...because of commit 7444cca876cf ("Allow compilation on platforms
without USBHID support").
It does fix building on MacOS, but it's completely and utterly broken,
because it disables USBHID *everywhere*. The USBHID define is simply
never defined.
The USBHID define is purely internal to the other uses, so
src/usbhid.c does this, for example:
#ifdef HAVE_CONFIG_H
#include "config.h"
#endif
#if defined(HAVE_HIDAPI)
#define USE_HIDAPI
#define USBHID
#elif defined(HAVE_LIBUSB) && !defined(__APPLE__)
#define USE_LIBUSB
#define USBHID
#endif
to make USBHID do something.
The correct thing to do would be to revert that commit, and then have
maybe something like the attached.
I never noticed this, because I haven't used my EON Steel in a while
due to missing a pressure sensor, and only tested that bluetooth
worked during the last set of changed.
But now it's almost spring break, and I got myself a new pressure
sensor, and checked that everything looked ok, and it doesn't.
Dirk - I can't test the OS X thing anyway, so I'm just reverting it in
my private tree. I *think* the attached patch "just works", but I'd
suggest you do the real revert and test, and push out the end result?
Linus
src/usbhid.c | 9 +++++++++
1 file changed, 9 insertions(+)
diff --git a/src/usbhid.c b/src/usbhid.c
index 52adf1f..45044a3 100644
--- a/src/usbhid.c
+++ b/src/usbhid.c
@@ -848,4 +848,13 @@ out:
return status;
}
+
+#else /* !USBHID */
+
+dc_status_t
+dc_usbhid_custom_io (dc_context_t *context, unsigned int vid, unsigned int pid)
+{
+ return DC_STATUS_UNSUPPORTED;
+}
+
#endif
_______________________________________________
subsurface mailing list
[email protected]
http://lists.subsurface-divelog.org/cgi-bin/mailman/listinfo/subsurface
