Re: [libvirt PATCH] scripts/rpcgen: fix 64 unsigned int test on macOS
On Thu, Nov 30, 2023 at 11:40:17AM +0000, Daniel P. Berrangé wrote:
On Thu, Nov 30, 2023 at 05:17:09AM -0500, Andrea Bolognani wrote:
> On Thu, Nov 30, 2023 at 09:24:15AM +0000, Daniel P. Berrangé wrote:
> > +++ b/scripts/rpcgen/tests/demo.c
> > +#ifdef __APPLE__
> > +# define xdr_uint64_t xdr_u_int64_t
> > +#endif
>
> This makes the compilation error go away, but I'm not convinced it's
> the right fix.
>
> IIUC demo.{c,h} are generated from demo.x, so wouldn't this be
> overwritten the next time we regenerate them?
No, the demo.{c,h} are reference implementations from the original
rpcgen, which we'll never change.
The demo.x is processed by our own RPC generator code, and then we
check that both the original demo.c, and our own generated version
produce the same data on the wire.
Your description doesn't seem to line up with what I'm seeing here:
$ git status
On branch rpcgen
nothing to commit, working tree clean
$ echo "/* TEST TEST TEST */" >>scripts/rpcgen/tests/demo.c
$ git status
On branch rpcgen
Changes not staged for commit:
(use "git add <file>..." to update what will be committed)
(use "git restore <file>..." to discard changes in working directory)
modified: scripts/rpcgen/tests/demo.c
no changes added to commit (use "git add" and/or "git commit -a")
$ VIR_TEST_REGENERATE_OUTPUT=1 meson test -C build
...
$ git status
On branch rpcgen
nothing to commit, working tree clean
> Also both Linux and macOS have xdr_u_int64_t, and we already
seem to
> use the u_ variant for other things (u_short, u_int), so couldn't we
> just use xdr_u_int64_t everywhere and avoid the conditional?
IIRC, that would then break Windows
You're right, portablexdr doesn't seem to have xdr_u_int64_t either.
--
Andrea Bolognani / Red Hat / Virtualization