[Rd] Compilation Error when DEBUG_approx Toggled on in RISC-V

Martin Maechler m@ech|er @end|ng |rom @t@t@m@th@ethz@ch
Fri Feb 24 16:03:06 CET 2023


>>>>> Jane He 
>>>>>     on Thu, 23 Feb 2023 15:33:23 -0800 writes:

    > Hi all,

    > While compiling R to RISC-V64 architecture and debugging in R's C source
    > codes, I think I have found a small bug. Can anyone please verify whether
    > it is a real bug?

    > The possible bug lies in the file `R-4.2.2/src/library/stats/src/approx.c`
    > in function `R_approxfun` around line 148:

    > #ifdef DEBUG_approx
    >    REprintf("R_approxfun(x,y, nxy = %.0f, .., nout = %.0f, method = %d, ...)",
    >             (double)nxy, (double)nout, Meth->kind);
    > #endif

    > However, there is no `Meth` defined in this function, causing a compilation
    > error when `DEBUG_approx` is toggled on. The real `Meth` is actually
    > defined in the `approx1` function, and the Meth seen here is probably a
    > historical artifact that was never erased. To fix the error, I suggest
    > changing `Meth->kind` to `method` in the `R_approxfun` function as follows:

    > #ifdef DEBUG_approx
    > REprintf("R_approxfun(x,y, nxy = %.0f, .., nout = %.0f, method = %d,
    > ...)",
    > (double)nxy, (double)nout, method);
    > #endif

    > If this is really an error, I hope this small fix will be helpful to the R
    > community.

well,  you are right this has been a cut'n'paste thinko (almost
surely by me) that has probably never shown but to the 0, 1, or
2 people in the world who may have manually compiled *and*
defined DEBUG_approx ..

    > Thank you for your time and consideration.

and to you, too!
The fixed is part of the source now.


Best,
Martin

    > Best,
    > Jane He

    > _________
    > University of California, Irvine
    > Student of Master of Software Engineering
    > 2022-2023 cohort

    > [[alternative HTML version deleted]]

    > ______________________________________________
    > R-devel using r-project.org mailing list
    > https://stat.ethz.ch/mailman/listinfo/r-devel



More information about the R-devel mailing list