On Sat, Sep 25, 2021 at 6:23 PM Rob Landley <[email protected]> wrote:
> On 9/25/21 3:02 PM, enh wrote: > > macOS prints "Darwin", which i assumed was what you wanted? (the > built-in macOS > > uname doesn't have -o.) > > That's the kernel type. (Default output is -s.) > > Added help text describing -o as the userspace type. Made it just always > say > "Toybox". > > I ripped out the #ifdef hack from the dawn of time (back when I left > busybox) > and added "linux32" (a kernel level hack from 2001, introduced in > 2.1.115pre3). > Which means there's a new "linux32" command (didn't make that up, it's on > debian > already) which probably doesn't build on macos? you know that's what the little green check mark or little red cross next to each commit on https://github.com/landley/toybox/commits/master tells you? you can click on the red crosses to see the specific build log. in this instance: https://github.com/landley/toybox/runs/3714437120 > I stuck it in uname.c because > literally all it does is tell uname to lie (man page even says so), but > macos > probably hasn't got the header, the function, or the macro. (Even though > it's > sys/personality.h rather than linux/personality.h, it SMELLS like a linux/ > header. Certainly not in posix.) > > I.E. It might need to have the #include moved, and a function prototype and > macro #define in the #else case. Lemme know if that's the case. (I would > have > already, but... sys? Where did they get sys/ from?) > does that make sense? or should we just `#ifndef __APPLE__` around the whole of linux32 in uname.c? (see attached patch in case that's not clear. tested and works on macOS locally.) > Rob >
0001-linux32-ifdef-out-to-fix-macOS-build.patch
Description: Binary data
_______________________________________________ Toybox mailing list [email protected] http://lists.landley.net/listinfo.cgi/toybox-landley.net
