#pentagram@irc.freenode.net logs for 9 Oct 2002 (GMT)

Archive Today Yesterday Tomorrow
Pentagram homepage


[13:08:24] --> Colourless has joined #Pentagram
[13:08:24] --- ChanServ gives channel operator status to Colourless
[17:47:49] --> wjp has joined #pentagram
[17:47:49] --- ChanServ gives channel operator status to wjp
[18:18:36] <wjp> "It's no longer painful to add new scalers" <-- sounds good :-)
[18:18:52] <Colourless> :-)
[18:18:58] <Colourless> look at the top of Scale.cc
[18:19:31] <wjp> hm, bilinear * interlacing? that interlaces but does bilinear interpolation afterwards?
[18:19:55] <Colourless> no, does bilinear interpolation first, then does interlacing
[18:20:22] <wjp> hm, the quarter means that one in every four lines is removed?
[18:20:57] <Colourless> no, what that means is the 'interlaced' line is 75% brightness
[18:21:11] <wjp> ah, I see
[18:21:18] <Colourless> HalfInterlaced is that the 'interlaced' line is 50% brightness
[18:21:31] <wjp> yay, lots of compile errors
[18:21:38] <wjp> </sarcasm> ;-)
[18:21:56] <Colourless> hmm, i was hoping that wouldn't happen :-)
[18:22:10] <wjp> couple of missing newlines at end of file too
[18:22:15] <wjp> viewer/Scale.inl:1721: parse error before `;'
[18:22:38] <wjp> and then there's two pages of warnings about signed/unsigned comparisons
[18:23:32] <Colourless> put a #if 0/#endif around everything in Scale.inl from 1647 to the end of the file
[18:24:22] * wjp blinks
[18:25:27] <wjp> viewer/Surface.cc:118: object-dependent reference to `display16' can
[18:25:27] <wjp> only be used in a call
[18:25:27] <wjp> viewer/Surface.cc:118: to form a pointer to member function, say
[18:25:27] <wjp> `&RenderSurface::display16'
[18:25:33] <wjp> several of those
[18:25:40] <wjp> kind of weird
[18:26:15] <Colourless> y
[18:26:19] <Colourless> ah ok
[18:27:08] <Colourless> would seem to be a difference between how msvc and gcc handle getting pointers to member functions/methods. In MSVC it would appear it's automatic, while in gcc it's not
[18:27:18] <wjp> so it would seem, yes
[18:27:27] <wjp> let me check Stroustrup
[18:28:49] <wjp> it only mentions the & form
[18:30:34] <wjp> ok... getting there
[18:30:41] <Colourless> :-)
[18:31:02] <wjp> __fastcall?
[18:31:16] <Colourless> pass args in registers, not on stack
[18:31:58] <wjp> not in gcc :-)
[18:32:21] <Colourless> it shouldn't have added that unless it was supported
[18:32:40] <wjp> display.cc line 50, btw
[18:33:21] <Colourless> hmmm, how did that get there
[18:33:28] <Colourless> remove that entire func
[18:33:50] <wjp> it's used in line 61
[18:33:55] <Colourless> yeah
[18:34:06] <wjp> (line 61 after removing it :-) )
[18:34:31] <Colourless> i was testing out what exactly __fastcall did (using breakpoints in the compiler)
[18:34:43] <wjp> ok, so remove that cout too I guess?
[18:34:46] <Colourless> yep
[18:35:28] <wjp> ok, it actually builds now :-)
[18:35:39] <Colourless> btw gcc for windows does support __fastcall since it's a standard windows call type
[18:36:29] <wjp> makes sense
[18:36:55] <wjp> it runs too :-)
[18:37:06] <Colourless> :-)
[18:37:07] <Colourless> yay
[18:39:13] <wjp> hm, now I'm still getting pages and pages of signed/unsigned warnings
[18:39:48] <wjp> the "if (buff_size < sline_pixels+1) {" lines in Scale.inl
[18:40:08] <Colourless> just removed all 'unsigned' everywhere :-)
[18:40:31] <wjp> that would work if there was an 'int' following the 'unsigned' :-)
[18:40:46] <Colourless> annoying
[18:40:54] <Colourless> i already fixed up a few of those in other places
[18:41:34] <wjp> ok, much better. Only a couple of unused variable warnings now
[18:44:43] <Colourless> where?
