| 00:05.40 | *** join/#asterisk-dev coreyfarrell (~coreyfarr@rrcs-66-57-14-122.midsouth.biz.rr.com) |
| 00:05.40 | *** mode/#asterisk-dev [+o coreyfarrell] by ChanServ |
| 07:36.04 | *** join/#asterisk-dev sjm_uk (~sjm_uk@cpc76962-warw17-2-0-cust593.3-2.cable.virginm.net) |
| 09:36.42 | *** join/#asterisk-dev coreyfarrell (~coreyfarr@rrcs-66-57-14-122.midsouth.biz.rr.com) |
| 09:36.42 | *** mode/#asterisk-dev [+o coreyfarrell] by ChanServ |
| 13:17.01 | *** join/#asterisk-dev coreyfarrell (~coreyfarr@rrcs-66-57-14-122.midsouth.biz.rr.com) |
| 13:17.01 | *** mode/#asterisk-dev [+o coreyfarrell] by ChanServ |
| 13:45.04 | *** join/#asterisk-dev bford (uid283514@gateway/web/irccloud.com/x-rmnaeicwanmyvfsm) |
| 13:45.04 | *** mode/#asterisk-dev [+o bford] by ChanServ |
| 14:13.32 | *** join/#asterisk-dev kharwell (uid358942@gateway/web/irccloud.com/x-acxlpbpdmdjkfeqf) |
| 14:13.32 | *** mode/#asterisk-dev [+o kharwell] by ChanServ |
| 14:19.36 | *** join/#asterisk-dev teitoklien (~teitoklie@2401:4900:314d:5ef9:7831:a3fe:4482:4faf) |
| 14:20.34 | *** part/#asterisk-dev teitoklien (~teitoklie@2401:4900:314d:5ef9:7831:a3fe:4482:4faf) |
| 15:59.03 | *** join/#asterisk-dev Dovid (~dovid@ool-45738ae3.dyn.optonline.net) |
| 16:00.13 | Dovid | hi. I don't know much c. I am looking at funcs/func_curl.c and I want to catch where a response from a web server is a 404. Currently if port 80 is closed on a server it gets caught on line 705 under |
| 16:00.50 | Dovid | hi. I don't know much c. I am looking at funcs/func_curl.c and I want to catch where a response from a web server is a 404. Currently if port 80 is closed on a server it gets caught on line 705 under if "(curl_easy_perform(*curl) != 0) {" can anyone clue me in how to check the response code that curl got? |
| 16:11.51 | gtjoseph | <PROTECTED> |
| 16:12.02 | gtjoseph | Here's a sample: https://curl.haxx.se/libcurl/c/CURLINFO_RESPONSE_CODE.html |
| 16:42.00 | *** join/#asterisk-dev AsteriskRoss (~AsteriskR@r01.nt-r1.nor.gb.voicehost.co.uk) |
| 17:01.52 | jkroon | file, gtjoseph - thanks for the +1's on my submissions. |
| 17:02.44 | gtjoseph | :) |
| 17:02.47 | gtjoseph | no problem |
| 17:02.53 | file | beep boop |
| 17:06.24 | Dovid | gtjoseph: TY. let me tr y that |
| 17:06.32 | gtjoseph | sure! |
| 17:23.25 | Dovid | gtjoseph: I am trying to get asterisk to fail if the curl response is a 404. i know a bit about coding but my knoweldge on C is limited. Looking at the curl code regardless of what happens in acf_curl_helper it seems (if its connection refused or if it gets a 404) We always get to return ret; at the end. The line before I added: "ast_log(LOG_WARNING, "DOVID: UPTO ret which is %d <<", ret); which outputs "acf_curl_helper: |
| 17:23.25 | Dovid | <PROTECTED> |
| 17:24.33 | gtjoseph | let me check |
| 17:27.31 | gtjoseph | We're checking for the return from curl_easy_perform which indicates if the call was "successful". |
| 17:27.48 | gtjoseph | that doesn't mean that the server returned a "200 OK" though. |
| 17:28.06 | gtjoseph | I think it just means that the server returned a valid response. |
| 17:29.18 | gtjoseph | You then have to do a curl_easy_getinfo to see what the HTTP result was. |
| 17:31.33 | gtjoseph | it's odd that that function doesn't actually do anything other than print an error if the call fails though. |
| 17:31.45 | Dovid | gtjoseph: I was just typing that now |
| 17:34.23 | Dovid | seems ret is now giving me a -1 or a 0. 0 seems to be if it gets any response. its about time I learn c. one sec |
| 17:35.29 | gtjoseph | Ha! |
| 17:38.35 | Dovid | seems like if we satisfy "if (args->cb_data.str && ast_str_strlen(args->cb_data.str)) {" then ret gets set to 0 so I need to check it there. would Digium accept a patch to consider it a failure if the response is a 404? |
| 17:39.13 | gtjoseph | At first glance, I don't see why not. |
| 17:39.25 | gtjoseph | I'm sure others will weigh in. |
| 17:40.31 | file | yeah, it would be good to get feedback... because failure can mean different things to different people in that case |
| 17:43.08 | gtjoseph | file: it _is_ kinda odd that we don't consider a failure of the curl_easy_perform to be an issue. :) |
| 17:43.24 | Dovid | file: Or perhaps a config that says to fail on a 404 but fething configs is way above my pay grade |
