Ah, true. What confused me is that I do not get libusb_pollfd_added_cb() call
for the hotplug pipe - so I assumed such pipe is not used.
But what happens is that - libusb_init() that creates the context, calls
usbi_io_init() which creates the pipe and adds the fd's. Thus it is not
possible to g
I just wanted to make a note on your issue tracker about a problem I
encountered. Solution is included.
**Problem:** I have compiled libusbx and installed on a machine running RedHat
6.3. I compiled a program which requires hot plug support without issue.
However, when running the program, hot
Hi, I'd like to follow up on this. I have added in one more submit based on the
review feedback on the thread. Is there any more change that I should do? Is
this change going to be merged in with the main branch?
Thanks,
Kuangye
---
Reply to this email directly or view it on GitHub:
https://gi
The hotplug thread writes a message to an internal hotplug file descriptor.
This message is read by the thread that calls libusb_handle_events() which then
call the callback. The callback is never made by the hotplug thread.
---
Reply to this email directly or view it on GitHub:
https://github.c
When creating fully asynchronous application, the guarantee that all callbacks
happen in libusb_handle_events_*() is good. However, the hotplug implementation
breaks this guarantee - new thread is forked and the hotplug callback is called
from alternate thread. This means application needs to do
Signed-off-by: Moritz Fischer
You can merge this Pull Request by running:
git pull https://github.com/mfischer/libusbx whitespace_fixes
Or you can view, comment on it, or merge it online at:
https://github.com/libusbx/libusbx/pull/157
-- Commit Summary --
* lin