February 2017 Archives by author

• Messages sorted by: [ thread ] [ subject ] [ date ]
• More info on this list...

Starting: Wed Feb 1 03:10:05 UTC 2017
Ending: Tue Feb 28 23:31:05 UTC 2017
Messages: 203

• [master] d5d5637 Redundant assert   Dridi Boukelmoune
• [master] 06108a1 Test what logexpect can't   Dridi Boukelmoune
• [master] 1706409 Half of #2088 is solved   Dridi Boukelmoune
• [master] 0d63862 Include vcl.use in varnish -vcl failure detection   Dridi Boukelmoune
• [master] 24da967 Forgot to remove -err   Dridi Boukelmoune
• [master] 06108a1 Test what logexpect can't   Dridi Boukelmoune
• [master] e300b98 Assume an error when terminated by a signal   Dridi Boukelmoune
• [master] e5edb7c Simplify "Should not happen" checks   Dridi Boukelmoune
• [master] cca7700 Boolean 101   Dridi Boukelmoune
• [master] 03f65b1 Revert "Boolean 101"   Dridi Boukelmoune
• [master] 4d446f4 Don't logexpect partial regular expressions   Dridi Boukelmoune
• [master] 412961a Polish shell usage in test cases   Dridi Boukelmoune
• [master] 45190d6 Code style polish mostly around vtc_fatal   Dridi Boukelmoune
• [master] 8f34e3a Remove duplicated word to improve reading pleasure   Dridi Boukelmoune
• [master] 7a75565 GC obsolete function detection   Dridi Boukelmoune
• [master] 39551cd Check for python >= 2.7   Dridi Boukelmoune
• [master] d437198 Dogfooding   Dridi Boukelmoune
• [master] 276f021 Revert "Document VRT_ipcmp wrt bumping VRT API version"   Dridi Boukelmoune
• [master] b5a5122 Move "older version" VRT changes where they belong   Dridi Boukelmoune
• [master] c0a6194 Inaugurate ongoing VRT changes after 5.0   Dridi Boukelmoune
• [master] 6db7d75 Use closefd() throughout, take 2   Dridi Boukelmoune
• [master] 97c1722 Restore regression test for #2184   Dridi Boukelmoune
• [master] 62219a5 Restore missing newline (follow up to 497b076)   Dridi Boukelmoune
• [master] bad5e55 Attempt a succinct explanation of test gunzip   Dridi Boukelmoune
• [master] 6db7d75 Use closefd() throughout, take 2   Dridi Boukelmoune
• [master] 89a3f0e Use closefd() throughout, take 3   Dridi Boukelmoune
• [master] 6db7d75 Use closefd() throughout, take 2   Dridi Boukelmoune
• [master] 983555d Introduce a ZERO_OBJ macro similar to bzero   Dridi Boukelmoune
• [master] 9230b61 Polish   Dridi Boukelmoune
• [master] 227920a Keep track of WS_Assert_Allocated   Dridi Boukelmoune
• [master] 1e75419 Track VRT_hit_for_pass   Dridi Boukelmoune
• [master] b28c352 Introduce a `vxid` LHS for VSL queries   Dridi Boukelmoune
• [master] 0930c1e s/VXID/vxid/   Dridi Boukelmoune
• [master] 936c339 Use code litterals for the grouping modes   Dridi Boukelmoune
• [master] 83cbb7f Document vxid VSL queries   Dridi Boukelmoune
• [master] 4522084 Typo   Dridi Boukelmoune
• [master] 7d4b4b0 Rename t.vcl to bad.vcl   Dridi Boukelmoune
• [master] 4a98171 Test multiple -f options for varnishd   Dridi Boukelmoune
• [master] c58b30e Test multiple -f options with a bad VCL   Dridi Boukelmoune
• [master] 917773a Test multiple -f options with an empty value   Dridi Boukelmoune
• [master] 296d83b Plug leak   Dridi Boukelmoune
• [master] cdb2edf Give more room to slower platforms   Dridi Boukelmoune
• [master] dcdb250 Add a setenv command to varnishtest   Dag Haavi Finstad
• [master] 272fc95 give some advise on adjusting vsl_buffer   Nils Goroll
• [master] 403c75e Bring back STRING + DURATION   Nils Goroll
• [master] 978d0a8 include config.h first   Nils Goroll
• [master] bfab0a1 check for python >= 2.7   Nils Goroll
