EDuke32 Changelog Archivio non ufficiale

Build 4777

28 novembre 2014 · 81 commit di 3 autori · SVN

Build per mese 276 build

  1. gen

  2. feb

  3. mar

  4. apr

  5. mag

  6. giu

  7. lug

  8. ago

  9. set

  10. ott

  11. nov

  12. dic

File di log originale · 20141128-4777.txt 
------------------------------------------------------------------------
r4777 | hendricks266 | 2014-11-28 22:30:42 -0800 (Fri, 28 Nov 2014) | 1 line

Fix the crash in the ASM relating to the dynamic allocation of ylookup.
------------------------------------------------------------------------
r4776 | hendricks266 | 2014-11-28 19:43:11 -0800 (Fri, 28 Nov 2014) | 1 line

Make sure the mouse is captured when the main menu transitions into MODE_GAME. DONT_BUILD.
------------------------------------------------------------------------
r4775 | hendricks266 | 2014-11-28 19:21:56 -0800 (Fri, 28 Nov 2014) | 1 line

Fix mousewheel under SDL 1.2. DONT_BUILD.
------------------------------------------------------------------------
r4774 | hendricks266 | 2014-11-28 19:07:33 -0800 (Fri, 28 Nov 2014) | 1 line

Restore SDL 1.2 on Windows to a working state. DONT_BUILD.
------------------------------------------------------------------------
r4773 | hendricks266 | 2014-11-28 18:29:42 -0800 (Fri, 28 Nov 2014) | 1 line

Fix POLYMER=0 GTK C++ build. DONT_BUILD.
------------------------------------------------------------------------
r4772 | hendricks266 | 2014-11-28 16:32:28 -0800 (Fri, 28 Nov 2014) | 1 line

Restructure dynamicgtk to comply with C++ casting rules. DONT_BUILD.
------------------------------------------------------------------------
r4771 | hendricks266 | 2014-11-28 15:57:21 -0800 (Fri, 28 Nov 2014) | 1 line

Eliminate -Wno-narrowing. DONT_BUILD.
------------------------------------------------------------------------
r4770 | hendricks266 | 2014-11-28 15:47:46 -0800 (Fri, 28 Nov 2014) | 1 line

GTK: Fix warnings and errors related to the banners. DONT_BUILD.
------------------------------------------------------------------------
r4769 | hendricks266 | 2014-11-28 00:30:31 -0800 (Fri, 28 Nov 2014) | 1 line

Fix C/GTK2/SDL1.2 build. DONT_BUILD.
------------------------------------------------------------------------
r4768 | hendricks266 | 2014-11-28 00:14:00 -0800 (Fri, 28 Nov 2014) | 1 line

Some minor fixes. DONT_BUILD.
------------------------------------------------------------------------
r4767 | hendricks266 | 2014-11-28 00:12:40 -0800 (Fri, 28 Nov 2014) | 1 line

Fix the Wii build with the refactored sdlayer. DONT_BUILD.
------------------------------------------------------------------------
r4766 | hendricks266 | 2014-11-25 20:39:23 -0800 (Tue, 25 Nov 2014) | 1 line

Make a bunch of things extern "C". This should fix C++ builds on OS X and with Lunatic. DONT_BUILD.
------------------------------------------------------------------------
r4765 | hendricks266 | 2014-11-25 20:37:22 -0800 (Tue, 25 Nov 2014) | 1 line

Fix Lunatic C build. DONT_BUILD.
------------------------------------------------------------------------
r4764 | helixhorned | 2014-11-25 15:46:28 -0800 (Tue, 25 Nov 2014) | 1 line

Fix POLYMER=0 build and a couple of unused-variable warnings with it. DONT_BUILD.
------------------------------------------------------------------------
r4763 | hendricks266 | 2014-11-25 13:08:58 -0800 (Tue, 25 Nov 2014) | 1 line

Eliminate the printf_nowarn stuff that has long since been supplanted by TrackerCast. DONT_BUILD.
------------------------------------------------------------------------
r4762 | hendricks266 | 2014-11-24 00:19:50 -0800 (Mon, 24 Nov 2014) | 1 line

