Commit Graph

160 Commits (f82c42a22e6a0883c8b51b269d3be9c50e4cd6be)

Author SHA1 Message Date
Russell Bryant f52c068605 Merged revisions 90967 via svnmerge from
18 years ago
Olle Johansson 25cbb792b9 (closes issue #11422)
18 years ago
Mark Michelson e3c122863d Removing a pointless check of option_debug
18 years ago
Mark Michelson 0b3ab17bfa Merged revisions 90059 via svnmerge from
18 years ago
Jason Parker aa0f63d65f Remove "old"-style CLI handler, since nothing uses it anymore.
18 years ago
Russell Bryant 880fb1ece9 Merged revisions 89893 via svnmerge from
18 years ago
Russell Bryant 79bc5ede5f Merged revisions 89839 via svnmerge from
18 years ago
Steve Murphy 1f792f1745 closes issue #11294; missed the conditional unlock of the contexts when the hash table is used instead; also, used the ast_free_ptr as advised.
18 years ago
Russell Bryant 21a22fa584 Merged revisions 89790 via svnmerge from
18 years ago
Russell Bryant b11f846e7e Merged revisions 89594 via svnmerge from
18 years ago
Steve Murphy 2ec4b57622 Thanks to pnlarsson for noting the spelling error in the cli commands. Also, added some verbage about the new algorithm to CHANGES.
18 years ago
Steve Murphy a63f6be669 closes issue #11363; where the pattern _20x. buried in an included context, didn't match 2012; There were a small set of problems to fix: 1. I needed NOT to score patterns unless you are at the end of the data string. 2. Capital N,X,Z and small n,x,z are OK in patterns. I canonicalize the patterns in the trie to caps. 3. When a pattern ends with dot or exclamation, CANMATCH/MATCHMORE should always report this pattern, no matter the length. With this commit, I also supplied the wish of Luigi, where the user can select which pattern matching algorithm to use, the old (legacy) pattern matcher, or the new, trie based matcher. The OLD matcher is the default. A new [general] section variable, extenpatternmatchnew, is added to the extensions.conf, and the example config has it set to false. If true, the new matcher is used. In all other respects, the context/exten structs are the same; the tries and hashtabs are formed, but in the new mode the tries are not used. A new CLI command 'dialplan set extenpatternmatch true/false' is provided to allow switching at run time. I beg users that are forced to return to the old matcher to please report the reason in the bug tracker. Measured the speed benefit of the new matcher against an impossibly large context with 10,000 extensions: the new matcher is 374 times faster.
18 years ago
Luigi Rizzo e0ff5fef5c remove a bunch of useless #include "options.h"
18 years ago
Steve Murphy 55e03ad681 closes issue #11290; the proposed patch was a good guess, and would solve the bug to some extent, but was really masking the real issue, that there were bad entries in the table. This fix removes the condition that the hashtab updates be done on exten removal only when the pattern_tree was present, which is silly. The operations that apply to the pattern tree are instead made conditional. Also, threw back in routines that kpfleming deleted because of probs in the 64-bit world. Tested on both 32 and 64-bit machines (compile). Tested the reload problem with over 20 reloads, and no problems. If you find more problems, please reopen 11290.
18 years ago
Mark Michelson 29ad24e37a I introduced a deadlock avoidance into 1.4, which I attempted to port to trunk as well.
18 years ago
Kevin P. Fleming 34ed13becd remove some debugging code that doesn't compile on 64-bit platforms
18 years ago
Steve Murphy 8e90425a6c A free in add_pri was ultimately the source of the grief we were having with parking. This set of changes fixes that problem, and introduces some more error messages, and puts debugs into ifdefs for what could be short-term usage. Txs to Terry W. for his help, guidance, and especially patience.
18 years ago
Luigi Rizzo a23c055c3d move asterisk/paths.h outside asterisk.h and into those files
18 years ago
Luigi Rizzo 915b97d300 move internal function declarations to include/asterisk/_private.h
18 years ago
Mark Michelson dda649ead2 Merged revisions 89457 via svnmerge from
18 years ago
Steve Murphy f62bf52e55 Get rid of some debug messages in pbx.c
18 years ago
Steve Murphy 4ea5eced10 These changes were made in response to niklas@tese.se's letter of 11-17-2007, where he had 20 and 201 in two different contexts, included in the same context. In that particular case, we were behaving the same as 1.4, but after experimenting, I quickly found that if 20 and 201 were in the same extension, 1.4 would return 201, and this code returns 20. These changes now enable the current code to replicate the behavior of 1.4 in respect to MATCHMORE in cases like this.
18 years ago
Luigi Rizzo 9335ace850 another bunch of include removals (errno.h and asterisk/logger.h)
18 years ago
Steve Murphy 406f6f2909 a correction to code involved in an extension removal
18 years ago
Steve Murphy 1a5cdeb138 a quick fix to pbx_dundi.c to make it so it will compile. Hope I did the right thing. And some additions to removal of extens to take care of hashtab pointers in all cases.
18 years ago
Steve Murphy 7953c2c46f My goodness, haven't handled an extension deletion. Add code to ast_context_remove_extension2() to remove an extension from the trie. Done by marking it deleted. The scoreboard won't update for it any more. Also, a couple of calls to insert hashtab had a spurious ->exten, which was removed.
