totalspectrum/loadp2

Loader for Parallax Propeller P2

loadp2's Issues

Improve error message for busy COM port

Currently it's just the same Could not find a P2 on port COMx as everything else, which may be somewhat confusing.

Add warning when trying to load a source file

More often than I want to admit, I've loaded a spin file instead of a binary file and then tried to troubleshoot why nothing works.

Simply printing a warning if the file extension to be loaded is spin/spin2/c/bas would be quite nice.

loadp2 use nanosleep for linux

in osint_linux.c ftime is used to implement msleep function
Trying to port flexprop to OpenBSD and this function (and related sys/timeb.h) has been removed a looong time ago there as it has been deprecated in POSIX standard. Even in Linux it is obsoleted according to the man page.

The usleep function it replaces is ALSO obsoleted and replaced by nanosleep instead as below. Sorry no patch/pull request and maybe too simplistic as it will fail horribly when using more then 999 ms. However for loadp2 it works without any problems on both OpenBSD and Linux (Mint & raspbian)

void msleep(int ms)
{
    struct timespec remaining, request = {0, ms * 1000000};
    nanosleep(&request, &remaining);
}

PS I'll come back with a pull request for all the changes needed to get it to compile on OpenBSD once I've completed it all. For now it is mainly in the Makefiles, but still struggling with proploader and port selection in flexprop (similar to MacOS though)

build_macos does not create same binary as Makefile

The loadp2 Makefile creates and executes the following compile command:

gcc -Wall -O -DMACOSX -o build/loadp2 loadp2.c loadelf.c osint_linux.c

The build_macos file executes the following compile command:

gcc -Wall -DMACOSX loadp2.c loadelf.c osint_linux.c -o loadp2

Looks like the Makefile variant is building with "-O' optimizations... Should these be the same and is build_macos even needed?

loadp2 should probably quit terminal mode when the output pipe breaks

Currently it just continues piping terminal data into the broken pipe....

I'd try looking into it myself, but

• I can't get it to build
• No idea why it isn't already dying from SIGPIPE (maybe it does on linux and mingw just needs special care?)

loadp2 terminal just kinda dies when there's too much data?

Here's a fun one:

_CLKFREQ = 100_000_000

DAT
        asmclk
        waitx ##_CLKFREQ*2

        mov pa,##1000
.loop
        debug("this is pretty mid tbh ",uhex_long(pa))
        djnz pa,#.loop

        debug("ok last one")
        jmp #$

Running this through loadp2's terminal will invariably fail somewhere around pa = $0000_02DC. Is fine if you use another terminal (tested with PST)