[chirp_users] New Daily Build

[Build System]

Greetings,

A new daily build of CHIRP is available.  This includes changes made
directly to the tree yesterday, and may include additional features,
bug fixes, and/or bugs.  If you are interested in trying it, grab it
from the following location:

  http://trac.chirp.danplanet.com/chirp_daily/daily-03142012


A list of the changes included in this build (since the last daily)
follows:

Changes for Build #35
[Dan Smith <dsmith at danplanet.com>] [icomciv] Don't attempt to communicate with the radio if we were not
passed a serial connection

[Dan Smith <dsmith at danplanet.com>] Fix setting the index and name fields editable and visible after the
recent change to the bank editor's column count

[Dan Smith <dsmith at danplanet.com>] Emit the changed signal when banks are changed

[Dan Smith <dsmith at danplanet.com>] Add a _wipe_memory_banks() function to the Yaesu base class and call that
when setting an empty memory on drivers that support banks (vx7, vx8,
ft7800)

[Dan Smith <dsmith at danplanet.com>] Report failure to add a memory to a bank to the user

Rev 2: Make the error dialog transient for the owning window so that
it's modal

[Dan Smith <dsmith at danplanet.com>] Fix bank UI to make empty memories insensitive. This improves the visual
nature, but also prevents the UI from asking the driver to add blank
memories to a bank if the user accidentally clicks in one of the boxes.

Rev 2: Ignore clicks on insensitive checkboxes

[Dan Smith <dsmith at danplanet.com>] Make Kenwood drivers refuse to talk to an incompatible model (since
you can now ask it to do so)

[Dan Smith <dsmith at danplanet.com>] Allow choosing a specific model even if the vendor supports a detection
scheme. This lets you choose, for example, a specific model Icom or
Kenwood radio and bypass the detection routine during download. This can
avoid certain issues with detection and/or speed up the process of
starting the download.

[Dan Smith <dsmith at danplanet.com>] Fix some places in the UI where we delete by setting a memory with
empty=True instead of calling erase_memory()

[Dan Smith <dsmith at danplanet.com>] Print a traceback of where a job was created from, when there is an
exception while running it

[Tom Hayward <tom at tomh.us>] [vx5] bugfix: get_raw_memory() was off by one