Fix some things when building on OS X. DONT_BUILD.
------------------------------------------------------------------------
r4761 | hendricks266 | 2014-11-23 22:30:47 -0800 (Sun, 23 Nov 2014) | 1 line

Fix build with x86_64-MinGW-w64. DONT_BUILD.
------------------------------------------------------------------------
r4760 | hendricks266 | 2014-11-23 22:03:58 -0800 (Sun, 23 Nov 2014) | 1 line

Fix oversight in mouse cursor code. DONT_BUILD.
------------------------------------------------------------------------
r4759 | helixhorned | 2014-11-22 10:37:24 -0800 (Sat, 22 Nov 2014) | 3 lines

In Lunatic build, make krand() a proper function again, called from Lua/FFI.

DONT_BUILD.
------------------------------------------------------------------------
r4758 | helixhorned | 2014-11-22 10:37:23 -0800 (Sat, 22 Nov 2014) | 3 lines

a-c.c: Compile vlineasm4nlogy() only if CLASSIC_NONPOW2_YSIZE_WALLS is #defined.

Also make three file-scope variables static.
------------------------------------------------------------------------
r4757 | helixhorned | 2014-11-22 10:37:21 -0800 (Sat, 22 Nov 2014) | 4 lines

engine.c: in wallscan(), fix oob of umost[] and dmost[] when x2 == xdim-1.

Analogously treat maskwallscan() and transmaskwallscan(), although I could
not get the respective accesses to be oob, too.
------------------------------------------------------------------------
r4756 | helixhorned | 2014-11-22 10:37:19 -0800 (Sat, 22 Nov 2014) | 11 lines

Makefile: do '(very)clean' with a plain "rm -f some/obj/*.o".

Notes:
- In the 'clean' target, the "core*" arg to 'rm' was removed as I find it too
  unspecific. (Someone could have an unrelated file with that prefix lying
  around.)
- The targets 'clean' and 'veryclean' now echo the commands they are about
  to execute.

Also, in Makefile.common, resurrect sanitized Clang builds if DEBUGANYWAY=2
is passed (previously, that was triggered with RELEASE=0 DEBUGANYWAY=1).
------------------------------------------------------------------------
r4755 | helixhorned | 2014-11-22 10:37:16 -0800 (Sat, 22 Nov 2014) | 7 lines

Fix C build and build on Linux/SDL2.

- for the first point: 'struct glattribs sdlayer_gl_attributes[]' was defined
  at file scope, but 'glmultisample > 0' is not a compile-time constant.
- for the second point: r4749 changed the meaning of some conditional
  compilation (supposedly, unintentionally). I'm reverting it to pre-r4749
  but keeping the rewritten style. It's concerned with mouse thumb buttons.
------------------------------------------------------------------------
r4754 | terminx | 2014-11-22 04:44:06 -0800 (Sat, 22 Nov 2014) | 1 line

Fix MinGW build. DONT_BUILD.
------------------------------------------------------------------------
r4753 | terminx | 2014-11-22 04:34:46 -0800 (Sat, 22 Nov 2014) | 1 line

Fixups to sync WIP SDL refactor up with r4738. DONT_BUILD.
------------------------------------------------------------------------
r4752 | terminx | 2014-11-22 04:34:29 -0800 (Sat, 22 Nov 2014) | 1 line

Minor cleanups, mostly just formatting. DONT_BUILD.
------------------------------------------------------------------------
r4751 | terminx | 2014-11-22 04:34:11 -0800 (Sat, 22 Nov 2014) | 1 line

Add missing "break" to M_RunMenuInput(), revert android gltexapplyprops() menu fuckery. DONT_BUILD.
------------------------------------------------------------------------
r4750 | terminx | 2014-11-22 04:33:57 -0800 (Sat, 22 Nov 2014) | 1 line

Fix _FORTIFY_SOURCE redefinition warning on Ubuntu's patched versions of GCC. DONT_BUILD.
------------------------------------------------------------------------
r4749 | terminx | 2014-11-22 04:33:47 -0800 (Sat, 22 Nov 2014) | 1 line

