Code review comment for lp://staging/~beehock/ubuntu/natty/lirc/lirc-fix-695767

Isnt this the current code?

EVENTS_BY_PATH=`ls /dev/input/by-path | sed 's/^/,\
\/dev\/input\/by-path\//' | tr '\n' '\0'`
23 - EVENTS=`ls /dev/input | grep -v "by-path" | sed 's/^/,\
\/dev\/input\//' | tr '\n' '\0'`
24 - db_subst lirc/dev_input_device EVENTS $EVENTS_BY_PATH $EVENTS

Only /dev/input/by-id is down and not the actual device.

Didnt `find /dev/input/* -not -type d -printf ", %p"` cover all the
devices in /dev/input excluding the directory?

I still dont understand what I need to fix.

On Wed, Mar 2, 2011 at 10:11 PM, Martin Pitt <email address hidden> wrote:
>> No, I meant to fix the issue where devices in the /dev/input/by-id are
>> not shown. The current code included the /dev/input/eventN.
>
> Hm, when I install current lirc, I get a full tree of /dev/input/, including by-id, by-path, etc.
>
>> Would there be cases where devices does not appear in in by-id and by-path?
>
> In theory yes, if a device isn't on USB, or doesn't have a vendor/product name. This shouldn't happen with USB devices, only perhaps with some builtin stuff like lid switches, etc.
>
> But in general I'd really avoid hardcoding /dev/input/eventX in a configuration file. It's inevitably going to break, as the enumeration is completely dynamic, and random up to some degree.
>
>> > Please note that your branch changes #DEBHELPER# into the dh_installinit
>> generated code, which should not be done. Please exclude that from forwarding
>> (I'll do that for sponsoring).
>> > --
>>
>> Can you let me know what I did to change that?
>
> I don't know either -- building a package does not change debian/lirc.postinst. The only plausible way I can imagine is that you first modified it in /var/lib/dpkg/info, and ran dpkg-reconfigure until it worked, and then just copied that into the package.
> --
> https://code.launchpad.net/~beehock/ubuntu/natty/lirc/lirc-fix-695767/+merge/51706
> You are the owner of lp:~beehock/ubuntu/natty/lirc/lirc-fix-695767.
>