Hitting backspace for cryptsetup passphrase introduces a character

[Emilio Pozuelo Monfort]

Binary package hint: usplash

If I hit backspace when usplash asks me for the passphrase to unlock my HD, it should remove the last char, and not introduce a new one (unless backspace is supposed to be a valid character, but I don't think so).

Actually if you press a wrong key and you know it, you can't hit backspace, but have to press enter so the passphrase fails and you can enter it again.

[Manuel Schmid (fliegenderfrosch)]

I can confirm this bug as I have the same problem.

[Saivann Carignan (oxmosys)]

I also confirm this behavior in ubuntu gutsy, can someone confirm if this problem also happens with hardy?

[Emilio Pozuelo Monfort (pochu)]

I do

[Martin Pitt (pitti)]

This annoys me as well, I'll have a look at it.

[Martin Pitt (pitti)]

Fixed in bzr head.

[Launchpad Janitor (janitor)]

This bug was fixed in the package usplash - 0.5.11

---------------
usplash (0.5.11) hardy; urgency=low

  * libusplash.c, usplash_timeout_get_string(): Fix wrong and confusing
    indentation.
  * usplash.c: Fix argument parsing for INPUTTIMEOUT (off-by-one error, and
    wrong string length).
  * libusplash.c, usplash_timeout_get_string(): If no char was entered during
    timeout, return an empty string instead of poking a 0 into random memory
    and returning a random string.
  * libusplash.c, handle_timeout_input(): Always output the string to the
    output FIFO, even if input timed out and the string is empty. Otherwise
    the usplash client cannot know whether it should starve and die waiting
    for the reading FIFO or not listen to it and let usplash starve and die
    for writing the FIFO.
  * libusplash.c: Interpret INPUTTIMEOUT value as deciseconds instead of
    seconds. This makes it much more useful for polling for input while doing
    something in the background.
  * usplash_write.8: Add documentation for INPUTTIMEOUT.
  * libusplash.c, usplash_timeout_get_string(): Handle backspace properly.
    (LP: #163793)

 -- Martin Pitt <email address hidden> Mon, 11 Feb 2008 18:06:37 +0100