• [master] aef7f5c try to stabilize this test   Nils Goroll
• [master] 5a0817d Move minimal reponses to the transport level and use them for H1 100/400/417   Nils Goroll
• [master] 70ba8e0 Send 100-continue after vcl_recv or when caching the request body   Nils Goroll
• [master] 5af2605 Fix cases when we do not want to send a 100 Continue   Nils Goroll
• [master] 88d1d1d Add std.late_100_continue to postpone sending a 100 Continue response   Nils Goroll
• [master] 341fd85 Back to Expect handling according to rfc7231   Nils Goroll
• [master] 4073a4b Parameter ban_cutoff to limit the number of active (incomplete) bans   Nils Goroll
• [master] 2a8398c For OC_EF_REMOVE, clear OC_EF_REFD under the lock to prevent a race with exp_mail_it   Nils Goroll
• [4.1] 51cbf50 Prepare for 4.1.5-beta1 release   PÃ¥l Hermunn Johansen
• [4.1] 5661746 Mention the join character   PÃ¥l Hermunn Johansen
• [4.1] efc000f Allow INT and REAL to be negative.   PÃ¥l Hermunn Johansen
• [4.1] 9920174 Increment fetch_failed for all causes   PÃ¥l Hermunn Johansen
• [4.1] e3e91d0 Insert space around '+' and '-' in generated C code.   PÃ¥l Hermunn Johansen
• [4.1] 9a5afcf Update devicedetect.vcl to current upstream   PÃ¥l Hermunn Johansen
• [4.1] 0c4c164 beresp.backend.ip is only available while the connection is open, which means not in vcl_backend_error{}   PÃ¥l Hermunn Johansen
• [4.1] 1c80c8f Document workuser and minor improvements   PÃ¥l Hermunn Johansen
• [4.1] dd27061 Update changelog   PÃ¥l Hermunn Johansen
• [4.1] 411fb55 Repair formatting mistake   PÃ¥l Hermunn Johansen
• [4.1] 47df804 Prepare for 4.1.5-beta2 release   PÃ¥l Hermunn Johansen
• [4.1] 2c82b1c Prepare for 4.1.5 release   PÃ¥l Hermunn Johansen
• [4.1] 1ff9df8 Minor changelog fix   PÃ¥l Hermunn Johansen
• [master] afeb1a2 Update changes.rst to include info on 4.1 releases   PÃ¥l Hermunn Johansen
• [master] f4dfdb0 No more "changes since" in the changelog   PÃ¥l Hermunn Johansen
• [master] e3ef0db Update links   PÃ¥l Hermunn Johansen
• [4.1] b9a6cf5 Remove duplicated word to improve reading pleasure   PÃ¥l Hermunn Johansen
• [4.1] 3e4cd0b Fix my previous cherry-pick   PÃ¥l Hermunn Johansen
• [master] 03602d9 Don't panic if people send us HTTP2.0 when it is not enabled.   Poul-Henning Kamp
• [master] c7e611c Log the details of the vcl-poker fails to set state in the worker process.   Poul-Henning Kamp
• [master] 0d400c3 Don't double backslash-escape panic messages, and be more generous with their length.   Poul-Henning Kamp
• [master] a3756e2 Don't forget final NL   Poul-Henning Kamp
• [master] 7347e93 beresp.backend.ip is only available while the connection is open, which means not in vcl_backend_error{}   Poul-Henning Kamp
• [master] 21c215f Always initialize ctx->handling to zero, and try to avoid setting it more than once.   Poul-Henning Kamp
• [master] 778dbd1 VSB_quote() is getting too complex, I'll need to revisit that...   Poul-Henning Kamp
• [master] aedc4d6 Also enable VCL_trace in vcl_init/fini which doesn't have buffered vsl.   Poul-Henning Kamp
• [master] 9849fd0 Fall out of compiled VCL code if any statement sets ctx->handling non-zero.   Poul-Henning Kamp
• [master] 3bd4f51 Now that we fall out of VCL the moment ctx->handling gets set, we can ditch the bogus return values and some decorations.   Poul-Henning Kamp
• [master] 52e49f1 Rename debug.init_fail() to debug.fail() and make it general purpose.   Poul-Henning Kamp
