- 31 Jul, 2019 3 commits
-
-
Nils Goroll authored
otherwise the topreq thread cleanup may race us and clean things under our feet. In this case it should not really have mattered because of the topreq waiting for the subreq, but anyway, this is how it should be.
-
Nils Goroll authored
-
Nils Goroll authored
-
- 30 Jul, 2019 37 commits
-
-
Nils Goroll authored
-
Nils Goroll authored
-
Nils Goroll authored
-
Nils Goroll authored
We now fini the root pesi while other tasks could still be running
-
Nils Goroll authored
This reverts commit 92621497.
-
Nils Goroll authored
-
Nils Goroll authored
-
Geoff Simmons authored
-
Geoff Simmons authored
-
Nils Goroll authored
-
Nils Goroll authored
-
Nils Goroll authored
-
Nils Goroll authored
-
Nils Goroll authored
-
Nils Goroll authored
-
Nils Goroll authored
-
Nils Goroll authored
-
Nils Goroll authored
because, as in the case of pretendgz, they may be chaging the next level up. e00023 passes now
-
-
Nils Goroll authored
-
Nils Goroll authored
-
Geoff Simmons authored
-
Nils Goroll authored
fiddling the filters manually really does not make much sense, in particular not for our test-cases which are supposed to closely match varnish-cache serial esi behavior
-
Nils Goroll authored
I _suspect_ this caused issues with make failing in dependency tracking for me, like: make[2]: Entering directory '/home/slink/Devel/varnish-git/libvdp-pesi/src' CC vdp_pesi.lo mv: cannot stat '.deps/vdp_pesi.Tpo': No such file or directory Makefile:700: recipe for target 'vdp_pesi.lo' failed make[2]: *** [vdp_pesi.lo] Error 1
-
Geoff Simmons authored
-
Geoff Simmons authored
-
Nils Goroll authored
-
Nils Goroll authored
but better safe than sorry
-
Nils Goroll authored
-
Nils Goroll authored
-
Nils Goroll authored
-
Nils Goroll authored
-
Nils Goroll authored
it's too early in set_delivered() because set_closed() calls set_delivered() on a nexus when all children are already devivered, but that can happen from within the nexus node's fsm. we will add extra finis when it's safe in a later commit
-
Nils Goroll authored
in the varnish-cache variant of the code this works because not VDP_Push is used but the internal push to the parent
-
Nils Goroll authored
When subreqs still used the topreq's VDPs, we could not do this, but now that we properly layer the VDPs again, we can (and should)
-
Nils Goroll authored
-
Nils Goroll authored
waiting for the boc happened at the wrong place *** v1 debug|Assert error in vped_gzgz_init(), foreign/cache_esi_deliver.c line 163: *** v1 debug| Condition(foo->start > 0 && foo->start < foo->olen * 8) not true.
-