[18:44:54] <wjp> viewer/Scale.inl:1564: warning: unused variable `uint16 *dest2'
[18:45:29] <wjp> hm, the rest seems to be in disasm
[18:46:23] <Colourless> remove that line
[18:46:54] <wjp> viewer/Application.cc:267: warning: unused variable `int counter'
[18:47:44] <Colourless> it's been in there a while, and it could go :-)
[18:49:05] <wjp> viewer/Surface.inl:117: warning: comparison between signed and unsigned integer expressions
[18:49:30] <wjp> viewer/Surface.cc:596: warning: comparison between signed and unsigned ...
[18:49:40] <wjp> viewer/music_proc.cc:321: warning: comparison between signed and ...
[18:51:31] <Colourless> Surface.inl:117
[18:51:43] <Colourless> the problem is because the FrameID type is an unsigned int
[18:51:51] <Colourless> while the Shape::get_count() function returns an int
[18:52:25] <Colourless> Surface.cc:596 same problem
[18:53:35] <Colourless> music_proc.cc:321: hmm, ignore please :-)
[18:54:45] * wjp looks at the if 8 lines higher
[18:54:59] <wjp> something like that? ;-)
[18:55:25] <Colourless> strange that :-)
[18:55:46] <Colourless> i really have to wonder what exactly i was thinking there
[18:55:47] <Colourless> :-)
[18:55:56] <Colourless> make the 'fix' in one place, but not the other
[18:56:20] <Colourless> actually i made the fix in a number of places
[18:56:31] <Colourless> it seems that I must have just forgotten in that one place :-)
[19:02:05] --> Dominus has joined #pentagram
[19:04:24] <wjp> :-)
[19:04:25] <wjp> hi Dominus
[19:13:22] <Colourless> so, are there still warnings?
[19:14:17] <wjp> let me try a clean make
[19:16:24] <wjp> yes, there are
[19:16:44] <wjp> viewer/../audio/xmidi.cc:1101: warning: comparison between signed and...
[19:17:01] <wjp> several in disasm
[19:17:18] <Colourless> well, *you* wrote disasm :-)
[19:17:19] <wjp> some in unpackshp *looks ashamed*
[19:17:30] <wjp> disasm? not recently :-)
[19:17:37] <Colourless> and you wrote that too :-)
[19:17:59] <wjp> tools/../usecode/UCCache.cc:135: warning: unused variable `unsigned int namelength'
[19:18:12] <Colourless> well, that would be one of darke's
[19:18:15] <wjp> I didn't write that one I think?
[19:18:18] <wjp> yeah :-)
[19:18:26] <wjp> tools/dcmpsound/x86er.cc:400: warning: unused variable `uint32 start'
[19:18:47] * wjp blinks
[19:18:51] <wjp> we have a usecode/UCMachine.cc?
[19:19:03] <Dominus> now let's see if I can get it to compile
[19:19:53] <Colourless> yeah we do
[19:20:04] <Colourless> it was a bit of a project that darke and i were working on a bit
[19:20:23] <Colourless> it's 'part' of a usecode interpreter. only does some of the opcodes :-)
[19:20:56] <Colourless> and it doesn't handle some of the 'really' important ones *cough* spawn *cough*
[19:21:10] <wjp> ah... the unused vars are inside a 'if (crusader)' in disasm.cc
[19:21:22] <wjp> *phew* ;-)
[19:21:46] * wjp most certainly did not write that ;-)
[19:22:19] <Colourless> :-)
[19:22:26] <Colourless> what line?
[19:23:37] <wjp> around 162
[19:24:13] <wjp> hm, the others are in the "Listing functions" code
[19:24:18] <wjp> did I write that?
[19:25:31] <Colourless> the 162 would seem to be some of drake's code ( i think)
[19:25:50] <Colourless> Listing functions is mine
[19:31:01] <Dominus> I get: In file included from viewer/Scale.cc:25:
[19:31:01] <Dominus> viewer/Scale.inl: In function `void BlurFilter(Pixel_type*, int, int, int, int,
[19:31:01] <Dominus> int, int, int, Pixel_type*, int)':
[19:31:01] <Dominus> viewer/Scale.inl:1721: parse error before `;' token
[19:31:01] <Dominus> viewer/Scale.cc:382:7: warning: no newline at end of file
[19:31:14] <Dominus> make: *** [Scale.o] Error 1
[19:31:22] <wjp> read the logs ;-)
[19:31:37] <wjp> the warnings are pointed out and fixed in order ;-)
[19:31:43] <Dominus> grumble
[19:31:53] <wjp> s/warnings/errors/
[19:32:31] <Colourless> well you could just commit them wjp :-)
[19:32:43] <Dominus> I guess I'm waiting for an improved cvs...
