davidkinder / git Goto Github PK

I'm a complete newbie at C, but I've found that I needed to add -lm to LIBS in the Makefile in order to compile it with gcc 5.4, which is the default for Ubuntu 16.04.

I don't know if that's a kind of change you want to have by default in the Makefile, or if it would cause problems for other compilers or other OSes. But I'm suggesting it here just in case.

I do not know how to code, but I attempted to follow the instructions as best I could and got this as an error upon attempting to build
1>------ Build started: Project: Git, Configuration: Debug Win32 ------
1>Building help file...
1>The system cannot find the path specified.
1>The system cannot find the file specified.
1>C:\Program Files (x86)\Microsoft Visual Studio\2019\Community\MSBuild\Microsoft\VC\v160\Microsoft.CppCommon.targets(241,5): error MSB8066: Custom build for 'help\Git.hhp' exited with code 1.
1>Done building project "Git.vcxproj" -- FAILED.
If it is something very obvious I am sorry for taking up your time.

The existing RNG is taken from Glulxe, which (as of Glulxe 0.6.1) has switched to a better algorithm. Glulxe also allows use of OS-specific RNGs. Both these changes should be adopted into Git, see erkyrath/glulxe#33.

hi, is it possible to change the UI Color before compiling? and if so where do i have to define the colors

Glulxe's random number generator is much better than calling rand(), so use that instead.

Direct threading, which uses GCC's label-as-value extension, does not work when git is compiled as a 64-bit application. This is because the label-as-value value, which is of type void*, cannot (in a 64-bit application) be converted to a git_uint32 without loosing some of the value, since the uin32 is 32 bits, but the pointer is 64 bits.

What on Earth?
1.3.5 is somewhat fine. But 1.3.6 is rigged with virus?

Windows Defender detected one.
Virus Total detected tons of it.

See www.intfiction.org/forum/viewtopic.php?t=20768

The problem is in the code at "resume_number_L7_digit_L6": this converts the number to be printed to a positive value:

L1 = (L7 < 0) ? -L7 : L7; // Absolute value of number.

But since L1 is a signed 32-bit variable, this won't work for -2147483648, as +2147483648 can't be represented as a signed 32-bit value.

See the equivalent change in Glulxe: erkyrath/glulxe@fffcf01

Visual C++ doesn't support GCC's label-as-value extension, so Git can only be compiled with Visual C++ without direct threading. The code mostly seems to work, but for some games (e.g. Alabaster) halts at some point with a "bad opcode" error. Need to investigate whether this is a problem with Git or with Visual C++.

If your game is too big for Git to save undos, @saveundo should return 1, but it doesn't. From the looks of the source code it isn't programmed to do that, the only return number it has is 0.

Also, I notice that gMaxUndoSize is set to 256Kb. That won't make your recent increases to the buffer size pointless will it? (I'm guessing that's the max size of an individual undo slot? Should it perhaps be increased too just in case?)

(Also also, when you said "Increased the default undo buffer size in all ports" you only changed code for Windows, had you already changed it for the others? Because it doesn't report failures I'm thinking I'll have to detect the version of Git in order to switch on Ultra Undo, and I'm wondering if 1.3.2 is safe or not.)

Compiling with clang 7.3.0 on Mac OS X 10.11.6 I get this error:

:76:/usr/include/assert.h5:
:/usr/include/assert.h66: 76:error: 66::use of undeclared identifier '__noreturn__' 76:66: error
: use of undeclared identifier '__noreturn__'

I'm able to resolve it by setting defining git_noreturn as __dead2 in config.h, but I don't really understand what's happening or why that works.