Would it not be feasible to retain *all* the panic stacks? That is, the
original panic stack; the stacks for the channel operations that
subsequently failed because the other end had been dropped; and so on?
Unless one has a reason one must throw away information about a crash, it's
always been my
To do that, we'd have to add some delay into sending the crash report, e.g.
generate the report a second after the first panic. My concern about doing
this is just an issue of adding code complexity and brittleness to the
constellation, which is our single point of failure.
On Jul 6, 2016 07:05, "J
Hi all,
I've made a few maps of the servo/servo contributors.
https://flic.kr/s/aHskDBNv66
This is done by downloading the list of contributors from GitHub API, get
their locations from GitHub, clean them up manually and import them into
Mapsdata [1].
The coordinates are from Mapsdata's geocodi
3 matches
Mail list logo
