- --
Viewing Issue Advanced Details
ID | Category [?] | Severity [?] | Reproducibility | Date Submitted | Last Update |
---|---|---|---|---|---|
05640 | Compiling | Major | Always | Jul 25, 2014, 12:07 | Apr 10, 2018, 08:55 |
Tester | gauel | View Status | Public | Platform | Other |
Assigned To | Resolution | No change required | OS | Linux | |
Status [?] | Closed | Driver | |||
Version | 0.154 | Fixed in Version | 0.197 | Build | 64-bit |
Fixed in Git Commit | Github Pull Request # | ||||
Summary | 05640: Build failed: array subscript is above array bounds | ||||
Description |
TARGET = ume gcc-4.9.1 $ make -j16 ======================= src/mame/video/model1.c: In function 'void draw_quads(model1_state*, bitmap_rgb32&, const rectangle&)': src/mame/video/model1.c:307:17: error: array subscript is above array bounds [-Werror=array-bounds] while(p[ps2+1].y == cury) ^ src/mame/video/model1.c:319:17: error: array subscript is below array bounds [-Werror=array-bounds] while(p[ps1-1].y == cury) ^ src/mame/video/model1.c:319:17: error: array subscript is below array bounds [-Werror=array-bounds] src/mame/video/model1.c:329:17: error: array subscript is above array bounds [-Werror=array-bounds] while(p[ps2+1].y == cury) ^ src/mame/video/model1.c:329:17: error: array subscript is above array bounds [-Werror=array-bounds] Compiling src/mame/video/segag80v.c... src/mame/video/model1.c:305:17: error: array subscript is below array bounds [-Werror=array-bounds] while(p[ps1-1].y == cury) ^ makefile:909: recipe for target 'obj/sdl64/mame/video/model1.o' failed ======================= |
||||
Steps To Reproduce | |||||
Additional Information | |||||
Github Commit | |||||
Flags | |||||
Regression Version | |||||
Affected Sets / Systems | |||||
Attached Files
|
|||||
Relationships
There are no relationship linked to this issue. |
Notes
7
No.10863
Tafoid Administrator
Jul 25, 2014, 14:34
|
Normally, we do not accept any bugs for cutting edge gcc versions. I know 4.8.x has been well tested as it is part of many linux distros. I"ll acknowledge for now. I guess. Don't expect help to correct until, I'm told, the next Fedora is released (months away)? |
---|---|
No.10864
R. Belmont Developer
Jul 25, 2014, 16:10
|
This is a double whammy as SDL UME builds are not supported, and GCC 4.9 will not be added to the supported compilers list until Fedora 21 is released (currently mid-November 2014). You can hack around this specific issue fairly easily by editing src/build/flags_gcc.mak and changing it so the 4.8 changes apply to 4.8 or 4.9, but anything that goes wrong after that (including if the games don't actually run) is entirely self-service I'm afraid. |
No.10865
gauel Tester
Jul 26, 2014, 01:31
|
> I know 4.8.x has been well tested as it is part of many linux distros. Thanks for the info. I'm using Arch Linux now. > SDL UME builds are not supported Sorry for annoying you. I always use UME and I thought model1.c was not a UME specific file. The build error happens on "TARGET = mame" too. |
No.10867
R. Belmont Developer
Jul 26, 2014, 17:47
|
4.9.0 and 4.9.1 have been reported (by Linus Torvalds himself!) to generate incorrect code for the Linux kernel. So it is strongly recommended you use 4.8.3 if it's available on your distro. |
No.10868
Haze Senior Tester
Jul 26, 2014, 18:34
|
This is nothing to do with UME, RB is trolling. You'll get the same with a regular MAME build as you observe. |
No.10896
gauel Tester
Aug 5, 2014, 04:06
|
1. We need to modify src/emu/cpu/m68000/m68k_in.c https://aur.archlinux.org/packages/sdlume/ > It is easy to fix: convert the file src/emu/cpu/m68000/m68k_in.c to use Unix (LF only) line endings. 2. We need to add options on gcc-4.9.1/Arch Linux. $ make NOWERROR=1 OPTIMIZE=2 PYTHON=python2 Now I can run UME on Arch Linux. |
No.14942
Robbbert Senior Tester
Apr 10, 2018, 08:55
|
GCC 4 is no longer supported, therefore this report is no longer relevant. |