Issue #1443 - Fix early or duplicate 404 error handling

[shinrich]

No description provided.

[zwoop]

What are the symptoms of this?

[atsci]

FreeBSD build successful! See https://ci.trafficserver.apache.org/job/freebsd-github/1544/ for details.

[atsci]

Linux build successful! See https://ci.trafficserver.apache.org/job/linux-github/1437/ for details.

[atsci]

clang-analyzer build successful! See https://ci.trafficserver.apache.org/job/clang-analyzer-github/109/ for details.

[shinrich]

An assert triggers. A stack trace in the issue description.

#1443

[SolidWallOfCode]

I looked at this as well and it definitely looks like it's an artifact of a remap plugin setting the transaction to be an error. You can see that this happens in a switch statement where t_state.next_action was SM_ACTION_REMAP_REQUEST on entry yet is SM_ACTION_SEND_ERROR_CACHE_NOOP. This could only have been updated via do_remap_request() and presumably via a remap plugin as RemapProcessor doesn't update that member. It also seems reasonable to check if a plugin has error'd out the transaction and give up on further transaction processing at that point.

[atsci]

Intel CC build failed! See https://ci.trafficserver.apache.org/job/icc-github/86/ for details.

[atsci]

FreeBSD build successful! See https://ci.trafficserver.apache.org/job/freebsd-github/1654/ for details.

[atsci]

Linux build successful! See https://ci.trafficserver.apache.org/job/linux-github/1550/ for details.

[atsci]

clang-analyzer build successful! See https://ci.trafficserver.apache.org/job/clang-analyzer-github/218/ for details.