Swordfish

Swordfish has worked well on my computer but has recently become unreliable. Sometimes when compiling, it just stops part way through, not completing the process. Sometimes it locks up during the process with the message at the bottom of the screen "compiling, please wait." Other times, it seems to compile properly but the resulting code doesn't work properly.

Exiting the program doesn't change anything. Rebooting allows compiling to complete but the results can be rubbish.

I have uninstalled the program using Windows tools, deleted everything remaining in the Mechanique folder in the Programs directory and re-installed Swordfish without improvement. I have the paid version of Swordfish, and the dialog box says it's up to date.

My computer runs Windows 8.1, with 8 GB of ram and the 64 bit OS.

Help! Any ideas?

I tried installing SwordfishSE to see if it would work. Same problem.

In this case, the compiling progress bar goes about a third of the way across the box in the first step, and I'm dumped back to the IDE screen. No errors or messages, and it's not locked up

It sounds like something has changed recently. I'd go through the usual list I do when a windows program quits working...
chkdsk, turn off virus scanner, disable firewall, disable anything that wants to auto update, etc.

I usually end up restoring to a previous point in time when things worked.

Thanks Jerry,

I have some troubleshooting to do.

Somehow the recovery feature wasn't on on my PC. Windows 8.1 does have a "refresh" feature that fairly quickly resets the computer to as-new settings without deleting files. Programs do have to be re-installed however.

I did that this morning. The first thing I installed was Chrome. The next Swordfish and then the PICkit GUI. Back in business.

I did "discover" an interesting problem just now...or call it a feature. I'm using the PICkit logic tool to exercise a couple bits on PortB to test my code, with the PICkit connected to the ICSP connector. This gives me RB6 and RB7. Pin 6 on the ICSP socket goes to something other than PortB.

Pin 1 of the PICkit logic tool can be used as a output only...which is connected to /MCLR on the ICSP connector. Guess what? It defaults to 0. I spent a few seconds panicking that my simple code still wasn't working, because the PIC was being held in reset by the logic tool!

Just FYI. With an 18F25K22, a 20MHz crystal and the PLL enabled, you can read PortB, test it against the previous valve and stuff it into an array if it's changed with a sample rate over 200kHz!