[00:55:41] (PS1) Yuvipanda: Catch OperationalError too [analytics/quarry/web] - https://gerrit.wikimedia.org/r/205798 [00:56:11] (PS2) Yuvipanda: Catch OperationalError too [analytics/quarry/web] - https://gerrit.wikimedia.org/r/205798 [00:56:19] (CR) Yuvipanda: [C: 2] Catch OperationalError too [analytics/quarry/web] - https://gerrit.wikimedia.org/r/205798 (owner: Yuvipanda) [00:56:24] (Merged) jenkins-bot: Catch OperationalError too [analytics/quarry/web] - https://gerrit.wikimedia.org/r/205798 (owner: Yuvipanda) [01:01:01] (PS1) Yuvipanda: Add title to query pages [analytics/quarry/web] - https://gerrit.wikimedia.org/r/205800 [01:01:37] (PS2) Yuvipanda: Add title to query pages [analytics/quarry/web] - https://gerrit.wikimedia.org/r/205800 [01:01:48] (CR) Yuvipanda: [C: 2] Add title to query pages [analytics/quarry/web] - https://gerrit.wikimedia.org/r/205800 (owner: Yuvipanda) [01:01:55] (Merged) jenkins-bot: Add title to query pages [analytics/quarry/web] - https://gerrit.wikimedia.org/r/205800 (owner: Yuvipanda) [08:08:58] Analytics-Tech-community-metrics, ECT-April-2015: Provide list of oldest open Gerrit changesets without code review - https://phabricator.wikimedia.org/T94035#1226997 (Qgil) Interesting! Let's fine tune this list. If there are a few false positives because of some Gerrit problem, I think this is ok. Stil... [08:09:30] Analytics-Tech-community-metrics, ECT-April-2015: Provide list of oldest open Gerrit changesets without code review - https://phabricator.wikimedia.org/T94035#1226999 (Qgil) [08:32:01] Analytics-Tech-community-metrics: https://www.openhub.net/p/mediawiki stats out of date - https://phabricator.wikimedia.org/T96819#1227012 (Qgil) [08:32:14] Analytics-Tech-community-metrics: https://www.openhub.net/p/mediawiki stats out of date - https://phabricator.wikimedia.org/T96819#1227014 (Qgil) p:Triage>Low [08:37:30] Analytics-Tech-community-metrics: https://www.openhub.net/p/mediawiki stats out of date - https://phabricator.wikimedia.org/T96819#1227018 (hashar) @Qgil you changed changed the git URL: ``` lang=diff - https://gerrit.wikimedia.org/r/mediawiki/core + https://git.wikimedia.org/git/mediawiki/core.git ``` I am... [08:47:32] Analytics-Tech-community-metrics: https://www.openhub.net/p/mediawiki stats out of date - https://phabricator.wikimedia.org/T96819#1227032 (Qgil) Huh, sorry and fixed. I was looking for a *.git URL because this is what they put in their example, but I didn't see I was changing from gerrit.* to git.* [10:37:04] (PS1) Joal: Patch pageview definition [analytics/refinery/source] - https://gerrit.wikimedia.org/r/205833 [11:39:00] Analytics-Tech-community-metrics, ECT-April-2015: "Who contributes code" page metrics are not updating - https://phabricator.wikimedia.org/T95166#1227262 (Dicortazar) [11:39:58] Analytics-Tech-community-metrics, ECT-April-2015: "Who contributes code" page metrics are not updating - https://phabricator.wikimedia.org/T95166#1182110 (Dicortazar) Kudos to @Acs. This should be fixed by now. Please have a look at it. [11:42:22] Analytics-Tech-community-metrics, Wikimedia-Git-or-Gerrit, ECT-April-2015: Active Gerrit users on a monthly basis - https://phabricator.wikimedia.org/T86152#962385 (Dicortazar) This panel is ready to go. Please have a look at it. Affiliations were added and bots removed as final steps [11:42:48] Analytics-Tech-community-metrics, Wikimedia-Git-or-Gerrit, ECT-April-2015: Basic metrics about contributors exercising +2/-2 permissions in Gerrit - https://phabricator.wikimedia.org/T59038#641085 (Dicortazar) This panel is ready to go (http://korma.wmflabs.org/browser/scr-contributors.html). Please... [11:43:49] halfak: you there ? [11:44:10] Analytics-Tech-community-metrics, ECT-April-2015: Migrate Korma identitites database to SortingHat - https://phabricator.wikimedia.org/T92953#1227278 (Dicortazar) This task was already done some days ago. Closing this. How do we know that identities were migrated to SortingHat? The big JSON file was prod... [11:44:19] Analytics-Tech-community-metrics, Wikimedia-Git-or-Gerrit, ECT-April-2015: Basic metrics about contributors exercising +2/-2 permissions in Gerrit - https://phabricator.wikimedia.org/T59038#1227280 (Dicortazar) [11:44:21] Analytics-Tech-community-metrics, ECT-April-2015: Migrate Korma identitites database to SortingHat - https://phabricator.wikimedia.org/T92953#1227279 (Dicortazar) Open>Resolved [11:44:23] Analytics-Tech-community-metrics, ECT-April-2015: Instructions to update user data in korma - https://phabricator.wikimedia.org/T88277#1227281 (Dicortazar) [12:34:25] o/ joal [13:21:12] Hi halfak, went for lunch :) [13:21:25] spark job finished in 15h roughly [13:21:48] With one specific task finishing 4hours after any other [13:21:49] joal, for Enwiki? [13:21:54] yessir [13:22:10] Cool! Yeah. There's some of the input files that are massively different than the others. [13:22:32] yeah, what that mean is that there is one page that contains an awful lot of refisions [13:22:43] OK. So now to run the test. The script I have for extracting tags should be 2.7 compatible. Let me check. [13:22:51] My guess is WP:AIV or WP:Sandbox. [13:23:03] https://en.wikipedia.org/wiki/Wikipedia:Administrator_intervention_against_vandalism [13:24:46] halfak: we definitelky should try ref extraction on simplewiki first, if you agree [13:25:10] joal, that'll work for me. [13:25:51] So, pageid = "id":5137507,"title":"Wikipedia:Administrators' noticeboard/Incidents" [13:26:00] 130GB of revisions :) [13:26:16] :) [13:26:23] Yeah. That's a big one. [13:26:34] The biggest we have :) [13:26:37] You should try https://en.wikipedia.org/wiki/Wikipedia:Sandbox [13:26:52] second biggest is 25gb [13:27:06] page_id = 16283969 [13:27:42] have not checked :) [13:27:53] kk [13:28:03] looking at file size :) [13:42:33] mornin' ottomata [13:42:42] Got some french trick for you ;) [13:43:04] ottomata: http://paste.rottenbytes.info/?3601c94f61e0b9d5#CY6Jgv7aqaDRIY+ADeQ8YNKnnAqR+XtDGBd7v5cTlMI= [13:43:35] ottomata: first line of IMPALA_SERVER_ARGS [13:43:43] should help [13:43:58] file location: /etc/impala/default [13:44:02] I htink [13:44:22] nuria: Hi, you there ? [13:44:49] ohhh boy [13:45:13] hahah yayyy, oh this is from your friend [13:45:18] and, it loks liek your friend is using puppet! [13:45:27] or, maybe chef, that probably uses erb templates too [13:45:32] thanks joal [13:46:06] there have beena lot of little configs like this that i've had to figure out too, that are not documented by cdh instructions at least [13:46:10] not a good sign! [13:46:21] ottomata: agreed from the not good sign [13:46:34] ottomata: the friend uses chef :) [13:48:37] :) [13:49:07] By the way ottomata, he'd like to know if we run into container release issues [13:49:09] Looks like python 2.7 is going to be a pain with my current utilities, but this is simple so I'll start from scratch. [13:49:20] Cause that's the ones he didn't manage to solve correctly [13:49:24] Why does pyspark have to only work with a version of python