Menu
▾
▴
1 Attachments
#175 Problems compiling r6987 CSL-Reduce on Macintosh
Hi Arthur,
trying to compile r6987 CSL-Reduce on my Macs I got the following behavour.
1) compilation of native X86_46 CSL-Reduce (macOS Big Sur 11.7.10, macports) fails,
2) compilation of native M1 CSL-Reduce (macOS Sequoia 15.3.1, macports) is OK,
3) compilation of universal M1 CSL-Reduce (macOS Sequoia 15.3.1, macports) fails.
The failure point is again related to the file 'jit.cpp', but for a different reason. As you can see from the attached log file, the compiler complains because there is an attempt to redefine the procedure 'plant'.
In my opinion this is due to the fact that in the Macintosh part of the file 'jit-x86-64.cpp' the procedure 'plant' is defined and after the same procedure is redefined inside the file 'jit.cpp'. The file 'jit-aarch64.cpp' does not contain a definition of the procedure 'plant' so that no problem arises.
The work around I used was to comment out lines 189-192 of the file 'jit-x86-64.cpp' and everything compiles without problem.
I don't know if this work around is correct one, but, as far as compilation of CSL Reduce is concerned, it works.
Best regards
Marco
Thanks and apologies again. That file has been in a atate of repid flux
and I have just checked in revision 6988 which has just built OK on
Linux/x86_64 but that I now try on the Mac - and then on Windows!
Yes your fix makes sense. In my first draft of that part of the code I had
separate versions of plant() for Intel and ARM but I then worked out how
to lift that code so I had only a single version - and what you observe
captured a state where the common version was mostly in place but traces
of the divergence not fully tidied up.
I now move to the Mac to see if I am in good shape now...
Arthur
On Sun, 16 Feb 2025, Marco Ferraris wrote:
Related
Bugs: #175