- 29 Mar, 2017 4 commits
-
-
Poul-Henning Kamp authored
Insist we get the frames we expect.
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
mark up functions where only it can go. Mark scheduled h2_req's.
-
Poul-Henning Kamp authored
-
- 28 Mar, 2017 5 commits
-
-
Poul-Henning Kamp authored
Fixes #1875 (maybe)
-
Poul-Henning Kamp authored
than to delay a random-ish number of seconds
-
Dridi Boukelmoune authored
A probe may still be running when its backend is deleted. In that case it needs to read the proxy_header field under the lock and give up if the backend was already gone. Not sure how to write a test case for this one. Fixes #2278
-
Poul-Henning Kamp authored
backend if still in mid-transfer. With sufficiently small -bodylen, as in this test, that might or might not happen. Explicitly close from the server side, to stabilize this test. Fixes #2280
-
Poul-Henning Kamp authored
-
- 27 Mar, 2017 1 commit
-
-
Dridi Boukelmoune authored
On systems where both 32 and 64-bit libraries can co-exist, simply inheriting Varnish's prefix is not enough because VMODs might land in the wrong directory. For instance, on rhel you will find both /usr/lib and /usr/lib64. This is usually not a problem with package managers, because they usually take care of everything at configure time, for example RPMs may use a %configure macro that defines all well-known directories. It can however be overlooked when a VMOD is manually built from source.
-
- 26 Mar, 2017 2 commits
-
-
Nils Goroll authored
-
Nils Goroll authored
-
- 25 Mar, 2017 2 commits
-
-
Federico G. Schwindt authored
-
Poul-Henning Kamp authored
-
- 24 Mar, 2017 3 commits
-
-
Dridi Boukelmoune authored
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
-
- 23 Mar, 2017 19 commits
-
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
This reverts the patch for #2207, which as martin points out, is not a good idea after all.
-
Poul-Henning Kamp authored
-
Dridi Boukelmoune authored
Original patch from @joshuaspence. Closes #2262
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
"leave this much free in workspace". Add debug.typesize() which can tell us the C-sizeof() various types.
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
-
Martin Blix Grydeland authored
This enables waiting for the logging thread to idle to make sure all generated log is caught.
-
Martin Blix Grydeland authored
This might be useful for debugging
-
Martin Blix Grydeland authored
The child expects mgt to be consistent in its VCL state changes, e.g. not ask the child to 'use' a VCL that isn't warm. Not doing so will cause the child to assert (see asserts in cache_vcl.c:ccf_config_use). This could trigger because the mgt erranously would set its recorded temp regardless of the child's reply in mgt_vcl_setstate. This patch changes it to only record the new temp in mgt if the child replied success. If the child isn't running the new temp is always recorded. Slightly update v00044.vtc test case to reflect the need to update the state before restarting the child.
-
Martin Blix Grydeland authored
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
not be able to get enough iovec's to do Chunked transmission. Fixes #2275 And #2207 was probably misdiagnosed, but the fix is correct anyhow.
-
- 22 Mar, 2017 4 commits
-
-
Nils Goroll authored
Previously, we would be left with a set up but uninitialized req->vfc when reembarking a request on the waitinglist because the set up was done in CNT_Request, but transport->req_body was only called in cnt_recv. Fixes #2266
-
Nils Goroll authored
-
Poul-Henning Kamp authored
-
Poul-Henning Kamp authored
Fixes #2260
-