At least, I _think_ that's what I did.
I concatenated three programs, ran them through the programmer at http://dm41.swissmicros.com/, got my dump text, slapped it onto the DM41L, and then hit the power key, and it never exited serial mode. I had to reset it using the reset button on the back to clear the memory.
Whacking up some of the stuff I wanted piecemeal worked. Interestingly, the shorter listing compiled by the web site has the same ending address as the one that crashed (0x198 registers), which is of course too many --- but the shorter listing comes out to 0xA8 registers, which would fit. (You may want to check my math, actually, the DM41L is on the other side of the table...)
Anyway, no harm done, because all the work is on the PC side. Some observations, though.
- it might be nice if the the web site encoder squawked out the number of registers the encoded core image would actually consume.
- The unit should probably reject loading files that exceed core.
Really cool. I am loving finding the programs online I've wanted to try and either not had listings (no pack available when I had my Dad's HP-41CX) or didn't have the time once I got a HP-41CX about ten years ago to key in.
If it helps, here's a zip of what's going on. All of the code was originally written by someone else --- the uac, uac-wka, and rem bits are from Geir Isene, and the MOONS and H2D and D2S are from various bits on hp41museum.org. Don't send crash.dm41 to a calculator unless you have a paper clip to reset it, though.