that is a decade old! [13:49:39] halfak: Maybe because it's a decade old ? [13:49:48] Oh. ha [13:49:55] halfak: let's go scala if rebuild is needed ;) [13:50:18] joal, I need to schedule about 10 hours to pick up scala and get competent :( [13:50:24] huhuhu [13:50:49] * halfak just got an email for annual reviews [13:50:50] halfak: send me your python, I'll write scala, and it will be a correct tuto, don't you think ? [13:50:54] which will be like 5 hours :( [13:51:06] Sure. [13:51:24] Depends on the complexity of the python though :) [13:51:27] https://github.com/halfak/mwrefs [13:51:30] It's really simple. [13:51:36] https://github.com/halfak/mwrefs/blob/master/mwrefs/extract.py [13:51:40] That does the hard bit. [13:51:42] Just a regex. [13:52:01] This glues it together: https://github.com/halfak/mwrefs/blob/master/mwrefs/utilities/diffs.py [13:52:01] ok [13:52:19] https://github.com/halfak/mwrefs/blob/master/mwrefs/utilities/diffs.py#L58-L61 [13:53:07] Ok got it [13:53:31] now, is a format excluding the revision text ok ? [13:53:40] ohhhh my bad [13:53:44] halfak: --^ [13:53:51] it's user_text, not text [13:54:25] joal, you're just talking about output, right? [13:54:35] based on that, I should have something running for simplewiki hopefully tomorrow [13:54:39] halfak: yup [13:54:59] halfak: output format in diffs.py comments [13:55:03] :) Glad to have you pick it up joal, but I'm worried about having you write me scala. [13:55:08] yeah joal [13:55:32] halfak: I am a bad programmer, but I wouldn't thought you had found it so fast ! [13:55:43] :P joal [13:55:54] joal, I need to be quick with this stuff -- that's why I've insisted for so long that I stick the environment that I'm agile in. [13:56:10] I have to backfill EventLogging, so while it runs, I'll write scala :) [13:56:33] I really do hope to have the first version of it by tomorrow [13:56:37] halfak: --^ [13:56:49] halfak: if not good enough, I'll let you go with pyspark :) [13:57:17] joal, I appreciate it, but that's not it. I need to write the analysis code! So I should be either scheduling time to level up in scala or writing this in pyspark. [13:57:37] Either that, or you write all my analysis code now >:) [13:57:49] muhahahaha [13:57:52] 'cause I have a big backlog :P [13:58:01] halfak: you're the best manager I've found ! [13:58:36] I can provide the tsv file as defined in diffs.py [13:58:42] how would that be ? [13:58:47] halfak: -^ [13:59:17] joal, that would be helpful, thanks :) [13:59:34] ottomata: I have some code review waiting for you (not the wikidumps one, the other one about pageviews) [13:59:45] * halfak promises himself that he'll run some pyspark on altiscale with the new input format this week. [13:59:52] ottomata: it includes maven poms improvements [14:00:25] * joal jsut wants to get to the point of computing wikicredits batch style :) [14:02:43] oh! [14:04:12] looking [14:04:41] hm joal [14:04:53] why can/should you remove the dependencies from the top level pom? [14:04:56] you are leaving some other deps [14:05:10] Just preventing some rewrite [14:05:22] oh, provided.... [14:05:24] I removed the ones we wnat to be scoped provided [14:05:25] ? [14:05:43] yeah --> maven doesn't propagate scope provided [14:05:50] ottomata: --^ [14:05:53] hm, that makes sense actually, so those won't be in the uber jar [14:06:00] beacuse they are already on the box [14:06:07] and get included in the CP automatically when given to hadoop [14:06:08] ja? [14:06:13] correct: refinery-hive.jar