Timeline
07/05/10:
- 23:51 Changeset [5014] by
- Start massaging the conditional parsing to handle non-variable left sides.
- 23:01 Changeset [5013] by
- Add a very rudimentary symbol table, and use it to find our variables.
- 22:56 Changeset [5012] by
- Add missing final newline.
- 22:14 Changeset [5011] by
- Move the variable set from being an argument to FindVar? to being a …
- 13:56 Changeset [5010] by
- generate section 1 man pages from rst fix some rst2man rendering issues
- 13:56 Changeset [5009] by
- generate vcl.7 from rst, include default.vcl in vcl.rst
- 12:10 Changeset [5008] by
- Improve usage restriction messages a little bit.
- 11:37 Changeset [5007] by
- Polishing
- 10:27 Changeset [5006] by
- Add a bitmap of allowed actions for the returns macro
07/04/10:
- 23:44 Changeset [5005] by
- Collapse the "VAR" token into ID.
- 22:07 Changeset [5004] by
- Comment out parse functions for SIZE type.
- 22:06 Changeset [5003] by
- Comment out unused SIZE type
- 21:55 Changeset [5002] by
- Various cleanups.
- 16:40 Changeset [5001] by
- Convert the "req.hash +=" pseudo variable handing of the hask-key build …
- 16:15 Changeset [5000] by
- The Tcl-Py conversion left out code to generate vrt_obj.h, add that. …
- 14:49 Changeset [4999] by
- In preparation for generic types functions, including conversion …
07/03/10:
- 12:23 Changeset [4998] by
- Some boilerplate stuff for VMOD's.
- 11:59 Changeset [4997] by
- FLexelinting
- 11:43 Changeset [4996] by
- prerequisites for installation, cache sizing guidelines + some general …
- 11:04 Changeset [4995] by
- Move the dlopen/dlsym/dlclose check of newly compiled VCL code to a sub …
07/01/10:
- 19:39 Ticket #731 (Can't remove illegal header without terminal colon) created by
- I needed to find a server-side solution for a broken client which …
06/30/10:
- 23:24 Ticket #730 (varnishd/cache_fetch.c:FetchBody dropping Content-Length on HEAD) created by
- Above is more or less it. Run a HEAD request against any webserver, and …
- 14:55 Changeset [4994] by
- Make vcl_dir work for vcl.load and include filenames.
06/29/10:
- 14:29 Changeset [4993] by
- Change API to VCC to be handle/object based.
- 14:19 Changeset [4992] by
- more on vcl: pipe, deliver, ACLs, purging
- 13:15 Changeset [4991] by
- Rename "struct tokenlist" to "struct vcc", it is much more than a simple …
- 12:01 Changeset [4990] by
- Add a vcl_dir parameter to resolve relative VCL pathnames from …
- 11:26 Ticket #723 (Incorrect SVN checkout URL in documentation) closed by
- fixed: Wiki is OK.
06/28/10:
- 16:28 Changeset [4989] by
- Fix a problem with accounting of ESI transactions, in particular make the …
- 13:57 Ticket #715 (CLI communication error (hdr) on vcl.discard) closed by
- fixed: Fixed under #722, see that. Will be fixed in 2.1.3
- 13:55 Ticket #723 (Incorrect SVN checkout URL in documentation) reopened by
- Oops, didn't mean to close this. Per: check other references (wiki ?)
- 13:53 Ticket #723 (Incorrect SVN checkout URL in documentation) closed by
- fixed: (In [4988]) Fix SVN urls. Fixes: #723
- 13:53 Changeset [4988] by
- Fix SVN urls. Fixes: #723
- 13:49 Ticket #725 (Assert error in _vsb_assert_integrity(), vsb.c line 84) closed by
- fixed: It looks like you are using synth{} in vcl_deliver{}, that is not …
- 13:47 Ticket #724 (Panic message: Assert error in VRT_synth_page(), cache_vrt.c line 918) closed by
- worksforme: I'm guessing you are using "synth {...}" in vcl_fetch{}, that's not legal, …
- 13:40 Ticket #728 (Assert error in vrt_selecthttp) closed by
- worksforme: I can make no heads or tails of that backtrace under the assumption that …
- 13:38 Ticket #726 (Assert error in http_SetH(), cache_http.c line 595) closed by
- fixed: It is a bug that you can set obj.http* in vcl_hit, please don't do that. …
- 10:26 Changeset [4987] by
- Explain the security-barriers influence on Varnish' design.