| 17:43.35 | file | gtjoseph: it is, the default return value is -1, it continues to do clean up afterwards and returns -1 in that case |
| 17:43.40 | file | only when there's data does it become 0 |
| 17:44.10 | gtjoseph | yeah i guess. |
| 17:45.02 | gtjoseph | in the case of a 404, i wonder what the contents of str are then. |
| 17:45.28 | gtjoseph | Dovid: you might want to check what's in "str" when there's a 404. |
| 17:46.04 | Dovid | https://pastebin.com/jqEUBmDi |
| 17:46.13 | Dovid | thats the response HOWEVER the http code response is 404. |
| 17:46.43 | gtjoseph | yep, that would be an issue. |
| 17:46.57 | Dovid | seems that if str !='' then we consider it a success |
| 17:47.07 | gtjoseph | yep |
| 17:47.38 | Dovid | file a bug report or whip up code that says if resposne is 404 to to set ret to -1 ? |
| 17:48.17 | gtjoseph | I'd do both. You can assign the issue to yourself |
| 17:48.32 | Dovid | I am not a developer by any means but I will try. |
| 17:48.58 | gtjoseph | we appreciate the effort and will help. |
| 17:49.26 | Dovid | I will open a ticket and try to write a patch. the question is which way is correct? is it it |
| 17:49.50 | Dovid | if(ret code !=200) then we have an issue or if(re code == 404) then we have an issue |
| 17:50.06 | gtjoseph | that's a good question. :) |
| 17:50.12 | seanbright | i think its a behavior change |
| 17:50.15 | seanbright | it's* |
| 17:50.34 | gtjoseph | we don't have a way to communicate the http response code back to the user. |
| 17:51.09 | Dovid | but in acf_curl_helper when should we set to ret to -1? if its a 404 or if its not a 200 |
| 17:51.53 | gtjoseph | I would say yes but continue to process and send back the text results. |
| 17:52.11 | gtjoseph | but as file and seanbright has said, it's a behavior change that we'd have to discuss. |
| 17:52.36 | Dovid | ok. let me open a ticket for now and I will let the powers that be decide. |
| 17:52.42 | *** join/#asterisk-dev CELYA_ (~Thunderbi@lfbn-nan-1-2297-47.w92-139.abo.wanadoo.fr) |
| 17:53.23 | gtjoseph | Go ahead and submit a patch to Gerrit as well. That will trigger discussion. |
| 17:53.46 | Dovid | I think a configurable option would be best with the default for it to be the way it is but thats above my head. |
| 17:54.26 | gtjoseph | another option would be to return the http error code in a channel variable |
| 17:55.30 | gtjoseph | well, i guess this is related to your realtime question on the users list so a channel vaiable isn't really going to help. |
| 17:56.13 | seanbright | i would prefer we store the last response code in a datastore for later retrieval by another dialplan function than setting a magic channel variable |
| 17:57.21 | Dovid | gtjoseph: yes it is. rather then have other people fix it for me I decided to dive in. |
| 17:58.07 | Dovid | seanbright: I understand your logic should it be where we are using res_func from the dialplan. however in this case its from real time + music on hold. my issue is since ret is 0 asterisk does not failover |
| 17:58.53 | seanbright | you're talking about res_config_curl? |
| 17:59.28 | Dovid | seanbright: in what regrard? using the it in the DB or asterisk realtime? |
| 17:59.40 | seanbright | you're talking about curl and realtime |
| 17:59.57 | seanbright | which is provided by res_config_curl.c |
| 18:00.01 | seanbright | unless i really don't understand |
| 18:00.20 | Dovid | seanbright: exconfig.conf which which is using I think res_curl |
| 18:00.31 | seanbright | it's using res_config_curl.c |
| 18:00.48 | seanbright | which makes this the 3rd time i have said that |
| 18:00.50 | seanbright | :) |
| 18:01.10 | Dovid | lol. I love using rt more and more (in select cases) |
| 18:01.28 | gtjoseph | it's kinda convoluted. :) |
| 18:01.49 | seanbright | so res_config_curl is "using dialplan" because it is awful |
| 18:02.03 | seanbright | res_config_curl does not use res_curl, it uses func_curl |
| 18:02.49 | gtjoseph | yeah. all res_curl does is run the curl init for the process. |
| 18:03.34 | seanbright | so this is what you do |
| 18:03.43 | seanbright | modify func_curl to provide a way to get the last response code |
| 18:03.49 | seanbright | via a dialplan function |