WIP refactor of SDL interface. DONT_BUILD.
------------------------------------------------------------------------
r4748 | terminx | 2014-11-22 04:33:15 -0800 (Sat, 22 Nov 2014) | 1 line

Clean up the sound remapping stuff... remove the bits relating to sounds we haven't actually used anywhere in the source directly. DONT_BUILD.
------------------------------------------------------------------------
r4747 | terminx | 2014-11-22 04:32:56 -0800 (Sat, 22 Nov 2014) | 1 line

Part 1 of attempting to adhere to C standard section 7.1.3. It turns out it's illegal to use identifiers that begin with _ or have __ anywhere in them. DONT_BUILD.
------------------------------------------------------------------------
r4746 | terminx | 2014-11-22 04:29:38 -0800 (Sat, 22 Nov 2014) | 1 line

Get rid of -Wdeclaration-after-statement as we're only targeting C99 and C++ now. DONT_BUILD.
------------------------------------------------------------------------
r4745 | terminx | 2014-11-22 04:29:25 -0800 (Sat, 22 Nov 2014) | 7 lines

Mainly misc cleanups (and a fix for the C++ build), but there are a few important changes in here.

VM_OnEvent() has become VM_OnEvent(), VM_OnEventWithReturn(), VM_OnEventWithDist(), and VM_OnEventWithBoth() (the latter of which is only ever used once...). Of course, this required every call to VM_OnEvent() be changed.

memberlabel_t and vmstate_t have been changed to use the regular "int" type versus explicitly specifying int32_t as they did previously. The rationale for this change is simply that it looks cleaner, and I think we should move toward just using "int" in most cases where there's no particular reason to specify an explicit data type.

Also changes CON_KILLIT to just "return" instead of "continue". DONT_BUILD.
------------------------------------------------------------------------
r4744 | terminx | 2014-11-22 04:28:52 -0800 (Sat, 22 Nov 2014) | 1 line

Inline krand() when not using KRANDDEBUG. DONT_BUILD.
------------------------------------------------------------------------
r4743 | terminx | 2014-11-22 04:28:34 -0800 (Sat, 22 Nov 2014) | 1 line

Add FORCE_INLINE to compat.h and apply it to a couple of things we'd strongly prefer be inlined. DONT_BUILD.
------------------------------------------------------------------------
r4742 | terminx | 2014-11-22 04:28:21 -0800 (Sat, 22 Nov 2014) | 1 line

Add my _clang-format file to the repository. These options are pretty close to what most of our code already looks like, the primary difference being that it enforces spaces between operators and operands. I was initially opposed to the spaces, but I would be lying if I said they didn't increase legibility of much of our code, in addition to making it a lot easier to identify the specific intent in cases where we have a lot of expressions nested within one another. DONT_BUILD.
------------------------------------------------------------------------
r4741 | terminx | 2014-11-22 04:28:12 -0800 (Sat, 22 Nov 2014) | 1 line

Fix up conveyor belts. This fixes breakage of the movement in general, as well as corrects the general behavior of the effect so that a player on a moving surface is moved at a constant rate, regardless of whether the player crouches, is shrunk, etc. DONT_BUILD.
------------------------------------------------------------------------
r4740 | terminx | 2014-11-22 04:27:57 -0800 (Sat, 22 Nov 2014) | 1 line

Fix Mapster32 fullscreen video modes. DONT_BUILD.
------------------------------------------------------------------------
r4739 | hendricks266 | 2014-11-17 00:14:33 -0800 (Mon, 17 Nov 2014) | 1 line

Fix minor menu bugs. DONT_BUILD.
------------------------------------------------------------------------
r4738 | hendricks266 | 2014-11-16 23:39:12 -0800 (Sun, 16 Nov 2014) | 1 line

Set up mouse cursor display in menus, with idle timeout fully implemented. No functionality yet. DONT_BUILD.
------------------------------------------------------------------------
r4737 | helixhorned | 2014-11-07 14:07:14 -0800 (Fri, 07 Nov 2014) | 1 line

Fix C++ build. DONT_BUILD.
------------------------------------------------------------------------
r4736 | helixhorned | 2014-11-07 14:07:13 -0800 (Fri, 07 Nov 2014) | 3 lines

In release builds, #define Xaligned_alloc as xaligned_malloc.