18 years ago
Luigi Rizzo fdb7f7ba3d Start untangling header inclusion in a way that does not affect
18 years ago
Steve Murphy ab32ff2ad8 This corrects a hashtab removal, given a bad argument
18 years ago
Steve Murphy 542f29ea0e This fixes a problem with pattern ranges; and corrects a situation in res_features, where an extension would be created with the name Zap/51, as an example. THe / is bad because it would tend to mean that the 51 is to be cid matched.
18 years ago
Steve Murphy eee92d0881 Get rid of a previously missed ast_log call for debug, no longer nec.
18 years ago
Steve Murphy 0bad49b3dd Perhaps I went overboard on initializing things. I can remove unnecc. stuff later. A few bug fixes. Killing small bugs on the way to killing bigger ones. Removed locking on hashtabs; there's plenty of locks already being taken. A small bug in the root_tree hashtab compare func.
18 years ago
Steve Murphy 620e580ed6 Had trouble playing with parking; spent a long time trying to reason out MATCHMORE mode. made these updates and xfers on zaptel lines seem to work ok now
18 years ago
Steve Murphy 20d24e1027 Rescaled the weights of the patterns to give something more independent of pattern length; and make . less likely to win. Question: which should win for 14102241145-- _1xxxxxxx. or _XXXXXXXXXXX -- right now, the pure X pattern will win.
18 years ago
Steve Murphy 225c94d950 A further problem highlighted by 11233 has been resolved; a certain combination of patterns in a certain order, led to a malformed trie, due to a ptr not being initialized in the loop. Also, some tree printing prettifications.
18 years ago
Kevin P. Fleming a34bbab643 use simpler technique for removing known entries from lists
18 years ago
Steve Murphy fd1f6df25f This hopefully will fix the re-opened 11233. Hadn't covered the case of a context with no patterns. (blush)
18 years ago
Steve Murphy f0a7e86177 closes issue #11233 -- where some fine points in the algorithm to build the tree needed to be corrected. Many thanks for the test case, jtodd
18 years ago
Jason Parker 50879dcb1d Merged revisions 89194 via svnmerge from
18 years ago
Steve Murphy 98429d37b2 Based on a note in asterisk-dev by Brian Capouch, I determined I too agressive in not initializing arrays passed to pbx_substitute_variables_xxxx; I reviewed the code (again) and hopefully found every possible spot where substitute_variables is called conditionally, and made sure the char array involved was set to a null string.
18 years ago
Steve Murphy a897556f7f This is the perhaps the biggest, boldest, most daring change I've ever committed to trunk. Forgive me in advance any disruption this may cause, and please, report any problems via the bugtracker. The upside is that this can speed up large dialplans by 20 times (or more). Context, extension, and priority matching are all fairly constant-time searches. I introduce here my hashtables (hashtabs), and a regression for them. I would have used the ast_obj2 tables, but mine are resizeable, and don't need the object destruction capability. The hashtab stuff is well tested and stable. I introduce a data structure, a trie, for extension pattern matching, in which knowledge of all patterns is accumulated, and all matches can be found via a single traversal of the tree. This is per-context. The trie is formed on the first lookup attempt, and stored in the context for future lookups. Destruction routines are in place for hashtabs and the pattern match trie. You can see the contents of the pattern match trie by using the 'dialplan show' cli command when 'core set debug' has been done to put it in debug mode. The pattern tree traversal only traverses those parts of the tree that are interesting. It uses a scoreboard sort of approach to find the best match. The speed of the traversal is more a function of the length of the pattern than the number of patterns in the tree. The tree also contains the CID matching patterns. See the source code comments for details on how everything works. I believe the approach general enough that any issues that might come up involving fine points in the pattern matching algorithm, can be solved by just tweaking things. We shall see. The current pattern matcher is fairly involved, and replicating every nuance of it is difficult. If you find and report problems, I will try to resolve than as quickly as I can. The trie and hashtabs are added to the existing context and exten structs, and none of the old machinery has been removed for the sake of the multitude of functions that use them. In the future, we can (maybe) weed out the linked lists and save some space.
18 years ago
Kevin P. Fleming edc78d6023 improve linked-list macros in two ways:
18 years ago
Joshua Colp f440f8ba48 Fix up some PBX logic that became broken. The code would exit prematurely when it should have been collecting more digits.
18 years ago
Mark Michelson 5a4867543d "show application <foo>" changes for clarity.
18 years ago
Russell Bryant 9bfd82cec2 Merged revisions 88805 via svnmerge from
18 years ago
Luigi Rizzo 08b10da53b Simplify the implementation and the API for stringfields;
18 years ago
Tilghman Lesher e8c781b215 Add pbx_lua as a method of doing extensions
18 years ago
Tilghman Lesher b5c2bcb169 'h' extension doesn't execute past first priority
18 years ago
Tilghman Lesher 90db1316ed Modify WaitExten to include an optional dialtone
18 years ago
Steve Murphy 63f2f04cf4 This commits the performance mods that give the priority processing engine in the pbx, a 25-30% speed boost. The two updates used, are, first, to merge the ast_exists_extension() and the ast_spawn_extension() where they are called sequentially in a loop in the code, into a slightly upgraded version of ast_spawn_extension(), with a few extra args; and, second, I modified the substitute_variables_helper_full, so it zeroes out the byte after the evaluated string instead of demanding you pre-zero the buffer; I also went thru the code and removed the code that zeroed this buffer before every call to the substitute_variables_helper_full. The first fix provides about a 9% speedup, and the second the rest. These figures come from the 'PIPS' benchmark I describe in blogs, conf. reports, etc.
18 years ago
Joshua Colp 284721fdb1 Drop any more references to type in the Exception dialplan function.
18 years ago