| 18:04.04 | seanbright | modify res_config_curl to call that dialplan function after it calls the CURL one |
| 18:04.13 | seanbright | act appropriately upon the results |
| 18:04.25 | Dovid | seanbright": How would that help me where where I am not using curl in the dialplan? |
| 18:04.48 | Dovid | real time is calling func_curl |
| 18:04.51 | gtjoseph | You are behind the scenes. You just don;t know it. :) |
| 18:05.01 | seanbright | stop getting hung up on the "dialplan" part of "dialplan function" |
| 18:05.22 | seanbright | https://github.com/asterisk/asterisk/blob/master/res/res_config_curl.c#L300 |
| 18:05.39 | seanbright | res_config_curl (the glue between cURL and realtime) calls the CURL dialplan function "manually" |
| 18:06.01 | Dovid | gtjospeh: I get what you are saying. when you say dialplan in my head I need to evaluate the variable afterwards. I guess we would set the lat response and then in the real time check that response and decide there if it's a failure? |
| 18:06.57 | gtjoseph | yeah in res_config_curl as seanbright said |
| 18:07.55 | Dovid | so I need to capture the response code (which I know I need for curl_easy_getinfo) and some how set it so I can look at it in res_config_curl. correct? |
| 18:08.04 | gtjoseph | yep |
| 18:08.18 | Dovid | so who can I throw money at for quick lesson in C? |
| 18:08.27 | gtjoseph | seanbright :) |
| 18:08.54 | Dovid | seanbright: I would offer a round of drinks but who knows if Asterison is even happening at this point :( |
| 18:09.25 | seanbright | so you offer money |
| 18:09.27 | seanbright | find out it's me |
| 18:09.31 | seanbright | offer drinks instead |
| 18:09.34 | seanbright | real nice |
| 18:09.38 | Dovid | nope. I would in addition :) |
| 18:09.42 | seanbright | i'm kidding |
| 18:09.48 | seanbright | but i can't help |
| 18:09.53 | seanbright | no time |
| 18:09.59 | Dovid | I was just going to say that. |
| 18:10.54 | Dovid | can anyone send me an example of how variables are passed to functions so I can crate one and then evaluate it? |
| 18:12.43 | gtjoseph | There's actually an example in func_curl, kind of... |
| 18:13.16 | gtjoseph | AST_THREADSTORAGE(thread_escapebuf); |
| 18:13.40 | seanbright | TLS FTL |
| 18:15.18 | Dovid | what does AST_THREADSTORAGE(thread_escapebuf); do ? that stores a variable for later? |
| 18:15.57 | gtjoseph | that creates a variable named thread_escapebuf that the same thread can access |
| 18:16.46 | Dovid | ok. so I would need to create my own, give it a default value and then evaulate it later? |
| 18:17.05 | gtjoseph | since the same thread would be calling CURL and then your new function... CURL_RESPONSE or something like it, youd have acxf_curl_helper store the response then create a new function in func_curlk that retrieves it. |
| 18:18.12 | gtjoseph | actually you could call your new function CURL_HTTP_RESPOSE_CODE that returns just the 200, 404, etc. |
| 18:18.23 | seanbright | i guess there are potential implcations if CURL() was called in a global context |
| 18:18.47 | gtjoseph | like from the main thread? |
| 18:19.03 | seanbright | wherever dialplan globals are called |
| 18:19.12 | seanbright | evaluated, rather. |
| 18:19.35 | gtjoseph | hmmm. |
| 18:19.42 | seanbright | sorry, i'm just thinking out loud. carry on. |
| 18:20.20 | gtjoseph | i think those are all evaluated on the channel's thread. |
| 18:20.38 | gtjoseph | but no channel. |
| 18:20.42 | gtjoseph | so let me think |
| 18:20.54 | Dovid | too much potential for me to mess this up. If anyone wants to write up a patch and I will pay let me know otherwise will wait to see what happens on jira |
| 18:21.07 | gtjoseph | :) fair enough. |
| 18:21.14 | seanbright | i think it would be fine to use TLS for it |
| 18:24.20 | gtjoseph | actually, maybe not. |
| 18:26.18 | gtjoseph | well, yeah. if there's no channel, pbx_builtin_setvar_helper which uses a global lock but it's still executed in the calling thread. |
| 18:29.36 | *** join/#asterisk-dev Corydon76 (~quassel@96.69.98.139) |
| 18:29.37 | *** mode/#asterisk-dev [+o Corydon76] by ChanServ |
| 19:27.38 | *** join/#asterisk-dev DivideBy0 (~DivideBy0@unaffiliated/divideby0x0) |
| 19:27.38 | *** mode/#asterisk-dev [+o DivideBy0] by ChanServ |