michael

💯 Travis CI needs to be avoided like the plague. There's major problems in that organization these days. Their response to this recent incident in particular was unacceptable: https://cycode.com/blog/github-oauth-compromise-affecting-heroku-and-travis-ci-users/

How about storing the sent transactions counter in a Config program account on-chain!

@bellbind - can you review this PR please?

I think all we need here so to wrap the body of glue/gonk/hardware/ril/rild/Android.mk in a "ifneq ($(TARGET_NO_RILD),true) .... endif" block, then add "TARGET_NO_RILD := true" to glue/gonk/device/toro/maguro/BoardConfig.mk, and any other...

Does the "rild(220)" abort occur immediately at boot? If not, can you please share steps to reproduce.

ok, thanks. my team can debug this today

@jaoo -- https://github.com/andreasgal/B2G/commit/738a74d039701d9d555864a2aa988aa7bef3daa2 pushed for the product version mismatch

@ferjm -- rild needs a bunch of groups (ref: https://github.com/andreasgal/B2G/blob/master/glue/gonk/device/toro/maguro/rootdir/init.rc#L359) so running it from the command-line will usually cause the errors that you've seen. We successfully made a call on...

Issue has been traced down to out-of-date firmware on Akami. Please return the devices that manifest the rild failure for reflashing.

How about we leverage the existing eng vs user/userdebug build flavors? |make config-xyz| seems likely to go away when TGRS occurs, in fact it already has for me: $ ....