- 09:46 Ticket #711 (64bit Catch 22 on Solaris (gcc _builtin_xxx functions)) closed by
- invalid: Thanks a lot! That was the kind of issue I would not haven gotten around …
- 09:45 TroubleLog edited by
- (diff)
- 09:40 Ticket #729 (Varnish loosing all its cache content) closed by
- worksforme: Hi Chris, Those messages (not responding to ping, killing it.) are indeed …
06/23/10:
- 14:23 Ticket #729 (Varnish loosing all its cache content) created by
- We are using Varnish (on 2 servers) to cache our images coming from Apache …
- 11:41 Changeset [4986] by
- Add a code coverage test for the classic hasher, now that critbit is the …
06/22/10:
- 21:28 Ticket #728 (Assert error in vrt_selecthttp) created by
- Using latest 2.1.2 RPMs on RHEL 5.5 Using the default "default.vcl" and …
- 15:36 Changeset [4985] by
- Make it possible to declare stand-alone backend probes and reference them …
- 15:01 Changeset [4984] by
- Emit probe-specs as separate free-standing structs, and point to them from …
- 14:22 Changeset [4983] by
- Add testcase to check that backends named "default" get used as such.
- 14:10 Changeset [4982] by
- If there is a backend or director named "default", make that the default, …
- 09:47 Ticket #695 (Varnish trunk 4516 segmentation fault in FreeBSD) closed by
- fixed: This was fixed in r4547 in another context.
- 09:37 Ticket #681 (problems using regular expressions with varnishlog) closed by
- fixed: (In [4981]) Typo in -X matching in varnishapi Fixes: #681
- 09:37 Changeset [4981] by
- Typo in -X matching in varnishapi Fixes: #681
- 09:25 Ticket #714 (VRT_l_* calls cause memory exhaustion under high load) closed by
- worksforme: It does not when I try it.
- 09:24 Ticket #717 (obj.hash and beresp.hash do not work inside vcl_deliver) closed by
- worksforme: The hash string is no longer available for reading, it was a waste of …
- 09:23 Ticket #718 (obj.hits or beresp.hits not available in vcl_deliver) closed by
- worksforme: obj.hits is available in vcl_deliver{} and vcl_hits{}
- 09:15 Ticket #709 (After SessionClose c EOF there is a second ReqEnd c 0) closed by
- fixed: (In [4980]) Emit Length for client side right before ReqEnd?, to summarize …
- 09:15 Ticket #720 (esi confuses varnishncsa) closed by
- fixed: (In [4980]) Emit Length for client side right before ReqEnd?, to summarize …
- 09:15 Changeset [4980] by
- Emit Length for client side right before ReqEnd?, to summarize ESI …
06/21/10:
- 17:33 Changeset [4979] by
- Emit a Length record on fetches from the backend.
- 15:53 Ticket #727 (Tiny Patch: Human readable nagios output) created by
- as mentioned in #710, I suggest a little change to make big number easier …
- 14:31 Changeset [4978] by
- bereq.connect_timeout should be available in vcl_pipe
- 10:44 Changeset [4977] by
- Check FreeBSD_version for memory barrier definition
- 10:16 Changeset [4976] by
- Silence a bogus pthread warning
- 10:15 Ticket #719 (Panic message: Missing errorhandling code in parse_esi_tag(), cache_esi.c ...) closed by
- fixed: (In [4975]) Fix a bug when ESI elements span storage elements, which only …
- 10:15 Changeset [4975] by
- Fix a bug when ESI elements span storage elements, which only the …
- 10:01 Changeset [4974] by
- Minor nitpicking
- 09:48 Changeset [4973] by
- Use WSP instead of VSL for ordering and performance
06/19/10:
- 18:51 Changeset [4972] by
- Too many bogus SVID versions out there, don't assert set_curs() return.
- 00:50 Installation edited by
- Adding my recepie from Solaris10 (diff)
06/18/10:
- 11:32 TroubleLog edited by
- (diff)
- 11:12 Changeset [4971] by
- Use ExpectErr?() instead of Expect, so that we get compiler error messages …
- 10:46 Ticket #726 (Assert error in http_SetH(), cache_http.c line 595) created by
- Given simple VCL like that: […] Im getting panic message (varnishd …
06/17/10:
- 16:14 Ticket #725 (Assert error in _vsb_assert_integrity(), vsb.c line 84) created by
- I am getting two different panic messages when try to test the performance …
- 14:17 Ticket #724 (Panic message: Assert error in VRT_synth_page(), cache_vrt.c line 918) created by
- We are getting errors like this from time to time. I cannot reproduce them …
- 13:50 Ticket #723 (Incorrect SVN checkout URL in documentation) created by
- svn co http://varnish-cache.org/svn/varnish/trunk svn: URL …
- 13:06 VCLExampleSaintMode edited by
- removed bogus req. (diff)
- 10:47 Changeset [4970] by
- Rename main counter struct from VSL to VSC namespace
- 10:46 Changeset [4969] by
- FlexeLint? inspired garbage collection
- 10:45 Changeset [4968] by
- Destroy directors before their embedded backends.
