While I like the addition of this feature, I have concerns as it breaks the convention of re_extract where it is expected to return a string. It might be better to change this into a new function (re_extract_hash, maybe?) and then have it always return that indexed container?
At that point you could always have it return that format and be able to adjust the number of expected matches that return in that format as well since you don't have to leave the existing format in place. Sent from my HTC Inspire™ 4G on AT&T ----- Reply message ----- From: "Pavel Levshin" <[email protected]> To: "[email protected]" <[email protected]> Subject: [rsyslog] re_extract to give out all submatches Date: Fri, Oct 25, 2013 1:17 am Something I did not noticed yesterday. There is, very likely, a memory leak here in re_extract: grammar/rainerscript.c: if(bMustFree) free(str); if(r[0].datatype == 'S') es_deleteStr(r[0].d.estr); if(r[2].datatype == 'S') es_deleteStr(r[2].d.estr); if(r[3].datatype == 'S') es_deleteStr(r[3].d.estr); finalize_it: if(bHadNoMatch) { Shouldn't the memory be free when finalizing after an unsuccessful match? But then, it turns out applicable to most functions: What if there is an argument which is a JSON object? It looks like it will leak every time. There should be some varFree() function for this. And, at last, what if r[0].datatype is unitialized and equals 'S' by accident? Shouldn't there be initialization for every struct var? -- Pavel Levshin 25.10.2013 1:18, Pavel Levshin: > Hello. > > I hope it will be useful. Now, when local variables are real JSON > containers, why don't extract all submatches at once? I've given it a > try and it works. The patch is based on a current master-var-refactor, > I am sorry for this mess. I will redo it if needed when the > refactoring will be done. > > I've extended re_extract to achieve this behaviour. Creation of a new > function (re_extract_all) just for this would lead to code > duplication. The modified function returns a container with all > submatches on request. Submatches in the container are identified by > index, where "0" is whole match. New behaviour is achieved by setting > 'submatch' parameter to any negative number. For non-negative numbers, > it works as it used to. Example: > > # re_extract(expr, re, match, submatch, no-found) > > set $.str = "Testing 31 32 33 and so on"; > set $.var = re_extract($.str, "^([[:alnum:]]+) ([0-9]+) ([0-9]+)", > 0, -1, "nothing"); > set $!matches = $.var; # container here > set $!match2 = $.var!2; # 2nd submatch > > And here is the output of all these variables (with some magic template): > > _local_ { "str": "Testing 31 32 33 and so on", "var": { "0": "Testing > 31 32", "1": "Testing", "2": "31", "3": "32" } } > _msg_ { "matches": { "0": "Testing 31 32", "1": "Testing", "2": "31", > "3": "32" }, "match2": "31" } > > > By the way, this patch also fixes a bug with buffer overflow in > re_extract. > > > -- > Pavel Levshin > > > > _______________________________________________ > rsyslog mailing list > http://lists.adiscon.net/mailman/listinfo/rsyslog > http://www.rsyslog.com/professional-services/ > What's up with rsyslog? Follow https://twitter.com/rgerhards > NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad of > sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you DON'T > LIKE THAT. _______________________________________________ rsyslog mailing list http://lists.adiscon.net/mailman/listinfo/rsyslog http://www.rsyslog.com/professional-services/ What's up with rsyslog? Follow https://twitter.com/rgerhards NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad of sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you DON'T LIKE THAT. _______________________________________________ rsyslog mailing list http://lists.adiscon.net/mailman/listinfo/rsyslog http://www.rsyslog.com/professional-services/ What's up with rsyslog? Follow https://twitter.com/rgerhards NOTE WELL: This is a PUBLIC mailing list, posts are ARCHIVED by a myriad of sites beyond our control. PLEASE UNSUBSCRIBE and DO NOT POST if you DON'T LIKE THAT.
