* Turn some signed integers ("int" or "SDWORD") into unsigned integers where it made more sence (and signed integers caused warnings about comparing signed with unsigned integers)
* At some places cast to an unsigned integer (where the type of the expression is a signed integer but the result of it logically _always_ is an absolute value)
* Remove unused struct-type EVENT_INIT
* Convert use of malloc->memcpy->free to one single realloc call (and now make sure to actually check realloc's return value, which wasn't done with the malloc call)
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@3037 4a71c877-e1ca-e34f-864e-861f7616d084
- Less strict warnings in debug-mode (still unusable, because ICC warns about _a lot_)
- Explicit conversions
- Clarify parameter evaluation order
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@2550 4a71c877-e1ca-e34f-864e-861f7616d084
Any fixes, if necessary, should be applied on loading.
In the image code (IMAGEDEF) this did not work, so OLD_TEXTURE_SIZE_FIX is applied during rendering.
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@2258 4a71c877-e1ca-e34f-864e-861f7616d084
- POINT -> Vector2i
- changed internalformat = getPixelFormat() to GL_RGBA in one place
- removed unused variables/functions
- proper intialisation
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1792 4a71c877-e1ca-e34f-864e-861f7616d084
- Renamed pie_PNGLoadFile, because I got confused about pie_ vs. iV_ when working on my terrain stuff, and did not fix it when backporting to trunk
- iTexture (now aka iV_Image, s.a.) is given a depth field to enable support for different number of channels. This is required to support RGB only textures (without adding a useless alpha channel) and eg. for greyscale images like heightmaps
- Simplyfied iV_loadImage_PNG (formerly known as pie_PNGLoadFile, s.a.) and also removed lots of seemingly unnecessary conversions
- Converted page-8 from paletted to RGB for simplification. This multiplied the size by about 3x (still < 100kB)
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1580 4a71c877-e1ca-e34f-864e-861f7616d084
* Turn some implicit conversions of void* to something* into an explicit cast
* remove some redundant commented out code
* refactor some error checking code ( was: if (!error) normal exec path; else deal with error; now is: if (error) deal with error; normal exec path )
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1534 4a71c877-e1ca-e34f-864e-861f7616d084
1. Fixed a bug (I made some parenthesis order mistakes, near predictX and predictY, which might corrupt predictX, predictY)
2. Fixed a value corruption when predictX or predictY becomes negative, since proj_SendProjectile only expects UDWORD (now set them to 0 when they become negative, so the turret wont try to fire projectiles to some off-map coords)
3. Fixed zero height building bug (invincible rearm pads) when an imd's ymin is equal to ymax...
lib/ivis_opengl/screen.c: \param channels
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1524 4a71c877-e1ca-e34f-864e-861f7616d084
allocated on the stack instead of the heap, which usually is a lot faster => stack alloc = O(1), heap (malloc=heap allocater) usually O(n) or O(log n)
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1522 4a71c877-e1ca-e34f-864e-861f7616d084
* (r1502 had a NULL pointer to many passed in the argumentlist)
thanks to vs2k5 for this fix (sorry couldn't find any other name anywhere on the mailinglist than your email userid)
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1503 4a71c877-e1ca-e34f-864e-861f7616d084
* Remove some dead (commented out) & dirty code from multiint.c which used JPEG code (image_load_from_jpg)
* Remove dead (commented out) code from screen.c which used JPEG code (function static GLuint image_create_texture)
* Remove unused backdrop types from pieblitfunc.[ch]
* Convert last JPEG files into PNG, add a "comment" property on those: "All of these are converted from JPEG files, so this is JPEG quality."
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1490 4a71c877-e1ca-e34f-864e-861f7616d084
* change pie_ReloadTexPage to use files instead of memory buffers as well
* change data.c's resource handler mappings to now load some more fileformats directly from file (all that directly or indirectly used to depend upon pie_PNGLoadMem)
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1483 4a71c877-e1ca-e34f-864e-861f7616d084
* do the same with pie_ReloadTexPage which calls pie_PNGLoadMem
* char* -> const char* in tex.c
* fix a syntax error from r1481
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1482 4a71c877-e1ca-e34f-864e-861f7616d084
* dump less information in the debug log if the assert is triggered (just "integer overflow; no more filenumbers available.")
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1481 4a71c877-e1ca-e34f-864e-861f7616d084
* make screen dumping code (screen shotstuff) more readable and friendly to edit (behaviour should be unchanged)
* change screenDumpToDisk:
* now returns void (rather than the target file name of the screen shot) since its only callee (a keymap callback) ignored the return value
* use snprintf now instead of sprintf (to prevent buffer overflows)
* add code to detect integer overflows (maybe we should just abort() on this condition?)
* remove an empty and commented out function (screenTextOut)
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1480 4a71c877-e1ca-e34f-864e-861f7616d084
This patch will create crashes where the code expects FREE to set the pointer to NULL! (Those problems should be fixed anyway.)
The only occassion where I saw this is when quiting a game, so it is not too much of an issue.
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1459 4a71c877-e1ca-e34f-864e-861f7616d084
- Remove "-%token <sval> TEXT" from audp_parser.y since it clashed with declarations in MinGW's Windows headers. It does still compile and run, but this may have harmed something anyway. Please check!
- Hack around DATADIR being defined by objbase.h in MinGW by undefining it before the inclusion of shlobj.h (incl. in main.c) This only works since the case that DATADIR is actually used on Windows is extremely rare.
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1437 4a71c877-e1ca-e34f-864e-861f7616d084
only compress the texure pages, and not backdrops and other textures. This this should give a framerate increase of 5-10 fps on fillrate limited
systems (cpu usage < 100%).
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1289 4a71c877-e1ca-e34f-864e-861f7616d084
Fixing type-punned pointers is difficult, especially with that HEAP_ALLOC function...
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1252 4a71c877-e1ca-e34f-864e-861f7616d084
- Vector3i from lib/ivis also used in lib/audio, maybe want to move it to some more generic place?
- Replace iPoint with Vector2i
- Remove additional integer type declarations from pietypes.h
- Rename all iSprite to iTexture
- Various warning fixes: Unsigned/signed comparisons, uninitialized variables, declared but not defined symbols
- Replace DEBUG_BLOCK with DEBUG_MALLOC (it was virtually the same)
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1245 4a71c877-e1ca-e34f-864e-861f7616d084
checks for lack of double buffering and too small
texture support. Use --debug texture to see your
texture size (first line in output).
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1156 4a71c877-e1ca-e34f-864e-861f7616d084
yes (as before), relaxed (-O2 & -Wextra), and no
(as before). Fix two more reported warnings.
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1151 4a71c877-e1ca-e34f-864e-861f7616d084
- Add 2 new backdrops from Terminator (http://wz2100.net/forum/index.php?topic=109.0)
- Remove unused pal parameter from pie_PNGLoadMem()
git-svn-id: svn+ssh://svn.gna.org/svn/warzone/trunk@1075 4a71c877-e1ca-e34f-864e-861f7616d084
Dennis Schridde
Per Inge Mathisen
Giel van Schijndel
Freddie Witherden
Gerard Krol