- 00:15 Ticket #722 (Director cleanup issue on vcl.discard) closed by
- fixed: (In [4967]) Fix a problem in director teardown at vcl.discard time: We …
- 00:15 Changeset [4967] by
- Fix a problem in director teardown at vcl.discard time: We didn't …
06/16/10:
- 23:18 Ticket #722 (Director cleanup issue on vcl.discard) created by
- We don't clean up directors and the backends they depend on in the right …
- 22:23 Changeset [4966] by
- Introduce the first per-backend stats counter: number of VCLs referencing …
- 14:47 Changeset [4965] by
- Add mark/clean facility to VSM, so that all dynamic allocations gets …
- 14:25 Changeset [4964] by
- Better error handling, limit how long time we wait for initialization to …
- 13:16 Changeset [4963] by
- Add a state field for deletes.
- 12:40 Changeset [4962] by
- Make varnishstat discover dynamic allocations in curses mode.
- 12:39 Changeset [4961] by
- Move SMA stats allocation to child process
- 12:39 Changeset [4960] by
- Preincrement seq#, otherwise it makes no difference.
- 12:19 Changeset [4959] by
- Move VSM allocation to common code
- 11:51 Changeset [4958] by
- VSM allocations/frees need to happen from both manager and child process …
- 10:44 Changeset [4957] by
- Add a comment about desirable but complex error recovery
06/15/10:
- 13:38 Ticket #721 (hash and client director docs.) created by
- Document the "hash" and "client" directors in the VCL man page.
- 13:11 Ticket #720 (esi confuses varnishncsa) created by
- Varnishlog reports Length multiple times for an esi-processed request, …
- 10:16 Ticket #719 (Panic message: Missing errorhandling code in parse_esi_tag(), cache_esi.c ...) created by
- We are getting the following (at the bottom of the mail) error on an ESI …
06/14/10:
- 11:35 Changeset [4956] by
- Pull in srandomdev() compat if needed.
- 11:17 Changeset [4955] by
- Make VSL_Dispatch reopen the vsm when idle-timeout kicks in.
- 10:40 Changeset [4954] by
- Remove the -L option stuff, it will not be needed for docs.
- 10:03 Changeset [4953] by
- Cleanup around arg processing. Need to think a bit more about usage …
- 09:41 Changeset [4952] by
- Rename vsm stuff from vsl to vsm for consistency
- 09:32 Ticket #716 (varnishstat assert error in do_curses()) closed by
- fixed: This should be fixed in r4951
- 09:31 Changeset [4951] by
- Don't check return value of curs_set(), appearantly the old SVID bug is …
- 09:24 Ticket #711 (64bit Catch 22 on Solaris (gcc _builtin_xxx functions)) reopened by
06/12/10:
- 10:29 Changeset [4950] by
- syntax error
- 10:26 Changeset [4949] by
- advanced topics pointed out + some minor fixes
06/11/10:
- 11:07 Changeset [4948] by
- Bans via VCL/HTTP
06/10/10:
- 23:19 Changeset [4947] by
- Some text on purging and bans. I guess we would need to rename url.purge …
- 19:03 Changeset [4946] by
- man varnishd was missing the -C option
- 16:13 Ticket #718 (obj.hits or beresp.hits not available in vcl_deliver) created by
- # add hit or miss in header if (beresp.hits > 0) { set …
- 12:51 Ticket #717 (obj.hash and beresp.hash do not work inside vcl_deliver) created by
- Message from VCC-compiler: Unknown variable 'beresp.hash' At: (input Line …
06/09/10:
- 17:40 Ticket #716 (varnishstat assert error in do_curses()) created by
- Whenever I try to run varnishstat and assert error is thrown saying: …
- 16:02 Changeset [4945] by
- Add a couple of missing includes
- 14:13 Changeset [4944] by
- Another necessary <stdint.h>
- 14:08 Changeset [4943] by
- Include stdint.h in varnishapi.h, that's easier.
- 13:28 Changeset [4942] by
- Add necessary include
- 11:37 Changeset [4941] by
- Hmm, not sure SVN and I fully agree on "svn cp" and "svn mv" commands. …
- 11:34 Changeset [4940] by
- Delaminate the VSL api from the VSM api
06/08/10:
- 14:07 Changeset [4939] by
- FlexeLint? VSC api
- 13:43 Changeset [4938] by
- Document VSC api
- 13:34 Changeset [4937] by
- Redo the VSC api new-style.
