Hey! I've been trying to use SSL for rBootHttpUpdater on my private (public-facing) Nginx server without luck. I've already tried settings SSL buffer size to 4k (as someone mentioned that's the size used by google), but it still seems to fail on memory allocation after some packets received. Maybe our SSL implementation is leaking, or it's just leaking in case of rBootHttpUpdater?
It was BOOT_GPIO_SKIP_ENABLED. I think maybe we should disable this by default. Out of the box, rboot will use GPIO16 to decide whether to select another ROM at boot. I think this should probably be something a developer consciously chooses.
What the hell. I've just wanted to subclass rBootHttpUpdate and stub out flash writing as it might've had something to do with the crashes (but didn't do that, just made an empty subclass), and now it actually managed to do a full update without crashing for the first time. But now it keeps failing again. :D
simple downloadString seems to properly download whole firmware (but getResponseString() returns ~6kb of data, which makes sense)
Oh crap, nvm, it fails now because I updated to old firmware, and Basic_rBoot by default switches to rom 1 (and default make flash only writes rom 0) - this actually sounds like a bug for me. IMO flash should overwrite rBoot config to boot from rom0...