Does not compile on Windows
Kylogias opened this issue · comments
On windows, using make that comes with mingw-w64 (i think), the compilation fails with error 'unistd.h' file not found
Windows 10
make version 4.3
Clang version 11.0.0
EDIT :: Apparently my clang toolchain was incorrect, but now I'm missing poll.h instead of unistd.h
It's a missing header, are you sure you installed everything you need?
It's a missing header, are you sure you installed everything you need?
I searched it up, and it's for accessing POSIX os procedures
I feel like it's because since Windows it's exactly POSIX compliant.
Does MinGW have a POSIX compatibility layer?
Using the MSYS2 version of clang worked, but now I'm getting a 404 for poll.h
This requires complete POSIX compliance from the host system, sorry. Will add the info into. the README.
This requires complete POSIX compliance from the host system, sorry. Will add the info into. the README.
Can you give me a list of POSIX functions used so I can TRY to make it compatible with windows?
This requires complete POSIX compliance from the host system, sorry. Will add the info into. the README.
Can you give me a list of POSIX functions used so I can TRY to make it compatible with windows?
Nuke the posix headers see the errors.
There's already POSIX compatibility layers for Windows. Just look online for the one that suits you.
Can I write Windows specific instructions on the README?
You should fork the repository and edit it that way. If our lord and savior jdh likes your contribution he may push it.
Can I write Windows specific instructions on the README?
Feel free to make a PR if you can get this compiling on Windows. Good luck though since the emulator uses readline which is very POSIX-specific, among other things. The test runner also relies a lot on POSIX functions (mktemp, opendir, etc.).
I'm trying to make a Windows "compatibility patch". Is the history feature in Readline ever used outside of adding things to history? If so, where
EDIT :: Just did a find in files, history is unused so I can safely add an empty add_history function
I GOT THE EMULATOR TO COMPILE!!! Halfway done (I might not compile the tests, although I will have to look at them first)
The full thing compiled (even the tests). The emulator was the hardest to compile, I had to make some compatibility code. There was only 1 error in the assembler which I don't think was POSIX specific. The tests compiled first try, no error correction necessary