Otherwise, the out-of-memory handler is not called in case of failure.
------------------------------------------------------------------------
r4735 | helixhorned | 2014-11-07 14:07:12 -0800 (Fri, 07 Nov 2014) | 6 lines

C-CON: fix 'readgamearray' and 'resizearray' on Windows.

We must not use Xrealloc(), since gamearrays are now allocated with
Xaligned_alloc().
NOTE: I did not check all potential problematic cases, only those to get
LNGA3 up and running.
------------------------------------------------------------------------
r4734 | helixhorned | 2014-11-07 14:07:12 -0800 (Fri, 07 Nov 2014) | 3 lines

C-CON: fix 'readgamearray' on 64-bit platforms.

Also prettify P_DisplaySpit().
------------------------------------------------------------------------
r4733 | helixhorned | 2014-11-07 14:07:11 -0800 (Fri, 07 Nov 2014) | 1 line

a-c.c: don't USE_VECTOR_EXT with Clang, since it doesn't compile here.
------------------------------------------------------------------------
r4732 | helixhorned | 2014-11-07 14:07:10 -0800 (Fri, 07 Nov 2014) | 1 line

LunaCON: player[].loogcnt, userdef[].color, minor fixes.
------------------------------------------------------------------------
r4731 | helixhorned | 2014-11-07 14:07:07 -0800 (Fri, 07 Nov 2014) | 1 line

LunaCON: fix tilesizx[] and tilesizy[] system gamearrays.
------------------------------------------------------------------------
r4730 | hendricks266 | 2014-11-06 23:53:28 -0800 (Thu, 06 Nov 2014) | 1 line

Replace DOS box-drawing characters in the ASM with sane ASCII. DONT_BUILD.
------------------------------------------------------------------------
r4729 | helixhorned | 2014-11-06 15:58:33 -0800 (Thu, 06 Nov 2014) | 1 line

Lunatic: some minor doc tweaks. DONT_BUILD.
------------------------------------------------------------------------
r4728 | helixhorned | 2014-11-06 15:43:52 -0800 (Thu, 06 Nov 2014) | 5 lines

Fix voxels in classic -- my error, confusing divscale() and mulscale() in r4576.

Optimizations are left to the reader :P. I would prefer to look for ways of
giving the compilers hints about loops to unroll instead of having to read
the resulting mess of manual unrolling, though.
------------------------------------------------------------------------
r4727 | helixhorned | 2014-11-06 15:43:51 -0800 (Thu, 06 Nov 2014) | 7 lines

gameexec.c: In VM_OnEvent_, don't form address like &sprite[-1].

The practical rationale: Clang-sanitize catches this, so this is bad.
The real rationale: I *think* it is undefined behavior to even form such a
pointer in C99. However, I would be hard pressed to provide a nice formal
argument in terms of the Standard wording right now. It looks like
6.5.3.2#4 is to blame.
------------------------------------------------------------------------
r4726 | helixhorned | 2014-11-06 15:43:50 -0800 (Thu, 06 Nov 2014) | 1 line

engine.c: prevent an oob read in drawsprite_classic() w/ floor-aligned sprites.
------------------------------------------------------------------------
r4725 | helixhorned | 2014-11-06 15:43:49 -0800 (Thu, 06 Nov 2014) | 1 line

Fix oob write with stepping 'up' in the OSD history.
------------------------------------------------------------------------
r4724 | helixhorned | 2014-11-06 15:43:47 -0800 (Thu, 06 Nov 2014) | 5 lines

baselayer.c: fix an oob write to ylookup[] intoduced in r4695.

