServ sets mode: -o TrueBrain
02:00:24 *** TrueBrain has joined #openDune
02:14:08 *** TrueBrain has joined #openDune
02:23:56 *** TrueBrain_ has joined #openDune
02:26:38 *** TrueBrain has joined #openDune
02:40:52 *** ChanServ sets mode: +v Xaroth
09:36:00 <Xaroth> I hate setting up gateways
09:36:07 <Xaroth> especially when they can't be 'dumb' gateways
09:36:15 <Xaroth> but actually have to have 10 layers of firewalling
10:30:00 *** TrueBrain_ is now known as TrueBrain
11:19:39 *** planetmaker has joined #openDune
11:19:49 *** planetmaker is now known as Guest80
11:24:06 *** Guest80 is now known as planetmaker
12:24:18 *** proyvind has joined #openDune
12:36:46 *** proyvind has joined #openDune
12:54:19 *** proyvind has joined #openDune
13:21:04 *** proyvind has joined #openDune
14:17:32 <TrueBrain> proyvind: in or out, make a choice! :p
14:32:36 <TrueBrain> how is your GUI naming going?
14:32:54 <glx> nothing more than my last commit :)
14:33:14 <TrueBrain> haven't seen a commit ... stupid IRC :p
14:33:59 <TrueBrain> what pointer is given to those functions?
14:34:18 <glx> dunno, depends on the window
14:35:05 <TrueBrain> how did you find those functions?
14:35:19 <TrueBrain> oh, because of that big switch ;)
14:35:59 <glx> first was game options, which gave me options, which send me to the switch :)
14:36:08 <glx> then breakpoints and click on buttons :)
14:42:51 <glx> and I guess the unnamed functions in this switch are button handlers too :)
15:19:52 <DorpsGek> SVN: glx (r760) -Add: named more GUI functions
15:20:14 <glx> all the switch is named :)
15:23:38 <glx> Name is the name of the unit/structure (in command window)
15:23:51 <glx> click on it to center the view on selection
15:24:21 <Xaroth> :o and suddenly the naming gets busy :P
15:24:33 <TrueBrain> it was never not busy Xaroth :p
15:24:42 <TrueBrain> glx: did you check if these are all the functions called?
15:25:32 <TrueBrain> okay, if I read this correct, the 'csip' is the window pointer indeed ..
15:25:53 <TrueBrain> glx: I meant did you check if the switch could ever handle any other functions?
15:26:49 <glx> but I think all are found
15:27:11 <glx> else we would still get unresolved calls ;)
15:27:25 <TrueBrain> glx: it can be nobody pressed on here or there ;)
15:27:57 <glx> I clicked on all things indicated in the manual
15:28:14 <TrueBrain> 2E6B seems to contain the window information
15:28:57 <TrueBrain> at 02/04 is the click pointer
15:29:19 <glx> depends on the window I think
15:30:07 <TrueBrain> possible it first copies stuff there .. as I think it is 'widget' information
15:30:09 <glx> oops I forgot I should not do "find references" for a function
15:30:20 <TrueBrain> f__B4B8_0BC1_000C_1630 <- I think that creates a window
15:33:31 <TrueBrain> although never writes that memory, and it always runs from 0 to end
15:34:16 <TrueBrain> hmm ... all windows are created on startup? :s
15:36:25 <TrueBrain> f__B4B8_0ED9_001E_DC63 <- emu_Window_Allocate
15:36:33 <TrueBrain> f__B4B8_0BC1_000C_1630 <- emu_Window_Create
15:36:42 <TrueBrain> 0x2E6B, WindowInfo, 0x18 in size
15:37:42 <glx> ,...emu_push(emu_get_memory16(emu_ss, emu_bp, 0x8));
15:37:43 <glx> ,...emu_push(emu_get_memory16(emu_ss, emu_bp, 0x6));
15:37:43 <glx> ,...emu_lfp(&emu_es, &emu_bx, &emu_get_memory16(emu_ss, emu_bp, 0x6));
15:37:43 <glx> ,.../* Call based on memory/register values */
15:37:43 <glx> ,...emu_ip = emu_get_memory16(emu_es, emu_bx, 0x30);
15:38:11 <TrueBrain> 2e6B contains WindowInfo
15:40:08 <TrueBrain> any clue about the Window size?