• [master] 61a15cb Implement premature VCL failure in vcl_recv{} and vcl_synth{}   Poul-Henning Kamp
• [master] e5efc2c Implement failure in vcl_hash{}   Poul-Henning Kamp
• [master] 851be37 Add a canary in case of failure-failure.   Poul-Henning Kamp
• [master] ba54dc9 add failure handling to vcl_pipe{}   Poul-Henning Kamp
• [master] 6f50a00 Implement fail in vcl_(miss|pass|pipe|purge){}   Poul-Henning Kamp
• [master] 1567375 Untangle the VCL per test-case   Poul-Henning Kamp
• [master] b881699 Implement fail in vcl_hit{} and vcl_deliver{}   Poul-Henning Kamp
• [master] 365c745 Add failure handing in vcl_backend_*{}.   Poul-Henning Kamp
• [master] 81bc4ce Make sure we always see vdef.h before vrt.h   Poul-Henning Kamp
• [master] 25c9e9b Pull in <stdio.h> to maximize chance of finding platform __printflike()   Poul-Henning Kamp
• [master] ebfcb92 Introduce VRT_fail() which fails VCL processing with a single-line message which either goes into CLI (vcl_init{}) or VSL (all others).   Poul-Henning Kamp
• [master] e9e1373 Fail VCL if attempts are made to set any of the top-line fields (url, proto, reason, status etc.) to an empty string.   Poul-Henning Kamp
• [master] 23705c4 Use VRT_fail in case silly HTTP status numbers are set, and test that.   Poul-Henning Kamp
• [master] a7ceb25 Support IP + STRING   Poul-Henning Kamp
• [master] d25ce0b Rather than add to a never ending list of things we cannot add/subtract, examine the list of things we can handle, and if not, and we are looking for strings anyway, convert to string right away.   Poul-Henning Kamp
• [master] 5337a9a Finally nail the race which made H2 tests randomly "just not work":   Poul-Henning Kamp
• [master] 175e711 Version numbers are unsigned   Poul-Henning Kamp
• [master] 4f03022 Avoid sideeffects in assert   Poul-Henning Kamp
• [master] d76e051 Plug a harmless resource leak to silence Coverity   Poul-Henning Kamp
• [master] e914a75 CLI errors are not fatal if we are in shutdown   Poul-Henning Kamp
• [master] 6ab7447 This assert is no longer valid   Poul-Henning Kamp
• [master] b722682 Make it clear to coverity that we are not doing anything untoward with the frame length.   Poul-Henning Kamp
• [master] c2b0624 Clarify that usage() doesn't return for Coverity.   Poul-Henning Kamp
• [master] e2090b4 This cannot overflow the buffer, but Coverity can't guess that.   Poul-Henning Kamp
• [master] c1cf8d3 Turn the snapshot tokens from WS into uintptr_t to make it very clear that you are not supposed to use them as pointers.   Poul-Henning Kamp
• [master] 92e77b8 Introduce WS_Inside() to reduce the amount of grubbing around inside struct ws.   Poul-Henning Kamp
• [master] 4e0004e Eliminate some grubbing around inside struct ws   Poul-Henning Kamp
• [master] 6440798 Introduce WS_ReserveLumps() to allocate an array of something.   Poul-Henning Kamp
• [master] a6696fc Fix an off-by one causing extremely rare asserts: We must have two free slots, one for the data we want to write and one for the end of chunk bits.   Poul-Henning Kamp
• [master] 77cab36 Start hammering away on H2 again   Poul-Henning Kamp
• [master] 3f46e4f Add a field which says if H2 error is stream and/or connection error.   Poul-Henning Kamp
• [master] a62435b Introduce a selfdescribing h2_error struct   Poul-Henning Kamp
• [master] 7144af0 More compact H2 error names   Poul-Henning Kamp
• [master] 7a834f4 Have frame-rx functions return an error   Poul-Henning Kamp
• [master] b167b52 Use table based dispatch for H2 rx   Poul-Henning Kamp
