On 30.01.2020 17:54, Anthony PERARD wrote: > On Wed, Jan 29, 2020 at 03:19:05PM +0100, Jan Beulich wrote: >> On 17.01.2020 11:53, Anthony PERARD wrote: >>> +# Handle objects in subdirs >>> +# >>> --------------------------------------------------------------------------- >>> +# o if we encounter foo/ in $(obj-y), replace it by foo/built_in.o >>> +# and add the directory to the list of dirs to descend into: $(subdir-y) >>> +__subdir-y := $(filter %/, $(obj-y)) >>> +subdir-y += $(__subdir-y) >> >> I realize I'll be called guilty of bike-shedding again, and I also >> realize this is the way Linux does it, but what use is the >> intermediate __subdir-y? Linux has no 2nd use, and hence I also >> don't see why we would gain one. I further think according to our >> style there should be no use of tabs here. > > I though the extra __subdir-y that Linux does was to do the filtering on > obj-y right way and not at a later time when subdir-y is used. But in > Linux (now that I look more closely at it), subdir-y is initialised with > := to have the right type, so the extra __subdir-y doesn't appear to be > useful. (And I didn't find any subdir-y=) > > So, I'll add a "subdir-y :=" somewhere and remove the need for > __subdir-y. (And hopefully, no one will add a subdir-y=dir somewhere and > break the build.)
Alternatively, to retain this "latching" effect, how about subdir-y := $(subdir-y) $(filter %/, $(obj-y)) ? Jan _______________________________________________ Xen-devel mailing list Xen-devel@lists.xenproject.org https://lists.xenproject.org/mailman/listinfo/xen-devel
