Because people just can't stop naming the partitions on every single one of their hard drive images `DH0` and `DH1`.
If something doesn't work with this approach, I have no idea why, because as far as I can tell in the example boot ROM I have available to me, this is all that one does.
* Too many 't's in "Shutting down"
* Properly truncate directory that is too long
* Add more space in boxes for text
* Don't screw up destination data if the text is too long
If a path was used as source it was appended to the destination, which
most of the time caused an error. This is because the filename was
isolated but then not used.
When executed under sudo the wrong shell is used by Debian, causing an
error. This patch fixes the script so it specifies which shell to use.
Note that you don't need to run flash.sh under sudo because it sudos
itself.
* Change to Libra Chat
* Change a couple of things that imply Rev B means EPM240 (and therefore
Rev A means EPM570)
* `chmod +x` no longer needed as the git tree contains the exec bit set
* The output of the nprog commands has changed so updated appropriately
nprog now checks if emulator is running and gives a much clearer success
/ fail with a log file for the more verbose output.
Also buptest now won't run if emulator is running.
The separate writes functionality of Musashi was severely getting in the way of any kind of caching of bus longwords, and I found that the negatvie performance impact of disabling separate writes could be negated by creatively applying the address translation cache to the FC reads/writes instead.
This commit requires a `make clean` before rebuilding the emulator with `make`.
this way we can pass around one pointer and access them all in the future.
This happens to win 5-10% performance, because now the code generator can generate a single
global variable pointer to get to all of the translation cache/range fields at once.
