Étienne Bersac wrote:
> Hi,
>
> Uploaded the 0.5 version which includes ift-export to export to Intel
> HEX format as well as info and manpages for ift-extract and ift-export.
> ift-load now uses syslog facility.
>
> The TODO list is empty. :)
>
> http://bersace03.free.fr/ift/
>
> Regards,
> Étienne.
A few things:
1) The build system seems broken to me:
* It doesn't set proper permissions (e.g root for ift-load)
* It doesn't default to paths that make sense and I had to explicitly
configure as shown below to get the thing to install properly:
./configure --prefix=/ --libdir=/lib --bindir=/bin --sbindir=/sbin
--datarootdir=/usr/share --datadir=/usr/share --sysconfdir=/etc
And that would still generate "//lib/firmware..." path. Because of the
-prefix=/ I pressume.
2) The udev rule doesn't work for me in terms of ift-load doesn't load
the firmware. Probably the USB_BUS and USB_DEV env variables are not set
as you seem to expect them to be in ift-load. May be you should either
pass those explicitly in the udev rule or just make ift-load find the
device itself basid on the device and vendor id?
3) In ift-load (load.c) you are using "ift-extract" instead of
"ift-load". :) E.g:
g_set_prgname("ift-extract");
...
gchar *desc = _("Extract firmware using ift-extract.");
etc...
4) ift-load (load.c) syslog logging won't work because you are:
* Meant to use syslog/vsyslog functions (instead of g_error) to send
the log messages to the syslog.
* openlog("ift-extract", 0, LOG_SYSLOG); - That should probably be
something like:
openlog("ift-load", LOG_PERROR, LOG_USER);
The man page says LOG_SYSLOG is for generated internally by syslogd
messages. LOG_PERROR will also print to stderr.
Also on a side-note: g_error is a bit excessive probably because it
aborts immediately.
Regards,
Ivan
_______________________________________________
Linux-uvc-devel mailing list
[email protected]
https://lists.berlios.de/mailman/listinfo/linux-uvc-devel
