On 02/14/2017 05:02 AM, Yann Ylavic wrote:
The previous test (mod_include) does use mod_bucketeer (including with subrequests), possibly some remaining buckets (EOR?) from there somewhere in the stack (core or some downstream filter's f->bb)? That'd be very wrong too (but I can't see something like that with gdb, though)...
Ah, it does seem to be related to mod_include, not mod_info. If I run the mod_include tests followed by mod_lua tests, the mod_lua tests report a segfault halfway through...
The crash looks like it's actually coming from a mod_include connection that, for some reason, takes a very long time to complete. mod_reqtimeout is also involved in the stack.
Admittedly bucketeer_out_filter() is not very nice because it does not "consume" its given brigade (nor does default_handler() clear it afterward), but that shouldn't be an issue since anyway nothing should use them once the request is destroyed. Do you have a backtrace of the crash (and/or a breakpoint in bucketeer_out_filter() for the test)?
Attached. --Jacob
backtrace
Description: application/info