[19:32:45] <Dominus> :-)
[19:33:54] * wjp takes a hint
[19:37:10] <wjp> hm, nobody tried to compile the new unpackshp.cc yet, I see? :-)
[19:37:31] * wjp just noticed he forgot to commit an updated read_utils.h
[19:38:04] <Colourless> :-)
[19:38:19] <Colourless> well, i guess it might depend on who has a compress shapes.flx file :-)
[19:38:39] <wjp> or on who doesn't have an uncompressed one, more correctly :-)
[19:39:32] * Dominus just likes to point out that U8 and Crusader:No Regret work in Virtual PC 5 :-)
[19:39:49] <Colourless> :-)
[19:39:50] * Dominus likes to point out again...
[19:41:31] <wjp> *sigh*.. one more exercise and I'll be done with my 'formal languages' homework
[19:41:42] <wjp> 11 full pages of text... ack
[19:42:10] <wjp> (um, 11 pages for all of it, not just that 1 exercise, btw :-) )
[19:54:00] <Colourless> :-)
[19:54:22] <Colourless> hmm, my reply took just a bit longer than i thought it would
[19:56:56] <Dominus> Surface.o(.text+0x196c):Surface.cc: undefined reference to `xformblend'
[19:56:57] <Dominus> Surface.o(.text+0x1a24):Surface.cc: undefined reference to `xformblend'
[19:56:57] <Dominus> Surface.o(.text+0x1c43):Surface.cc: undefined reference to `xformblend'
[19:56:57] <Dominus> Surface.o(.text+0x1ed2):Surface.cc: undefined reference to `xformblend'
[19:56:57] <Dominus> Surface.o(.text+0x1f8b):Surface.cc: undefined reference to `xformblend'
[19:56:57] <Dominus> Surface.o(.text+0x21c0):Surface.cc: more undefined references to `xformblend' follow
[19:56:59] <Dominus> make: *** [mapdisp.exe] Error 1
[19:57:29] <Colourless> is xform_blend.cc being compiled?
[19:59:14] <Dominus> I have this warning before:
[19:59:15] <Dominus> In file included from viewer/Surface.cc:21:
[19:59:15] <Dominus> viewer/xform_blend.h:25:1: warning: "FASTCALL" redefined
[19:59:15] <Dominus> In file included from c:/programme/div/curios/mingw/include/g++-v3/mingw32/bits/gthr-default.h:335,
[19:59:15] <Dominus> from c:/programme/div/curios/mingw/include/g++-v3/mingw32/bits/gthr.h:98,
[19:59:15] <Dominus> from c:/programme/div/curios/mingw/include/g++-v3/mingw32/bits/c++io.h:37,
[19:59:17] <Dominus> from c:/programme/div/curios/mingw/include/g++-v3/bits/fpos.h:44,
[19:59:19] <Dominus> from c:/programme/div/curios/mingw/include/g++-v3/bits/char_traits.h:45,
[19:59:21] <Dominus> from c:/programme/div/curios/mingw/include/g++-v3/string:47,
[19:59:23] <Dominus> from headers/common_types.h:26,
[19:59:25] <Dominus> from viewer/Surface.h:22,
[19:59:27] <Dominus> from viewer/Surface.cc:19:
[19:59:29] <Dominus> c:/programme/div/curios/mingw/include/windef.h:95:1: warning: this is the location of the previous definition
[19:59:51] <wjp> curios?
[20:00:06] <Dominus> my file structure is strange
[20:00:35] <Colourless> ok
[20:00:53] <Colourless> just a second
[20:00:58] <Colourless> i'll commit a change
[20:01:40] <Dominus> all the programs I install go to logical sud dirs of /programme/Div, in /Div I have them sorted in online, graphic, video, music and such. If something does not fit those general description t goes into curios :-)
[20:01:53] <Dominus> logicl sub dirs I mean
[20:01:58] <Dominus> bah
[20:02:04] <Dominus> logical sub dirs
[20:02:57] <Dominus> Colourless: thanks
[20:05:51] <Colourless> committed
[20:07:02] <Dominus> trying again
[20:17:31] <Dominus> still get those undefined references to xformblend and it seems that xform_blend.cc doesn't get compiled
[20:18:14] <Dominus> ah, it's missing from makefile.mingw
[20:18:30] <Colourless> yeah :-)
[20:18:37] <Colourless> my mistake
[20:18:44] <Colourless> i'll fix it
[20:18:47] <Dominus> k
[20:22:05] <Colourless> done
[20:23:46] <Dominus> then I get viewer/xform_blend.cc:158: invalid conversion from `uint32 (*)(unsigned int)'
[20:23:46] <Dominus> to `uint32 (*)(unsigned int)'
[20:23:59] <Dominus> make: *** [xform_blend.o] Error 1
[20:24:13] <Colourless> now that is impressing :-
[20:24:19] <Colourless> impressive :-)
[20:24:30] <Dominus> ?