• [master] 9a13765 Start working on systematic error handling   Poul-Henning Kamp
• [master] bdf117b More error-handling work   Poul-Henning Kamp
• [master] d8efa56 Unify sendhex and fix at least one bug therein.   Poul-Henning Kamp
• [master] 9d4797a Allow use of symbolic H2 error names in expect rst.err == BLA   Poul-Henning Kamp
• [master] 142e918 Send RST_STREAM and GOAWAY, and add a testcase to see that it works.   Poul-Henning Kamp
• [master] e223671 Handle NULL pointer structs with pan_already()   Poul-Henning Kamp
• [master] 5455830 More H2 coverage   Poul-Henning Kamp
• [master] 4dbe1de Try to ensure the core dump happens before the test ends   Poul-Henning Kamp
• [master] 614156b Don't dump core in files testing the panic code, to steals our gcov numbers.   Poul-Henning Kamp
• [master] 656982a Some experiemental code to deconfigure a thread pool.   Poul-Henning Kamp
• [master] f8ea87c Forgot this in previous commit.   Poul-Henning Kamp
• [master] 6db7d75 Use closefd() throughout, take 2   Poul-Henning Kamp
• [master] 5f0bc82 Make this test less sensitive to scheduling order.   Poul-Henning Kamp
• [master] b00cbb2 Allow user defined sub{}'s to return without specifing a handling.   Poul-Henning Kamp
• [master] cbe8047 Reintroduce hit-for-pass with new and better syntax:   Poul-Henning Kamp
• [master] 09f96b5 Add a couple of (needless) asserts for Coveritys sake.   Poul-Henning Kamp
• [master] d8cadd6 Ignore HUB and PIPE signals before starting child   Poul-Henning Kamp
• [master] 436426e Also add more delay for the first varnishd   Poul-Henning Kamp
• [master] 9577a7a Add missing newline, move thread/errno info after backtrace.   Poul-Henning Kamp
• [master] 7388e70 Point out upper-case characters, I'm sure this will be a common mistake.   Poul-Henning Kamp
• [master] 1b7e1d6 Forgot this part of last commit.   Poul-Henning Kamp
• [master] 0690efa Remove duplicated asserts   Poul-Henning Kamp
• [master] 987d090 Ignore upper/lower case when comparing header names   Poul-Henning Kamp
• [master] b4a959a Make a hole through H2 POST.   Poul-Henning Kamp
• [master] a8b56df Add a debug option to slow down acceptor threads so that we can stabilize test c00080.   Poul-Henning Kamp
• [master] 09df260 Allow and ignore a leading '-' on cli commands.   Poul-Henning Kamp
• [master] 9e27227 Implement varnishd -I <cli_file>   Poul-Henning Kamp
• [master] 7c43a11 Add test case for -I cli_file   Poul-Henning Kamp
• [master] f885344 Use different workdirs to avoid old SHMfiles confusing varnishadm   Poul-Henning Kamp
• [master] 42dc406 Don't start varnishd's with shell if we can avoid it, the required delays are execessive (The varnish testdriver already has code to deal with this)   Poul-Henning Kamp
• [master] 1910640 Encapsulate Expect: and Connection: processing in a transport agnostic function.   Poul-Henning Kamp
• [master] 876d42b Make sure the req always has a transport   Poul-Henning Kamp
• [master] 7910232 Reset the headers when we get a HEADERS frame   Poul-Henning Kamp
• [master] bdd84ff Make 100-continuation work in H2   Poul-Henning Kamp
• [master] bd0ce36 Silence flexelint a bit   Poul-Henning Kamp
• [master] 1289746 Tighten asserts related to pipelined rxbuf.   Poul-Henning Kamp
• [master] 9d73cd1 Make param::nuke_limit a total count of nukes allowed for each object creation.   Poul-Henning Kamp
• [master] 0b7b8e7 Comment out premature assert, I had forgotten the XXX over in H2   Poul-Henning Kamp