On Linux, it started manifesting itself since r4719 as crashes whose
backtraces pointed to malloc() and were not helpful. Valgrind finally
uncovered this.
------------------------------------------------------------------------
r4723 | terminx | 2014-11-01 22:44:27 -0700 (Sat, 01 Nov 2014) | 1 line

Useless change I forgot. DONT_BUILD.
------------------------------------------------------------------------
r4722 | terminx | 2014-11-01 22:38:57 -0700 (Sat, 01 Nov 2014) | 1 line

Sometimes I wonder why MSVC neglects to provide useful warnings. DONT_BUILD.
------------------------------------------------------------------------
r4721 | terminx | 2014-11-01 22:36:53 -0700 (Sat, 01 Nov 2014) | 1 line

Minor tweaks to VM_OnEvent_(). DONT_BUILD.
------------------------------------------------------------------------
r4720 | terminx | 2014-11-01 22:36:41 -0700 (Sat, 01 Nov 2014) | 1 line

Remove references to nonexistent enet_mmulti.h from Visual Studio project. DONT_BUILD.
------------------------------------------------------------------------
r4719 | terminx | 2014-11-01 22:36:28 -0700 (Sat, 01 Nov 2014) | 1 line

Fix a handful of crash bugs I caused with the 32-bit ASM version of the classic renderer. DONT_BUILD.
------------------------------------------------------------------------
r4718 | terminx | 2014-11-01 22:36:05 -0700 (Sat, 01 Nov 2014) | 1 line

Fix a crash I caused with MSVC in OSD_SetLogFile(). Apparently, the behavior of setvbuf() when passed a NULL ptr for a buffer is actually left up to the implementation regarding what it does with the size parameter... DONT_BUILD.
------------------------------------------------------------------------
r4717 | terminx | 2014-11-01 22:35:50 -0700 (Sat, 01 Nov 2014) | 1 line

Clean up classic renderer header (a.h) a bit. DONT_BUILD.
------------------------------------------------------------------------
r4716 | terminx | 2014-11-01 22:35:36 -0700 (Sat, 01 Nov 2014) | 1 line

A couple of additional compat.h formatting changes. DONT_BUILD.
------------------------------------------------------------------------
r4715 | terminx | 2014-11-01 22:35:22 -0700 (Sat, 01 Nov 2014) | 1 line

clang-format a couple of pragma headers. DONT_BUILD.
------------------------------------------------------------------------
r4714 | terminx | 2014-11-01 22:35:06 -0700 (Sat, 01 Nov 2014) | 1 line

LUT for slope recip calculations in classic. DONT_BUILD.
------------------------------------------------------------------------
r4713 | terminx | 2014-11-01 22:34:49 -0700 (Sat, 01 Nov 2014) | 1 line

Fix a couple of Makefile errors Helixhorned noticed. DONT_BUILD.
------------------------------------------------------------------------
r4712 | helixhorned | 2014-11-01 11:01:28 -0700 (Sat, 01 Nov 2014) | 3 lines

menus.c: prevent using a null pointer as arg to "%s" format for sprintf().

DONT_BUILD.
------------------------------------------------------------------------
r4711 | helixhorned | 2014-10-31 14:09:26 -0700 (Fri, 31 Oct 2014) | 1 line

Add forgotten events_defs.h, oops. DONT_BUILD.
------------------------------------------------------------------------
r4710 | helixhorned | 2014-10-30 14:44:22 -0700 (Thu, 30 Oct 2014) | 6 lines

Lunatic: fix building and startup up to tilesizx[] failure. DONT_BUILD.

- Needed to rip out enum GameEvent_t into own header file to prevent what
  would be a circular #include otherwise
- Added some forgotten #ifdef LUNATIC
- updated structure and function declarations on the Lua FFI side
------------------------------------------------------------------------
r4709 | helixhorned | 2014-10-30 14:44:21 -0700 (Thu, 30 Oct 2014) | 7 lines

C-CON: fix possible crash when a CON error is raised due to access of invalid sprite.

For example, in
    CON_ERRPRINTF("invalid target sprite (%d) %d %d\n", iActor, vm.g_i, TrackerCast(vm.g_sp->picnum));
vm.g_i can be -1 and vm.g_sp can be NULL then. (Not anymore.)

