Commit graph

339 commits

Author SHA1 Message Date
Richard Goedeken 347b4e48c9
Merge pull request #130 from GhostlyDark/remove_boost
Remove Boost dependency
2024-05-09 19:41:59 -07:00
GhostlyDark bd747ff3e6 Remove Boost 2024-05-08 14:51:28 +02:00
GhostlyDark b44b0d1c43 Hotfix building with Boost v1.85.0 2024-05-05 18:12:37 +02:00
Richard Goedeken d900f21915
Merge pull request #125 from Jj0YzL5nvJ/up2023
CI/CD: Fixes Ubuntu i386 builds, among other changes
2023-06-11 12:20:38 -07:00
Jj0YzL5nvJ 5fb3193cff CI/CD: Fixes Ubuntu i386 builds, among other changes 2023-05-22 00:09:36 -06:00
Richard Goedeken 497a8255ac
Merge pull request #124 from Jj0YzL5nvJ/ci-schedule
CI/CD: Integrate scheduled builds and other...
2023-03-02 21:59:42 -08:00
Jj0YzL5nvJ 45a25a93ac CI/CD: Integrate scheduled builds and other necessary updates 2023-02-14 23:49:09 -06:00
Richard Goedeken a07050d143
Merge pull request #123 from ecsv/loadblock_signed_overflow
rdp: Avoid signed overflow during texture block loading
2022-11-16 20:55:21 -08:00
Sven Eckelmann 12bf43103e rdp: Avoid signed overflow during texture block loading
It isn't well defined how the compiler should behave when a signed integer
overlows. This can lead to artifacts like scrambled textures when such an
overflow happens.

For the input v16 == 1879048192 (0x70000000) and dxt == 268435456
(0x10000000) , it would calculate the result 536870912 (0x20000000) instead
of -1 (2147483648 aka 0x80000000). Which means that the flipping will not
be performed correctly.

