In https://docs.google.com/document/d/1hpiiywAWDXoPlXHQNPmduyjFTbK_sEFRc8JXV60_n3A we discussed the interaction of the crash reporter and the pingsender, and how networking should work in the future. We decided that the crash reporter should use Necko as the primary means of sending pings/reports, with a fallback using system network stacks (through curl/libcurl, which are available on all target platforms).

The crash reporter is currently using the fallback network stacks as the sole means of sending reports, and is using pingsender (which uses system network stacks itself) for pings.

Using Necko is convenient as it will take into account user settings, such as proxy configurations, and also relies minimally on the system environment. Having a fallback is important if whatever causes Firefox to crash also causes background tasks using Necko to crash.