• [master] f146f86 Start remodelling the argument parsing (again) so that special mode arguments (-V, -x and later -C) break free early.   Poul-Henning Kamp
• [master] ac9a479 Renovate varnishd's usage message, and don't emit it until asked for it.   Poul-Henning Kamp
• [master] b2f32cd Try to get http->protover into some kind of consistent state so we do not send HTTP/2.0 headers to backends on pass from HTTP/2.0 clients   Poul-Henning Kamp
• [master] 213939a Don't issue Unset VSL records for the three top-line fields.   Poul-Henning Kamp
• [master] 9c4e8c8 Make vdir_remove_backend return void   Guillaume Quintard
• [master] 2f05442 Let remove_backend tell if the current dir changed   Guillaume Quintard
• [master] c5966cb Add optional sticky arg to fb director   Guillaume Quintard
• [master] af3657e Sticky fallback directors vtc   Guillaume Quintard
• [master] ddb2125 Rename fallback variables   Guillaume Quintard
• [master] c9fcdf2 Create new changelog section   Guillaume Quintard
• [master] d7c6036 Start inactive server   Guillaume Quintard
• [master] 9a8279b Assert on the right variable   Guillaume Quintard
• [master] 607e189 -some must have a non-zero arg   Guillaume Quintard
• [master] eaca897 Check pthread_cond_init's return   Guillaume Quintard
• [master] 82ad97a Make sure we ask for at least one frame   Guillaume Quintard
• [master] 8e4be28 Remove useless memcpy   Guillaume Quintard
• [master] 13716ff Silent coverity   Guillaume Quintard
• [master] d1d1b78 calloc instead malloc to calm coverity   Guillaume Quintard
• [master] 6017d5f Fix typo   Guillaume Quintard
• [master] 280da2d times are unsigned long, not int   Guillaume Quintard
• [master] 14be194 Use do/while to reassure coverity   Guillaume Quintard
• [master] 06108a1 Test what logexpect can't   Federico Schwindt
• [master] ca42915 Always call va_end()   Federico G. Schwindt
• [master] a7ceb25 Support IP + STRING   Federico G. Schwindt
• [master] f4b367e Ensure VFP is correctly set when unbusy'ing   Federico G. Schwindt
• [master] 95d8b34 Document VRT_ipcmp wrt bumping VRT API version   Federico G. Schwindt
• [master] 8dfa782 Dump loaded VMODs and related details under panic   Federico G. Schwindt
• [master] b907f30 More consistent panic output   Federico G. Schwindt
• [master] f0085ce Make this test less time sensitive   Federico G. Schwindt
• [master] 9ab3376 More version numbers are unsigned   Federico G. Schwindt
• [master] 929254c Polish   Federico G. Schwindt
• [master] baee057 Consistency   Federico G. Schwindt
• [master] e012464 Whitespace   Federico G. Schwindt
• [master] 89619ad Instead of exiting tell the loop we are quitting   Federico G. Schwindt
• [master] bb79466 Remove dead code   Federico G. Schwindt
• [master] 8723d4f Add missing header   Federico G. Schwindt
• [master] 950439f Correct prototype   Federico G. Schwindt
• [master] 6bc3cb8 Uneeded returns are unneeded   Federico G. Schwindt
• [master] 20c5f37 Polish   Federico G. Schwindt
• [master] f9348e4 Whitespace OCD   Federico G. Schwindt
• [master] bfaa16e Simplify   Federico G. Schwindt
• [master] 8f22397 Correct check when parsing the query string.   Federico G. Schwindt
• [master] 2285711 Accommodate for 32 bits archs   Federico G. Schwindt
• varnish-commit：人力资源效能方程式 45743   ks
• varnish-commit：身经百战的她为什么会失败？ 13962   ud
• varnish-commit:如何处理员工违纪问题? qexgmj   沃冰松

Last message date: Tue Feb 28 23:31:05 UTC 2017
Archived on: Wed Feb 11 18:06:45 UTC 2026

• Messages sorted by: [ thread ] [ subject ] [ date ]
• More info on this list...