(02-21-2020, 04:22 AM)TheOldPresbyope Wrote: @Tim Curtis
I think it succeeded but I accidentally closed the terminal before I checked that the compile was actually complete. The resulting binary seemed ok. I used readelf -A to check it, for example. One thing though, the binary was smaller by almost 1.5MB than the one in your release. I've just now started a second build from scratch to make sure everything was ok. I also saw that the protocol stuff was taking an inordinate amount of time.
I wasted what time I had available this afternoon and evening trying to set up and cross compile on my x86 linux laptop using https://github.com/rust-embedded/cross
After weeding out the usual blunders by successive approximation (I never claimed to be a Rustacean!) I ran into a build error I couldn't get past
Code:thread 'main' panicked at 'called `Result::unwrap()` on an `Err` value: "Cross compilation detected. Use PKG_CONFIG_ALLOW_CROSS=1 to override"', src/libcore/result.rs:1188:5
I see lots of posts about this Err value but no actionable answer. Setting the environmental variable in Linux doesn't resolve the error. The cross compile is done in a dynamically started Docker container and I probably need somehow to set the variable inside it.
Pity. I was hoping to be able to report a blazing fast build!
Maybe armed with what I learned about Rust today I can muddle through a cross compile on an RPi4B.
As for the RPi1B, doesn't it seem likely this is an just another memory size issue? When I'm running on the RPi0w, I see up to 60MB of swap being used. Of course using swap will cause the gears to grind even slower.
Regards,
Kent
You can always use the crossbuild framework that I sponsored some time ago. I'm currently using to build packages targeted for specific architecture (rpi, rpi2, rpi3) tuning relative compiler parameters, from my laptop. A full build takes (with all packages kernel, mpd, blue-alsa, shairport...) less than 1 hour on a core i7-7700...