goes from 120mb to 20mb :) [14:06:18] awesooome [14:06:19] nice [14:06:21] :) [14:06:31] * joal is happy ! [14:06:34] i did that for spark-assembly too i think, didn' tthink to do it for hadoopy jars [14:06:47] Plus some parameterization of version numbers [14:07:06] heheh, i'm fine with this, in the future though, you probably want things like pom optimizations and actual logic changes in different commits [14:07:18] Yeah, I nkow that [14:07:36] I thought about doing it, and had already a big bunch of small changes commited ... [14:07:43] I have been lazy on this one :( [14:08:45] sok i know the feeling [14:08:46] And by the way, I have run through many trouble running the udfs yesterday [14:09:08] hive error: function not defined [14:09:30] create temp function...:)? [14:09:33] I htink it must because of multiple jars with the same name uploaded [14:09:36] yeah [14:09:37] hmm [14:09:41] create is ok, running was not [14:09:52] except one time, it went throuhg it [14:09:58] ottomata: --^ weird [14:10:20] oh weird [14:10:22] that is weird [14:11:19] I'll restest renaming the jar [14:11:58] halfak: page 5137507 has 821551 revisions in my dataset :) [14:15:14] 588505 in WP:Sandbox [14:15:29] But this is not the first "Sandbox" [14:16:32] Hmm... maybe it is. The first revision was in "20021220215014" [14:16:41] but the page ID is in the 16 millions. [14:17:15] the one I'm talking about is page_id = 5137507 [14:17:36] joal, oh yeah. I know. I was just scoping out the other big pages. [14:17:44] oh, ok sorry :) [14:17:49] halfak: --^ [14:17:49] :) [14:18:04] is this one coherent with what you have ? [14:18:08] Yes [14:18:13] cool :) [14:18:17] good sign [14:18:37] I need to go for some EL in order to have that running, and then write some scala :) [14:18:44] will talk soon :) [14:19:41] joal: what's up with thie? [14:19:42] https://gerrit.wikimedia.org/r/#/c/205833/1/refinery-hive/src/main/java/org/wikimedia/analytics/refinery/hive/RefererClassifierUDF.java [14:19:43] this* [14:20:14] ottomata: I'd love to include that data in refined webrequest [14:20:26] But a map instead of a string is really to muvh ! [14:20:32] So I rewrote [14:20:40] indeed, why was it a map in the first place, do you know? [14:20:50] Maybe better to add a new one and not break somebody else code ... [14:20:55] :-S [14:21:05] I have no idea [14:21:16] i reviewed it..... [14:21:19] looking... [14:21:30] The original function in core is cool, but the udf using a map is overkill [14:21:54] no i just never asked [14:22:11] Particularly now that we have parquet ! [14:22:15] ok, also, why extends UDF instead of GenericUDF [14:22:15] ? [14:22:28] parquet doesn't parses map I think [14:22:51] While using strings makes good advantage of columnar format [14:23:10] hm, joal, are you sure it isn't possible that the map could contain multiple values [14:23:12] maybe that is why [14:24:15] ottomata: result is disjonctive [14:24:20] one or the other [14:24:24] seems like it [14:24:33] lines 101/102/103 [14:24:36] ok, hm, also, i think nuria has some big preference for GenericUDF over just UDF [14:24:51] Oh, really ? [14:24:57] I don't mind rewriting [14:24:59] oh ja, even the result just returns one of those [14:25:21] dunno though, maybe it was a special case, can't remmeber [14:25:29] I think Ellery did that [14:25:29] nuria's last day issss today? yesterday? :) [14:25:38] I think it's yesterday [14:25:40] aye [14:25:41] :) [14:25:49] qchris wrote this referer thing originally [14:25:53] But it's a