DONT_BUILD.
------------------------------------------------------------------------
r4708 | helixhorned | 2014-10-29 21:59:23 -0700 (Wed, 29 Oct 2014) | 1 line

compat.h: pedantic whitespace formatting. DONT_BUILD.
------------------------------------------------------------------------
r4707 | terminx | 2014-10-29 10:08:03 -0700 (Wed, 29 Oct 2014) | 1 line

Alter weapon switching to cycle through GROW_WEAPON when using next/previous weapon keys. DONT_BUILD.
------------------------------------------------------------------------
r4706 | terminx | 2014-10-29 10:07:47 -0700 (Wed, 29 Oct 2014) | 1 line

Default map zoom to something a bit more useful on Android. DONT_BUILD.
------------------------------------------------------------------------
r4705 | terminx | 2014-10-29 10:07:35 -0700 (Wed, 29 Oct 2014) | 1 line

CPLUSPLUS mode build fix. DONT_BUILD.
------------------------------------------------------------------------
r4704 | terminx | 2014-10-29 10:07:21 -0700 (Wed, 29 Oct 2014) | 1 line

CPLUSPLUS mode by default in Makefile.common, C99 otherwise. DONT_BUILD.
------------------------------------------------------------------------
r4703 | terminx | 2014-10-29 10:07:11 -0700 (Wed, 29 Oct 2014) | 1 line

Clean up some player code a bit... make bobposx/bobposy a vec2_t, make fricxv/fricyv a per-player vec2_t (TODO: CON access), promote angvel in input_t to int16_t and allow for player angle changes that result in odd numbered angles (we were effectively artificially limiting the angle to 1024 values before), fix some HUD model ID stuff that should help with the weapons in the HRP, clean up a bunch of random functions (P_FireWeapon(), P_DisplayTip(), P_DisplayAccess(), P_DisplayWeapon(), P_GetInput(), etc). Also clean up G_SetupFilenameBasedMusic() to loop through flac/ogg/mid when searching for usermap music replacements. Some of this really needs a BYTEVERSION bump, but these commits aren't for synthesis to build so we're not doing it yet. DONT_BUILD.
------------------------------------------------------------------------
r4702 | terminx | 2014-10-29 10:06:43 -0700 (Wed, 29 Oct 2014) | 1 line

Try out __restrict in the classic renderer. #define'd for easy removal in case it causes problems or is useless.
------------------------------------------------------------------------
r4701 | terminx | 2014-10-29 10:06:20 -0700 (Wed, 29 Oct 2014) | 1 line

Fix MSVC warning in Polymer
------------------------------------------------------------------------
r4700 | terminx | 2014-10-29 10:06:05 -0700 (Wed, 29 Oct 2014) | 1 line

Inline yax_getbunches with old map format. DONT_BUILD.
------------------------------------------------------------------------
r4699 | terminx | 2014-10-29 10:05:46 -0700 (Wed, 29 Oct 2014) | 1 line

More EDUKE32_PREDICT_FALSE, mostly in OSD_Printf(OSD_ERROR...) cases
------------------------------------------------------------------------
r4698 | terminx | 2014-10-29 10:05:29 -0700 (Wed, 29 Oct 2014) | 1 line

Aligned gamevar/gamearray allocation
------------------------------------------------------------------------
r4697 | terminx | 2014-10-29 10:05:15 -0700 (Wed, 29 Oct 2014) | 1 line

Fix a couple of clang warnings. DONT_BUILD.
------------------------------------------------------------------------
See http://svn.eduke32.com/listing.php?repname=eduke32 for more details.
Fix

Make a bunch of things extern "C". This should fix C++ builds on OS X and with Lunatic. DONT_BUILD.

Evan Ramos (Hendricks266) 25 nov 2014, 20:39 UTC−8 r4766

Eliminate the printf_nowarn stuff that has long since been supplanted by TrackerCast. DONT_BUILD.

Evan Ramos (Hendricks266) 25 nov 2014, 13:08 UTC−8 r4763

a-c.c: Compile vlineasm4nlogy() only if CLASSIC_NONPOW2_YSIZE_WALLS is #defined.

Philipp Kutin (Helixhorned) 22 nov 2014, 10:37 UTC−8 r4758

Also make three file-scope variables static.

FixEngine

engine.c: in wallscan(), fix oob of umost[] and dmost[] when x2 == xdim-1.

Philipp Kutin (Helixhorned) 22 nov 2014, 10:37 UTC−8 r4757

Analogously treat maskwallscan() and transmaskwallscan(), although I could not get the respective accesses to be oob, too.