[20:24:41] <Colourless> the error :-)
[20:26:00] <Colourless> try changing these lines:
[20:26:00] <Colourless> xformCol8, // 8 (green->dark grey)
[20:26:00] <Colourless> xformCol9, // 9 (black->vdark grey)
[20:26:00] <Colourless> xformCol10, // 10 (yellow)
[20:26:00] <Colourless> xformCol11, // 11 (white->grey)
[20:26:00] <Colourless> xformCol12, // 12 (red->orange)
[20:26:02] <Colourless> xformCol13, // 13 (blue)
[20:26:04] <Colourless> xformCol14, // 14 (blue)
[20:26:16] <Colourless> to
[20:26:17] <Colourless> &xformCol8, // 8 (green->dark grey)
[20:26:17] <Colourless> &xformCol9, // 9 (black->vdark grey)
[20:26:17] <Colourless> &xformCol10, // 10 (yellow)
[20:26:17] <Colourless> &xformCol11, // 11 (white->grey)
[20:26:17] <Colourless> &xformCol12, // 12 (red->orange)
[20:26:19] <Colourless> &xformCol13, // 13 (blue)
[20:26:21] <Colourless> &xformCol14, // 14 (blue)
[20:27:20] <Dominus> same error
[20:28:06] <Colourless> try adding (xformBlendFuncType) in the front of each line
[20:28:45] <Dominus> ->(xformBlendFuncType)&xformCol8, // 8 (green->dark grey) ?
[20:28:52] <Colourless> yeah
[20:29:34] <Dominus> that worked
[20:29:53] <Dominus> thanks
[20:30:26] <Colourless> actually sounds like a compiler bug. that error it spat doesn't actually make sense, since it's not converting anything :-)
[20:31:30] <Dominus> This is still with gcc 3.1 and it might be fixed in gcc 3.2 but I've been reluctant to convert to gcc 3.2 as we had (still have?) problems with that
[20:31:43] <Colourless> yes we do
[20:31:54] <Colourless> something that 'also' seems like a compiler problem :-)
[20:32:07] <Dominus> :-(
[20:32:46] <Colourless> hmmm....
[20:32:55] <wjp> could the fastcall in the xformBlendFuncType be a problem?
[20:33:09] <Dominus> ok, now it compiled
[20:33:17] <Colourless> shouldn't be
[20:33:23] <Colourless> we'll find out though :-)
[20:33:29] <Colourless> if it crashes, it had a problem :-)
[20:40:19] <-- Dominus has left IRC (Read error: 104 (Connection reset by peer))
[20:41:30] <wjp> hm, crash? :-)
[20:41:40] <Colourless> :-)
[20:41:54] <Colourless> if it crashed his system like that, i'm impressed :-)
[20:42:58] <Colourless> we'll have to wait till he's back
[20:44:22] --> Dominus has joined #pentagram
[20:44:41] <Colourless> so, why did ya disappear
[20:44:54] <Dominus> hi again
[20:45:04] <Dominus> mapdisp bluescreens me on exit in my adio driver
[20:45:08] <wjp> wb :-)
[20:45:25] <Dominus> tried it on reboot right away and it bs me again
[20:45:40] <Colourless> adio?
[20:45:49] <Dominus> audio
[20:45:56] <wjp> I have a feeling he means audio :-)
[20:46:05] <Colourless> just had to be sure :-)
[20:46:44] * Dominus sticks tongue out
[20:46:52] <Colourless> remove the -DHAVE_MIDI from the makefile
[20:49:00] <Dominus> I had the same problems with uwadv that it would crash me on exit
[20:50:29] <Colourless> wjp: i'm thinking it might be a good idea to have a #include "pentagram.h" at the top of each source file. For those of us who support it, that could act as a percompiled header, and it could also be used to do the definitions of things that we will always want (i.e. uint?/sint?) and could also be used to include config.h
[20:51:34] * wjp nods; ok
[20:51:54] * Colourless sends email to list to record said comment
[20:56:38] <Colourless> well, i've really got to be going
[20:56:55] <Dominus> ok
[20:57:07] <-- Colourless has left IRC ("cya")
[20:57:24] <Dominus> going to bother you again about midi music some other time, when you are actually present :-)
[21:25:55] <-- Dominus has left IRC ("enough for now")
[22:45:33] <-- wjp has left IRC ("Zzzz...")