[HECnet] Ersatz-11 PDP-11 emulator V7.0

[Cory Smelosky]

Next you just need to support non-x86 architectures. ;)

Sent from my iPhone

On 1 Nov 2013, at 05:28, "John Wilson" <wilson at dbit.com> wrote:

Ersatz-11 V7.0 is done.   This is the 20th anniversary release (development
began on 31-Oct-1993), and it's got a bunch of new goodies.   The free demo
versions (for DOS, Linux, OS/2, Stand-alone, and Windows) can be downloaded
from www.dbit.com.

There's a *lot* of new code.   If that means there's going to have to
be a quick update it's going to be hard to resist calling it "V7.0-07".

There's no one huge new feature, but lots of little ones:

- CR11/CD11 punched card readers

- DMR11/DMC11 network ports (speaks real DDCMP to DECnet/RSX's software DDCMP)

- DN11 autodialer

- KG11A CRC boards

- DU11/DUV11 sync SLUs proof-of-concept (but Z85230 driver needs finishing)

- SCC: driver for Sealevel sync boards (see above -- at least async works)

- Ridiculously improved VT100 emulation.   Smooth-scroll and blink on OS/2
and Windows.   Also SET-UP mode.   Try pressing SET-UP (default = Shift/Esc)
and then 0 on OS/2 or Windows for a good laugh, if you remember the real
thing.   VT100 w/o AVO, VT101, and VT102 emulations are now included (the
autowrap behavior differs on all three -- thank you Will Kranz for the
VT102 for testing).   This was motivated by finding that the VT100 emulation
was failing the VTTEST suite, but further investigation showed that a real
DEC VT100 "fails" too, in a fairly similar way (virtually identical now).

- ASSIGN ... /TTSYNC switch adds driver-level XON/XOFF flow control for
output to most serial device types (and is enabled temporarily during all
commands -- you really need it if smooth-scroll is turned on).

- Scrollback buffer in VT100 sessions -- finally

- HAYES: filter driver for data-leads-only modems with Hayes AT command set
(turns commands/responses into modem control/status leads, and connects to
the DN11 emulation) -- yes this would have been more useful in 1993

- OSPRINT: driver (emulates dumb line printer using OS-supplied printer
drivers in OS/2 and Windows).   Optionally adds green-bar-paper underlay,
which seems silly until you see how nice it looks in CutePDF.   OK it's silly.

- Unit number ranges in ASSIGN/DEASSIGN, MOUNT/DISMOUNT, and SHOW --

Tries to be clever about incrementing unit numbers for the PC device:
ASSIGN YZ0-3: COM1: (or /dev/ttyS0) gives you 4 DZ lines using 4 COM ports
MOUNT DU0-1: rsx0.dsk gives you two disks using rsx0.dsk and rsx1.dsk

*But* if there's nothing to increment then you get identical devices:
ASSIGN YV0-15: TELNET: gives you 16 DHU lines on the same Telnet server
MOUNT DC0-3: RAM: gives you a fully fleshed out RC11 (four RAM disks)

- Generic host-OS-independent names for several kinds of ports that can
be autodetected.   SHOW ASYNCPORTS, SHOW PRINTERPORTS, SHOW SYNCPORTS,
SHOW DIGPORTS, SHOW GPIBPORTS gives you a list of what ASYNCn: etc.
will get you right now.   This is supposed to make E11.INI files a little
more portable (so you don't have to change COM1: to /dev/ttyS0 etc.).

- Host-OS-independent NULL: devices for most device types.

- IEU11/IEQ11 dual GPIB port bare framework (proof of concept:   works only
with NULL: device, does no actual I/O, but accepts many commands and LOGs
their behavior)

- RAMdisks now take their default size from the drive type (which may
itself be defaulted)

- "Short" disk images have the missing part emulated as a null area
(so it's OK if the bad-block track wasn't included in the image --
you don't get errors if you touch it anyway)

- DB:/DR: (Massbus moving-head disks) now enabled in Demo version;
also OA:/XA: (DR11C/DR11W), used as Ethernet IPL, internal mP IPL,
or simple LPT port interface

- Raw floppies (PC formats, for now) on Linux, OS/2, and Windows

- /CYL/HEAD/SECTORS switches work on all C/H/S disks

- Disk logs now calculate the starting block number on C/H/S disks;
LOG /PC includes PC in log files

- SET ddcu: READONLY / READWRITE can write-lock or write-enable most
disks and tapes w/o re-mounting them

- SET IDLE DELAY=n adds a n-millisecond delay during WAITs

- More use of helper threads (keeps long searches in .TAP files from
interfering with PDP-11 execution -- and incidentally means that more
of the PC devices are mP-safe even on the DOS and Stand-alone versions,
since in that case "helper threads" means migrating DOS calls to CPA where
it's safe -- not that mP is officially supported of course)

- LoadModule, GetSymbol, and UnloadModule DLL calls (OS/2 and Windows)

- Many, many, many bug fixes -- hopefully more than were introduced by
the new code

- The update disks will be CDs this time -- yes finally entering the 1990s

John Wilson
D Bit