It sounds odd. The Try REBOL process is simply killed by Linux, so you're probably dealing with an incomplete HTTP request/response cycle
I've had http errors too when I've used the web interface
not offten ... and just print system/version
The attempt to create a small image on your server, good enough for an avatar image, for example, was OK. Also, it seems that for sites like imgur.com and imm.io you can upload without credentials.
That would be OK then
rebolbot might have to use rebol2 to run as I'm getting too many odd behaviours with rebol3
I have it performing quite well building websites, but not as a long-running process, nor doing network I/O
I did come across a series of debilitating bugs while developing the Red bridge
is anyone running a rebol3 long running process?
I have, but there are some bugs (like one Fork found) that can trip things up in some cases. Fork's bug is on my todo list.
Forks bug causes crashes when you do a lot of object allocation sometimes, even the internal objects allocated by binding loops.
At least I think so. I have code that can predictably crash R3 that I think might be related to Fork's crashes, but I don't have Fork's crashing code so I can't test it. But I do have my crashing code and know how to fix his bug, so I'm hoping that will make my crash go away too. If it does I'll let you know so you can include it in RebolBot.
Plugged in Martin Johannesson's shrink script ... note that "Appologies" is mispelled in the script