Fixes: mupen64plus/mupen64plus-video-glide64mk2#120
Signed-off-by: Sven Eckelmann <sven@narfation.org>
2022-11-15 22:00:41 +01:00
Richard Goedeken 95229fed6d
Merge pull request #122 from Jj0YzL5nvJ/up2022
CI/CD: Update MSVC
2022-09-30 16:59:46 -07:00
Jj0YzL5nvJ 340ab2c402 CI/CD: Update MSVC 2022-09-28 01:55:16 -06:00
Richard Goedeken 79809753e8
Merge pull request #121 from Jj0YzL5nvJ/ci-update
CI/CD: Update
2022-07-13 19:29:44 -07:00
Jj0YzL5nvJ bab96abd07 CI/CD: Update 2022-07-10 18:51:46 -06:00
Richard Goedeken 3266c61cf2
Merge pull request #118 from Jj0YzL5nvJ/ci-github
CI/CD: Implement GitHub Actions and public nightly builds
2022-04-11 20:52:35 -07:00
Jj0YzL5nvJ adb41a9e51 CI/CD: Implement GitHub Actions and public nightly builds 2021-12-05 08:59:54 -07:00
Pierce Andjelkovic 5ac60b38e2 RISC-V Support 2021-09-15 00:02:27 +02:00
orbea f0c92d93a2 build: Allow out of tree builds.
Also fixes a case of using $(SRCDIR) before its set.
2020-12-16 10:27:07 +01:00
Richard Goedeken aecf15ec20 Revert "Allow non-default compilers without resorting to symbolic links"
This reverts commit 2f9ccb13b3, as it conflicts with GNU make which always defines default values for certain variables
2020-09-16 22:38:29 -07:00
Richard Goedeken 629f2e99a3
Merge pull request #111 from Jj0YzL5nvJ/alter_cc_cxx
Allow non-default compilers without resorting to symbolic links
2020-08-30 18:54:06 -07:00
Jj0YzL5nvJ 2f9ccb13b3 Allow non-default compilers without resorting to symbolic links 2020-08-30 03:42:18 -06:00
Richard Goedeken 0274f4e64b
Merge pull request #110 from Jj0YzL5nvJ/alter
Allow to set custom platform toolset from commands
2020-08-19 23:36:45 -07:00
Jj0YzL5nvJ c277ac38e4 Allow to set custom platform toolset from commands 2020-08-13 11:54:35 -06:00
Jj0YzL5nvJ b4d4503dd0 AppVeyor with artifact packaging 2020-03-02 09:24:22 +01:00
Jj0YzL5nvJ 06b38186a8 Changes to have more generic project files 2020-03-02 09:24:22 +01:00
Gillou68310 3f7640881d Fix https://github.com/mupen64plus/mupen64plus-user-issues/issues/719 2019-11-15 09:09:33 +01:00
Gillou68310 8601df82e0 Migrate to VS2017 2019-11-13 17:17:08 +01:00
Richard Goedeken c31ce5ca6e fix warnings 2019-06-02 22:10:30 -07:00
Richard Goedeken a91b7c9bed
Merge pull request #106 from fzurita/fix_clang_warnigns
Fix clang errors
2019-03-17 22:41:42 -07:00
fzurita a61f12f78c Fix clang errors 2019-03-18 00:43:04 -04:00
Richard Goedeken 1931ececfc fix warnings 2019-03-03 15:22:44 -08:00
Richard Goedeken 93e506bd18 fix millions of bugs in string format functions with variadic arguments 2019-03-03 10:50:20 -08:00
Richard Goedeken 5cbf483d20 add function attributes so GCC can find bugs in calls to string formatting functions with variadic arguments 2019-03-03 10:05:46 -08:00
Richard Goedeken 7c74a052c5
Merge pull request #105 from djames1/master
Add ppc64le support
2019-02-21 21:44:12 -08:00
Daniel James 409eef76c9 Add ppc64le support 2019-02-20 12:03:39 -05:00
Richard Goedeken 55f330e6fb update version and release info for v2.5.9 BETA release 2019-02-10 10:15:45 -08:00
Richard Goedeken 816530b324 fix Travis CI test builds for MXE 32 and 64 bits 2019-01-29 22:19:44 -08:00
Richard Goedeken 1408ef1923 attempt to make Glide64mk2 work on big endian machines by correcting 244 expressions in which byte/short arrays would be indexed improperly on a big-endian machine 2018-11-20 20:45:05 -08:00
Richard Goedeken 1829868a68 Win32 build fixes 2018-10-12 21:16:29 -07:00
Richard Goedeken eb6484ee30
Merge pull request #102 from bkeys/makefile_aarch64
Added support for aarch64
2018-09-21 20:22:48 -07:00
Brigham Henry Keys f927d3c636 Removed commented out definitions 2018-09-21 02:38:32 -05:00
Brigham Henry Keys bfb5dc07ac Added support for aarch64 2018-09-19 12:12:32 -05:00
Richard Goedeken 21a911bd5e
Merge pull request #101 from bugfood/off-by-one
fix presumed off-by-one bug in TxQuantize::quantize
2018-08-27 21:44:50 -07:00
Corey Hickey 290d564ed6 fix presumed off-by-one bug in TxQuantize::quantize
The surrounding code is structured as if it expects the loop to run as
'< numcore'. Changing the loop conditional makes it consistent with
all the other 'numcore' loops in the file, and is presumably the correct
fix. Unfortunately, I do not know how to exercise this piece of code, so
the patch is given as-is.

Discussed in:
https://github.com/mupen64plus/mupen64plus-video-glide64mk2/issues/100
2018-08-27 20:07:27 -07:00
Richard Goedeken 3e62b86563
Merge pull request #99 from bugfood/rm-MAX_NUMCORE
Remove MAX_NUMCORE.
2018-08-05 12:59:33 -07:00
Corey Hickey ef7d39f464 Remove MAX_NUMCORE.
This fixes a bug I hit in TxQuantize.cpp: the params array is
initialized with length MAX_NUMCORE, but accessed at positions up to
numcore-1. The numcore variable is not constrained to MAX_NUMCORE, so on
systems with cores greater than MAX_NUMCORE, there can be a segfault.

Is there any downside? I don't see one; I'm not familiar with the code,
but it looks like this may cause a negligible decrease in memory usage
on systems with less than MAX_NUMCORE (8) cores, and negligibly higher
memory usage on systems with a higher number of cores as the price of
not segfaulting.

I do not know if I have exercised all code paths in testing this change,
but the usage of MAX_NUMCORE seems the same in all cases.
2018-08-04 23:21:26 -07:00
Richard Goedeken 6ae21010cb
Merge pull request #98 from bsmiles32/warnings
Warnings
2018-05-29 16:04:38 -07:00
Bobby Smiles c6f47d23f6 Fix _dbgfile usage before initialization. 2018-05-30 00:12:19 +02:00
Bobby Smiles 06ed7d55a5 Fix gcc warning about using src size instead of dst size as a limiting size 2018-05-30 00:05:54 +02:00
Bobby Smiles 79afad72c4 Fix GCC warning about unneeded parenthesis. 2018-05-29 23:59:51 +02:00
Bobby Smiles cf0977189c Fix ambiguous indentation warning triggered by GCC. 2018-05-29 23:59:42 +02:00