15:42:29 <TrueBrain> then I really wonder if 'window' or 'widget' is the better name
15:42:34 <TrueBrain> as there can only be one click handler
15:42:38 <TrueBrain> and we have Ok and Cancel as seperate
15:42:42 <TrueBrain> so it seems they are Widgets
15:43:02 <TrueBrain> but very specific ones
15:44:27 <glx> hmm 0x32 is handler array, so it's really minimum size
15:45:03 <TrueBrain> 0x30/0x32 is a single pointer
15:45:08 <TrueBrain> so it can also be the max size
15:45:26 <glx> accessed as emu_es, emu_bx, 0x30
15:45:40 <TrueBrain> emu_lfp(&emu_es, &emu_bx, &emu_get_memory16(emu_ss, emu_bp, 0x6));
15:45:45 <TrueBrain> where that is the Window *
15:45:45 <glx> size depends on number of widgets
15:45:55 <TrueBrain> there is no 'add' or what ever
15:48:00 <TrueBrain> @base 10 16 [calc 0x18 * 2]
15:50:29 <TrueBrain> all have click actions
15:50:33 <TrueBrain> so they are of some kind of special type
15:50:51 <TrueBrain> I am going to call them 'WidgetClick'
15:52:47 <TrueBrain> but it does not list all the actions in that switch
15:53:44 <TrueBrain> only the ones in r760, and one other
15:57:56 <glx> this switch is used by different windows
15:58:16 <TrueBrain> Mentat is also in this array
16:03:46 <DorpsGek> SVN: truebrain (r761) -Add: named and structurized WidgetClickInfo array
16:03:51 <TrueBrain> 18 widgets listed :)
16:11:49 <TrueBrain> it is a nice widget system, if I look at it like this :)
16:12:48 <TrueBrain> I was wondering how it was all done, and I was expecting hard-coded drawing and shit
16:12:52 <TrueBrain> but .. this is better .. nicer ...
16:13:01 <TrueBrain> and more advanced then I expected
16:39:03 <glx> ok found more gui functions :)
16:55:46 *** ChanServ sets mode: +v Yexo_
17:08:58 <DorpsGek> SVN: glx (r762) -Add: named more GUI functions
17:26:05 <DorpsGek> SVN: glx (r763) -Add: named 2 more functions (GameOptions save/load)
17:33:34 <glx> f__B4F2_0C43_0017_56DB is decompiled but never called
18:03:52 <TrueBrain> functions are only decompiled when they are marked as 'start', which only happens when they are called
18:05:24 <TrueBrain> hmmm ... it seems I have 2 variants of B4F2 functions .. odd ..
18:06:42 <glx> maybe an old decompiler bug
18:07:19 <TrueBrain> possible .. or a current ;)
18:07:24 <TrueBrain> as the function looks sane and valid
18:08:37 <TrueBrain> lol, check B4F2_1303
18:09:45 <TrueBrain> offsets and stuff are identical
18:09:54 <TrueBrain> so yes .. 0C43 is fucked up, and should be ignored :p
18:13:48 <DorpsGek> SVN: truebrain (r764) [JIT] -Fix: remove wrong function
18:14:32 <DorpsGek> SVN: truebrain (r765) -Update (r764): update decompiled code
18:45:59 <DorpsGek> SVN: glx (r766) -Add: named one more function
20:01:14 <TrueBrain> I should be studying more .. but I am sick of studing ...
20:01:17 <TrueBrain> I either know this or not ..
20:01:21 <TrueBrain> and I guess it is 'not'
20:01:43 *** Xaroth sets mode: +o TrueBrain
20:19:14 <DorpsGek> SVN: truebrain (r767) -Add: figure out size of most Widget members
20:21:36 <TrueBrain> oh, it seems all the Widget_Allocates are clickable thingies ..
20:22:11 <TrueBrain> maybe I am just reading this all wrong
20:24:02 *** Alberth has joined #openDune
20:25:09 <Alberth> "Can you give priority to tech tree and unit stats in the C-fication?" that sentence made me laugh :)
20:25:36 <Alberth> live is still quiet here?