Build system

do '(very)clean' with a plain "rm -f some/obj/*.o".

Philipp Kutin (Helixhorned) 22 nov 2014, 10:37 UTC−8 r4756

Notes: - In the 'clean' target, the "core*" arg to 'rm' was removed as I find it too unspecific. (Someone could have an unrelated file with that prefix lying around.) - The targets 'clean' and 'veryclean' now echo the commands they are about to execute. Also, in Makefile.common, resurrect sanitized Clang builds if DEBUGANYWAY=2 is passed (previously, that was triggered with RELEASE=0 DEBUGANYWAY=1).

Fix

Fix C build and build on Linux/SDL2.

Philipp Kutin (Helixhorned) 22 nov 2014, 10:37 UTC−8 r4755

- for the first point: 'struct glattribs sdlayer_gl_attributes[]' was defined at file scope, but 'glmultisample > 0' is not a compile-time constant. - for the second point: r4749 changed the meaning of some conditional compilation (supposedly, unintentionally). I'm reverting it to pre-r4749 but keeping the rewritten style. It's concerned with mouse thumb buttons.

Fix

Add missing "break" to M_RunMenuInput(), revert android gltexapplyprops() menu fuckery. DONT_BUILD.

Richard Gobeille (TerminX) 22 nov 2014, 04:34 UTC−8 r4751

Clean up the sound remapping stuff... remove the bits relating to sounds we haven't actually used anywhere in the source directly. DONT_BUILD.

Richard Gobeille (TerminX) 22 nov 2014, 04:33 UTC−8 r4748

Part 1 of attempting to adhere to C standard section 7.1.3. It turns out it's illegal to use identifiers that begin with _ or have __ anywhere in them. DONT_BUILD.

Richard Gobeille (TerminX) 22 nov 2014, 04:32 UTC−8 r4747
Fix

Mainly misc cleanups (and a fix for the C++ build), but there are a few important changes in here.

Richard Gobeille (TerminX) 22 nov 2014, 04:29 UTC−8 r4745

VM_OnEvent() has become VM_OnEvent(), VM_OnEventWithReturn(), VM_OnEventWithDist(), and VM_OnEventWithBoth() (the latter of which is only ever used once...). Of course, this required every call to VM_OnEvent() be changed. memberlabel_t and vmstate_t have been changed to use the regular "int" type versus explicitly specifying int32_t as they did previously. The rationale for this change is simply that it looks cleaner, and I think we should move toward just using "int" in most cases where there's no particular reason to specify an explicit data type. Also changes CON_KILLIT to just "return" instead of "continue". DONT_BUILD.

Add FORCE_INLINE to compat.h and apply it to a couple of things we'd strongly prefer be inlined. DONT_BUILD.

Richard Gobeille (TerminX) 22 nov 2014, 04:28 UTC−8 r4743

Add my _clang-format file to the repository. These options are pretty close to what most of our code already looks like, the primary difference being that it enforces spaces between operators and operands. I was initially opposed to the spaces, but I would be lying if I said they didn't increase legibility of much of our code, in addition to making it a lot easier to identify the specific intent in cases where we have a lot of expressions nested within one another. DONT_BUILD.

Richard Gobeille (TerminX) 22 nov 2014, 04:28 UTC−8 r4742
Fix

Fix up conveyor belts. This fixes breakage of the movement in general, as well as corrects the general behavior of the effect so that a player on a moving surface is moved at a constant rate, regardless of whether the player crouches, is shrunk, etc. DONT_BUILD.

Richard Gobeille (TerminX) 22 nov 2014, 04:28 UTC−8 r4741

Set up mouse cursor display in menus, with idle timeout fully implemented. No functionality yet. DONT_BUILD.

Evan Ramos (Hendricks266) 16 nov 2014, 23:39 UTC−8 r4738

In release builds, #define Xaligned_alloc as xaligned_malloc.

Philipp Kutin (Helixhorned) 7 nov 2014, 14:07 UTC−8 r4736

Otherwise, the out-of-memory handler is not called in case of failure.

C-CON FixCON

fix 'readgamearray' and 'resizearray' on Windows.

