Hey, Emacs, we're -*-Text-*- mode!

"-" means not done, "*" in progress, "+" done, "?" means contradictory
test results or error reports and total confusion of the Ftape
maintainer.

(-) Formatting cartridges with Colorado Jumbo 350 drives is broken. As
    usual, Colorado gives zero support.

(-) Use GNU autoconf to figure out all those compatibility problems
    and installation paths

(*) Maybe ftape's fast seek implementation is broken. There are tape
    drive/cartridge combinations that miss the desired location after
    a "seek N segments" or "seek N segments extended" by more than 100
    segments! This can lead to an "infinite" loop with the current
    fast seek implementation. The loop isn't really infinite as the
    driver aborts after some trials with an error. But the desired
    location isn't reached. Should improve the fast seek algorithm
    with a fall-back to a "slow seek", i.e. "logical forward" + reading
    the address marks until the desired location is reached.

(+) Add changes by "Harry G. McGavran Jr. " <mcgavran@rt66.com> to support
    1.2.13 kernels again.

(+) Add hints to the "isapnptools" package to the installation instructions.

(*) There are still "fdc_read timeout" errors. This is caused by
    insufficient communication between the driver and its isr().
    I think this is fixed now, so: "+"?

(*) document that depmod error messages can safely be ignored.
    Documented in the FAQ.

(+) change the ioctl interface for the formatting support. A user level
    program should decide how to format a tape cartridge, given the
    results of the "REPORT TAPE STATUS" and "REPORT DRIVE
    CONFIGURATION" ioctls. The driver module should also probe for a
    possibly supported "SET FORMAT" command, and try to determine the
    tape length. Then the user level routine should figure out what
    cartridge is inserted (if necessary with the help of the
    Experienced User (tm)) and then tell the driver about the
    parameters and format to use. Even the header segment information
    should be remembered in user mode. The user level program could
    even dump the header segment to disk, for error recovery (not done).

(*) write a man-page for the new formatting program. Partly done.

(*) There are still deadlocks possible that cannot be aborted with a
    signal. This really needs to be fixed.

(?) There are TR-3 EXtra cartridges out there that (sometimes?) are
    detected as QIC-3010 (should be QIC-3020). Strange.

(*) Push the whole thing into the kernel.

(+) Maybe get rid of the sftape module.

(+) Only overwrite the portions of existing volume table entries that
    are used by zftape when updating the volume table. Or flag new
    volume, and leave the others completely alone (the latter is
    probably a better solution).

(+) Investigate into probably broken handling of empty tape volumes
    (e.g. produced by "mt weof 3")

(*) Update the documentation. This is a never ending story.

(-) Maybe convert the documentation to sgml.
    Could be done. Best write a texi-to-sgml convert like the texi2www
    is a texi-to-html converter. I just have no time.
    Update the Ftape-HOWTO.

Claus-Justus Heine <claus@momo.math.rwth-aachen.de>

--
     LocalWords:  Ftape McGavran mcgavran rt com isapnptools fdc isr depmod FAQ
     LocalWords:  tm TR EXtra QIC sftape zftape mt weof sgml texi www html momo
     LocalWords:  HOWTO Justus claus rwth aachen autoconf LocalWords