20:26:24 <Alberth> I guess 'life' indeed
20:26:25 <glx> TrueBrain: if you have time, you can use keyboard to navigate in mentat window (there are some unresolved dynamic jumps there)
20:26:47 <TrueBrain> and we are never quiet :p
20:27:12 <glx> and I guess it's easier to just "play" in JIT than generating crash.bin :)
20:31:03 <glx> 0x16 possible values (at most), only 2 known for now
20:31:12 <TrueBrain> glx: if you can take a look at B4A2, around label 07B9
20:31:14 <TrueBrain> it loads in -0x2/-0x4 a pointer, from either 0x10, 0x14 or 0x17
20:31:16 <TrueBrain> but they end up in drawing routines
20:31:29 <TrueBrain> then my network dropped .. lovely :p
20:31:41 <TrueBrain> either way ... any idea why a widget could have 3 draw procs?
20:31:47 <TrueBrain> wait ... normal, clicked, and ..?
20:33:17 <TrueBrain> in 2DCE the 'default' procs seem defined
20:33:52 <glx> and state "should" be 0x2E
20:34:22 <TrueBrain> so first and third bit
20:35:12 <DorpsGek> SVN: truebrain (r768) -Add: some more variable naming for Widgets
20:36:25 <TrueBrain> I found 5 more for the case
20:36:50 <TrueBrain> otherwise I ened keys which trigger them
20:37:38 <glx> maybe space, numpad5 and enter
20:39:01 <DorpsGek> SVN: truebrain (r769) [JIT] -Add: mapped another 21 functions (keyboard actions in mentat)
20:39:21 <DorpsGek> SVN: truebrain (r770) -Update (r769): update decompiled code
20:42:21 <glx> 7 different values in total :)
20:42:33 <glx> I just checked the "memory"
20:42:39 <TrueBrain> so we have them all now?
20:44:12 <TrueBrain> it is nice, that the keyboard pressed the buttons too
20:44:16 <TrueBrain> so they are handled by one functions
20:49:04 <TrueBrain> I think one of those 'states' is in fact 'text'
20:49:13 <TrueBrain> as it seems that -0x2 is given to DrawText in one situation
20:50:28 <TrueBrain> no idea when which is what
20:53:31 <Xaroth> TrueBrain: Chunk-based? savegames?
20:54:25 <glx> so there's a switch to select between 5 values, and for one of them another switch with known functions :)
20:55:30 <TrueBrain> k, starting from GameOptions seems to give more clues how this works ..
20:55:52 <glx> hehe, I used the same flow to name all this stuff ;)
20:56:18 <glx> GameOptions -> Options -> the rest
20:57:02 <TrueBrain> f__B4F2_0855_0013_A7FD <- Draw_GameOptions
20:57:09 <TrueBrain> (at least, when disabled, it doesn't :p)
20:58:08 <TrueBrain> f__B4F2_0EE0_000E_BC8E <- RemoveCurrentWindow
20:58:15 <TrueBrain> but that should make 0855 <- ShowCurrentWindow
20:58:20 <TrueBrain> which doesn't make sense :p
21:01:39 <glx> <@TrueBrain> f__B4F2_0855_0013_A7FD <- Draw_GameOptions <-- used by all windows in this segment
21:02:00 <TrueBrain> but what defines the labels?
21:02:09 <TrueBrain> it btw draws everything
21:02:28 <TrueBrain> wait, other pointers
21:03:15 <TrueBrain> 0EE0 seems to do nothing
21:03:59 <TrueBrain> either way, 0855 draws the complete window
21:05:04 <TrueBrain> f__B4A2_0039_000B_EC51 seemly monitors mouse click
21:05:29 <TrueBrain> f__B48B_00F2_0005_601A converts that to a value between 0 and amount of buttons
21:06:04 <TrueBrain> wait, f__B48B_00F2_0005_601A does something completely different :p
21:06:14 <TrueBrain> 'GetCurrentWindow' or what ever :p
21:06:59 <TrueBrain> B4A2_0039 gives a value starting from 0x1E
21:08:30 <TrueBrain> 353F:264C <- Options dialog
21:10:22 <TrueBrain> most likely B4F2_0EE0 makes a copy of the current screen
21:10:25 <TrueBrain> so when closing it can easy recover
21:10:30 <TrueBrain> but that is just a guess
21:11:15 <TrueBrain> or maybe it draws the main window?