Philipp Kutin (Helixhorned) 7 nov 2014, 14:07 UTC−8 r4735

We must not use Xrealloc(), since gamearrays are now allocated with Xaligned_alloc(). NOTE: I did not check all potential problematic cases, only those to get LNGA3 up and running.

FixEngine

Fix voxels in classic -- my error, confusing divscale() and mulscale() in r4576.

Philipp Kutin (Helixhorned) 6 nov 2014, 15:43 UTC−8 r4728

Optimizations are left to the reader :P. I would prefer to look for ways of giving the compilers hints about loops to unroll instead of having to read the resulting mess of manual unrolling, though.

Engine

gameexec.c: In VM_OnEvent_, don't form address like &sprite[-1].

Philipp Kutin (Helixhorned) 6 nov 2014, 15:43 UTC−8 r4727

The practical rationale: Clang-sanitize catches this, so this is bad. The real rationale: I *think* it is undefined behavior to even form such a pointer in C99. However, I would be hard pressed to provide a nice formal argument in terms of the Standard wording right now. It looks like 6.5.3.2#4 is to blame.

Fix

baselayer.c: fix an oob write to ylookup[] intoduced in r4695.

Philipp Kutin (Helixhorned) 6 nov 2014, 15:43 UTC−8 r4724

On Linux, it started manifesting itself since r4719 as crashes whose backtraces pointed to malloc() and were not helpful. Valgrind finally uncovered this.

FixEngine

Fix a handful of crash bugs I caused with the 32-bit ASM version of the classic renderer. DONT_BUILD.

Richard Gobeille (TerminX) 1 nov 2014, 22:36 UTC−7 r4719
Fix

Fix a crash I caused with MSVC in OSD_SetLogFile(). Apparently, the behavior of setvbuf() when passed a NULL ptr for a buffer is actually left up to the implementation regarding what it does with the size parameter... DONT_BUILD.

Richard Gobeille (TerminX) 1 nov 2014, 22:36 UTC−7 r4718
Lunatic Fix

fix building and startup up to tilesizx[] failure. DONT_BUILD.

Philipp Kutin (Helixhorned) 30 ott 2014, 14:44 UTC−7 r4710

- Needed to rip out enum GameEvent_t into own header file to prevent what would be a circular #include otherwise - Added some forgotten #ifdef LUNATIC - updated structure and function declarations on the Lua FFI side

C-CON FixEngineCON

fix possible crash when a CON error is raised due to access of invalid sprite.

Philipp Kutin (Helixhorned) 30 ott 2014, 14:44 UTC−7 r4709

For example, in CON_ERRPRINTF("invalid target sprite (%d) %d %d\n", iActor, vm.g_i, TrackerCast(vm.g_sp->picnum)); vm.g_i can be -1 and vm.g_sp can be NULL then. (Not anymore.) DONT_BUILD.

Alter weapon switching to cycle through GROW_WEAPON when using next/previous weapon keys. DONT_BUILD.

Richard Gobeille (TerminX) 29 ott 2014, 10:08 UTC−7 r4707
FixCON

Clean up some player code a bit... make bobposx/bobposy a vec2_t, make fricxv/fricyv a per-player vec2_t (TODO: CON access), promote angvel in input_t to int16_t and allow for player angle changes that result in odd numbered angles (we were effectively artificially limiting the angle to 1024 values before), fix some HUD model ID stuff that should help with the weapons in the HRP, clean up a bunch of random functions (P_FireWeapon(), P_DisplayTip(), P_DisplayAccess(), P_DisplayWeapon(), P_GetInput(), etc). Also clean up G_SetupFilenameBasedMusic() to loop through flac/ogg/mid when searching for usermap music replacements. Some of this really needs a BYTEVERSION bump, but these commits aren't for synthesis to build so we're not doing it yet. DONT_BUILD.

Richard Gobeille (TerminX) 29 ott 2014, 10:07 UTC−7 r4703
Engine

Try out __restrict in the classic renderer. #define'd for easy removal in case it causes problems or is useless.

Richard Gobeille (TerminX) 29 ott 2014, 10:06 UTC−7 r4702