mt-st (1.9-1) unstable; urgency=medium

    This release fixes many cases where mt and stinit silently did something
    other than what was asked. Some of these fixes may affect existing
    setups and scripts:

    - mt now checks numeric arguments strictly. Before, it silently turned
      invalid, empty or out of range numbers into another value (often 0)
      and still ran the command, e.g. "mt fsf abc", "mt seek ''" or
      "mt defblksize 64k". These are now rejected with exit code 3, and the
      tape is not touched. Scripts that pass e.g. an empty variable will now
      fail.

    - stinit no longer drops a value placed right before a closing brace,
      e.g. "mode1 blocksize=0 compression=1}". These settings used to be
      ignored and are now applied, so check your stinit.def if it has
      definitions written this way.

    - stinit now detects when the SCSI INQUIRY command used to identify a
      drive fails, and may refuse to configure a drive that reports an
      error. If a drive that worked before is no longer configured, please
      report it with the error message stinit prints.

    - stinit without arguments now initializes all drives even if one of
      them fails, instead of silently stopping at the first failure and
      exiting 0. It exits with status 1 if any drive failed, including a
      drive with no matching definition. Check any boot scripts or services
      that run it.

    - Definition blocks without any modeN keyword are ignored, as they
      always have been. The documentation wrongly said they set common
      parameters for all devices. If you relied on this, move those
      parameters into each device block.

    For full details, please read the upstream changelog in
    /usr/share/doc/mt-st/changelog.gz.

 -- Iustin Pop <iustin@debian.org>  Mon, 16 Nov 2026 22:52:53 +0100

mt-st (1.3-1) unstable; urgency=medium

    The algorithm for activating tape device parameters has changed
    significantly. In previous versions (up to and including 1.1-6), there
    were three things which were used to do this. First, an init script
    run during system boot (rcS) was setting parameters for all tapes if
    the `st' driver was built-in to the kernel. Second, a modprobe
    configuration file was shipped which modified the `st' module
    insertion to also execute stinit afterwards (via the init script with
    "modload" parameter). Third, an udev configuration rule was shipped
    which called the init script with the modload parameter.

    This was an extremely complicated way to do a simple thing: after each
    tape device activation, call stinit with the device path (irrespective
    of how `st' driver is built, and independent of the boot process). So
    the init script and modprobe configuration file were dropped, and only
    the udev rule was retained, which now directly calls `stinit
    $device'. If you don't run udev, you need to ensure calling once
    stinit post-boot (e.g. from rc.local) or post-device insertion
    (manuall).

 -- Iustin Pop <iustin@debian.org>  Wed, 04 May 2016 01:35:47 +0200