21:12:55 <glx> 353F:26B5 GameOptions then
21:12:59 <TrueBrain> emu_push(emu_cs); emu_push(0x0939); emu_cs = 0x0FCB; emu_String_Get_ByIndex();
21:13:03 <TrueBrain> 337 / 0x151) There are no saved games to load.
21:13:30 <TrueBrain> but it shows up in 0855
21:14:19 <glx> as 0855 is used "locally"
21:15:17 <TrueBrain> it has exception code ;)
21:18:39 <TrueBrain> so they wrote this window manager for just a few windows :p
21:19:37 <glx> and mentat window is not a "real" window
21:19:42 <TrueBrain> French uses another font
21:19:47 <TrueBrain> we never played enough with French to map them all :p
21:20:20 <glx> IIRC french crash when I click on harvester in A1
21:21:59 <TrueBrain> okay .. the window appears to be just 7 bytes
21:22:05 <TrueBrain> after that are 15 bytes of 'widget' stuff
21:22:47 <TrueBrain> in ds:2A75 there is memory about the current widgets
21:22:57 <TrueBrain> or: a pointer to it :p
21:23:56 <TrueBrain> which seems to be of type Widget as in trunk
21:24:13 <TrueBrain> at least it puts the drawing routine to 34F2:61
21:25:51 <TrueBrain> seemly also does labels ;)
21:26:13 <TrueBrain> then this array only handles buttons :p
21:26:28 <TrueBrain> how to name f__B4F2_0855_0013_A7FD()? emu_Window_Draw?
21:27:40 <TrueBrain> text only is indeed handled differently
21:27:53 <TrueBrain> comes from the same array, but that function with the switch handles them differently
21:29:10 <glx> f__B4F2_0EE0_000E_BC8E backups screen under window
21:29:33 <DorpsGek> SVN: truebrain (r771) -Add: named 2 more GUI functions
21:29:49 <TrueBrain> went for emu_GUI_Option_CreateWindow for now
21:29:54 <TrueBrain> also named emu_GUI_Widget_Draw
21:30:14 <TrueBrain> I think next week orso I will do a lot of C-ifying of the window system, hoping to get a bit more knowledge about it
21:31:03 <glx> we already named a lot of it :)
21:31:19 <TrueBrain> but much more is needed :p
21:33:26 <TrueBrain> tried to reply to Skrilax_CZ in a nice way :)
21:34:08 <Xaroth> right, TrueBrain, which files were 'easy' files to c-ify?
21:34:18 <TrueBrain> not files, functions :p
21:34:29 <Xaroth> yeh, but surely there's a file full of those functions :P
21:34:30 <TrueBrain> emu_Unit_* in 1A34 should be doable
21:34:39 <TrueBrain> Deviate, Deviation_Decrease
21:34:45 <TrueBrain> SetDestionation, SetTarget
21:34:51 <TrueBrain> maybe even RemoveFog
21:35:40 <TrueBrain> just start with them very slowly
21:36:10 <Xaroth> O_O damage is quite big
21:36:10 <TrueBrain> at bp+6/bp+8 is always a Unit *
21:36:40 <TrueBrain> that is why Damage is _not_ in my list above :p
21:36:51 <TrueBrain> and that is why I started with Deviate ;)
21:37:15 <Xaroth> same as why you didn't mention FindStructure :P
21:37:31 <TrueBrain> in Deviate you will find a house can be more resistant to Deviation
21:37:37 <TrueBrain> also, that AIs have a higher chance to deviate
21:37:47 <TrueBrain> (or a lower, not sure :p)
21:38:08 <TrueBrain> also things you should/will notice:
21:38:16 <TrueBrain> 0x1E is the size of House
21:38:23 <TrueBrain> at 0x2D07 is g_unitInfo
21:39:09 <TrueBrain> the rest is pretty well named or other wise refered to
21:39:13 <TrueBrain> start slow, be consistant
21:39:59 <TrueBrain> all you need to know about f__B4CD_01BF_0016_E78F(), is that it removes the fog around the unit, updates the map, the viewport, shit like that .. just bookkeeping. Not of any interest to you except that it does that
21:40:25 <TrueBrain> that names all functions and variables in that function :)
21:41:36 <TrueBrain> hmm .. wasn't it a bug that a unit which un-deviates, still continues his current action? As the code has explicit thingies to avoid that
21:41:51 <Xaroth> if it un-deviates you can still have the command window open