- 12:35 Changeset [4936] by
- One more vsl->vsc rename
- 12:31 Changeset [4935] by
- More wanton renaming, this time varnishapi::VSC
- 12:19 Changeset [4934] by
- rename files to match subject matter
- 12:15 Changeset [4933] by
- Wanton renaming in libvarnishapi
- 12:01 Changeset [4932] by
- More wanton renaming
- 11:55 Changeset [4931] by
- Explicitly open the VSM
- 11:49 Changeset [4930] by
- Not sure why this didn't get included in last commit
- 11:46 Changeset [4929] by
- Split and rename shmlog.h to vsl.h and vsc.h
- 11:23 Changeset [4928] by
- Move vin stuff into vin.h and vsm stuff into vsm.h
- 11:09 Changeset [4927] by
- Begin splitting shared memory log stuff into its three distinct parts: …
- 10:46 Changeset [4926] by
- Rename vsl.h to vslapi.h
- 10:42 Changeset [4925] by
- Begin a major round of polishing of the VarnishAPI to make it sensible …
06/07/10:
- 22:32 WikiStart edited by
- (diff)
- 14:56 Changeset [4924] by
- Tighten on the relative accesses to wrap-sequence and special marker …
- 14:37 Ticket #715 (CLI communication error (hdr) on vcl.discard) created by
- When i try to delete old configs via vcl.discard I get an "CLI …
- 10:45 WikiStart edited by
- (diff)
- 10:43 WikiStart edited by
- (diff)
- 10:38 Changeset [4923] by
- lame typo
- 10:35 Changeset [4922] by
- intro + some layout
- 10:19 Changeset [4921] by
- Forgot to add flexelint files for varnishstat
- 10:12 Changeset [4920] by
- Whitespace cleanup
- 10:02 Changeset [4919] by
- More Id expansion
- 09:59 Changeset [4918] by
- Set Id expansion
- 09:46 Changeset [4917] by
- Sanitize macros
- 00:31 Ticket #714 (VRT_l_* calls cause memory exhaustion under high load) created by
- Writing to the req.url variable from inside a C block, e.g. using …
06/06/10:
- 15:54 Changeset [4916] by
- FlexeLint? cleanup
- 15:39 Changeset [4915] by
- FlexeLint? cleanup
- 15:14 Changeset [4914] by
- Clean up and unify shmlog writing in the worker process. Always wrap the …
- 14:39 Changeset [4913] by
- First part of shmlog rewrite: Handle shmlog in 32bit byte aligned words, …
- 11:40 Changeset [4912] by
- Revisit the memory barriers, and try to catch some common platforms with …
- 00:41 Changeset [4911] by
- Move the varnishstat -f argument into varnishapi, it is broadly applicable …
06/05/10:
- 22:50 Changeset [4910] by
- Split argv.h into its own header
- 22:17 Changeset [4909] by
- Split VSL_Arg() into VSL_Log_Arg() and VSL_Stat_Arg()
- 17:14 Changeset [4908] by
- More flexelinting
- 16:53 Changeset [4907] by
- Convert varnishtest to handle dynamic stats counters
- 16:33 Changeset [4906] by
- Flexelint polishing
- 16:24 Changeset [4905] by
- More flexelint cleanup
- 15:30 Changeset [4904] by
- More .lnt centralizing
- 15:21 Changeset [4903] by
- More FlexeLint? centralization
- 15:20 Changeset [4902] by
- Oops, missing {...}
- 15:03 Changeset [4901] by
- Start unifying FlexeLint? instructions, to curb spreading copy&paste-itis
- 14:51 Changeset [4900] by
- Just an XXX observation
- 14:48 Changeset [4899] by
- polish
- 14:40 Changeset [4898] by
- Wrap all the details in a struct for the VSL_IterStat() callback, it is …
- 13:05 Changeset [4897] by
- Teach varnishstat (curses mode) to reopen shmlog if varnishd master is …
- 13:04 Changeset [4896] by
- Add VSL_ReOpen() to detect shmfile changes (ie: master restarts)
- 12:16 Changeset [4895] by
- Rewrite curses support to use VSL_IterStat()
- 12:04 Changeset [4894] by
- Oops! r4893 contained most of what should go in this commit: Rewrite -1 …
- 12:03 Changeset [4893] by
- Tighten prototype consting.
- 11:58 Changeset [4892] by
- Add a new function for walking over all stats counters, including …
- 10:57 Changeset [4891] by
- Split stats to its own source file
- 10:48 Changeset [4890] by
- add more todo notes
- 09:26 Changeset [4889] by
- style changes, headers, etc
Note: See TracTimeline
for information about the timeline view.