21:41:57 <Xaroth> and complete giving it an order
21:42:02 <TrueBrain> shouldn't be in 1.07eu
21:42:04 <Xaroth> you can't click buttons anymore
21:42:11 <TrueBrain> oh, that might be possible
21:42:11 <Xaroth> but you can, for instance, target something
21:42:15 <TrueBrain> it untargets the unit ;)
21:42:24 <Xaroth> but the target cursor remains
21:42:29 <TrueBrain> when a unit deviates, it btw doesn't lose his current target :p
21:42:52 <Xaroth> if it auto-responds to being attacked it does
21:43:01 <TrueBrain> no, it never does :p Not _when_ deviated
21:43:13 <TrueBrain> it is only untargeted by _all_
21:43:40 <TrueBrain> so if I send a unit to attack (as H, to O)
21:43:43 <TrueBrain> they deviate my unit
21:43:47 <TrueBrain> it continues the attack :p
21:43:57 <Xaroth> yeh, but you can tell it to do otherwise :P
21:44:05 <TrueBrain> (if SetAction doesn't cancel it)
21:44:21 <TrueBrain> good luck Xaroth :)
21:44:43 <TrueBrain> glx: I really like this GUI stuff :) It is nice to unfold .. easier than Units / Structures :p
21:45:15 <TrueBrain> I guess much more well defined what it CAN do
21:45:16 <Xaroth> plus when we have all the UI stuff converted it's easy to go native SDL :)
21:45:35 <TrueBrain> that for sure .. and I think GUI stuff will be the first piece to move out of the emulated memory :p
21:45:42 <TrueBrain> much easier .. access is very well defined
21:45:48 <TrueBrain> but .. lot of work to do for that
21:46:21 <TrueBrain> I btw ran out of 1A34 functions to name :p
21:47:07 <glx> the one I named in it, you couldn't guess
21:47:54 <Xaroth> there's still a few unnamed :o
21:48:12 <TrueBrain> most are called rarely, or just from one function
21:48:21 <TrueBrain> almost at 1/3rd, so yes
21:50:25 <TrueBrain> and the longer I look at all this code, the less I have the need to C-ify :p
21:50:38 <TrueBrain> reading offsets is enough to have an idea what it means :p Sick .. I tell you .. sick
21:51:05 <glx> still it's easier to read when C-ified ;)
21:51:20 <TrueBrain> but I just read Deviate and Deviation_Decrease
21:51:23 <TrueBrain> and understood all of it :s
21:51:30 <glx> simplify a lot of emu_lfp stuff
21:51:38 <Xaroth> soon you'll be c-ifying random it straight away instead of naming it first :P
21:51:48 <TrueBrain> Xaroth: that is always possible, and that is how we started
21:51:57 <TrueBrain> GameLoops where all like that, for example
21:52:09 <TrueBrain> naming now only is easier, as for example the GUI is one big maze
21:52:23 <TrueBrain> and we also need to collect small functions for you, like Unit_Deviate :p
21:52:33 <Xaroth> nah, don't worry about me ! :P
21:52:39 <TrueBrain> nah; C-ifying takes (much) longer than reading through, and naming
21:52:45 <glx> I found GameOptions by luck
21:52:53 <glx> and it gave us a lot of info
21:53:23 <TrueBrain> C-ifying gets boring after a while :p
21:53:33 <TrueBrain> and now .. now I go to bed
21:53:35 <TrueBrain> hoping to sleep well
21:53:44 <TrueBrain> then I am done for this year
21:54:05 <glx> the year is quite finished ;)
21:54:27 <TrueBrain> Then I only have 2 pratical assignments I have to take care of before the 24th .. but okay :p
21:54:34 <TrueBrain> and 1 other big project ..
21:54:51 <TrueBrain> I will be at my parents .. no cooking .. no shopping .. a car ..
21:55:01 <TrueBrain> I will drown in time :p
21:55:12 <TrueBrain> but for now: good night!
21:55:17 <TrueBrain> glx: keep on going with the GUI :p
21:55:26 <TrueBrain> Xaroth: I expect a first draft tomorrow morning :p
22:47:47 <Xaroth> php extension libraries :o
22:48:36 <Xaroth> pain in the arse to install something, but still
23:41:28 <Xaroth> ooo, this pecl svn thingie is interesting
continue to next day ⏵