pity cause I could do with some help on EL :( [14:26:00] hm, i think UDF is fine. [14:26:04] ok. i'm for it! [14:26:05] k [14:26:10] :) [14:26:41] (CR) Ottomata: Patch pageview definition (4 comments) [analytics/refinery/source] - https://gerrit.wikimedia.org/r/205833 (owner: Joal) [14:26:42] some coments. [14:33:32] Analytics-Tech-community-metrics, ECT-April-2015, Patch-For-Review: "Age of unreviewed changesets by affiliation" shows negative number of changesets - https://phabricator.wikimedia.org/T72600#1227675 (Dicortazar) The dataset is now updated, but the pop-up does not work. Working on this now. As an est... [14:38:41] ottomata: holaaaaa [14:39:02] ottomata: i will be here today and probably part of tomorrow, need to commit spark stuff [14:39:19] cc joal [14:39:46] ottomata: ya all are going to love the craftiness to append to a file in spark [14:40:46] nuria: can't wait to read you :) [14:40:59] nuria: could go with some help on EL [14:41:05] let me know when you have time [14:42:30] Analytics-Tech-community-metrics, ECT-April-2015: Provide list of oldest open Gerrit changesets without code review - https://phabricator.wikimedia.org/T94035#1227743 (Dicortazar) Working on the last-patchset-upload order. [14:44:11] joal: now is good [14:44:24] k [14:44:27] batcave ? [14:45:21] give me 2 mins [14:45:26] sure, np [14:47:55] k batcave joal [15:02:08] Analytics-Tech-community-metrics, Wikimedia-Git-or-Gerrit, ECT-April-2015: Basic metrics about contributors exercising +2/-2 permissions in Gerrit - https://phabricator.wikimedia.org/T59038#1227834 (Dicortazar) @Qgil, we've (@Lcanasdiaz among others) discussed about the renames a bit more. * "Partici... [15:16:57] (PS2) Joal: Patch pageview definition [analytics/refinery/source] - https://gerrit.wikimedia.org/r/205833 [15:18:05] (CR) Joal: Patch pageview definition (4 comments) [analytics/refinery/source] - https://gerrit.wikimedia.org/r/205833 (owner: Joal) [15:31:43] joal: standupppp [15:31:52] joinnnnning ! [15:35:41] grep exits non-zero if it doesn't find a match! [15:35:57] Weird hadoop issues. [15:36:19] ha, halfak, ja then you can do fun things like [15:37:10] grep -q 'am I there?' file && echo "Yes I am" || echo "nope, didn't find it" [15:38:01] ottomata, I can't seem to figure out how to get it to not error out my hadoop streaming job. [15:38:07] It's OK to not find a match sometimes. [15:39:39] * halfak starts hair pulling. [15:40:05] * joal wants a picture :) [15:40:32] Analytics-EventLogging, Analytics-Kanban: Inserts events by scid - https://phabricator.wikimedia.org/T96872#1228006 (mforns) NEW [15:40:42] Analytics-EventLogging, Analytics-Kanban: Inserts events by scid - https://phabricator.wikimedia.org/T96872#1228013 (mforns) a:mforns [15:41:57] ha! "-D stream.non.zero.exit.is.failure=false" [15:42:15] (CR) Ottomata: Patch pageview definition (3 comments) [analytics/refinery/source] - https://gerrit.wikimedia.org/r/205833 (owner: Joal) [15:42:44] halfak: is this in a shell script? [15:42:48] Just a grep [15:42:55] in a hadoop job? [15:42:55] that's all I'm doing in my mapper. [15:42:57] oh as one of the maps [15:42:58] you could do [15:42:58] Yeah [15:42:59] uhhh [15:43:10] (grep .... ; true) or somethign like that [15:43:10] uhhh [15:43:16] Analytics-EventLogging, Analytics-Kanban, Patch-For-Review: Troubleshoot EventLogging missing chunks of data {oryx} - https://phabricator.wikimedia.org/T96082#1228017 (mforns) Open>Resolved [15:44:02] ah no [15:44:02] like this [15:44:02] https://github.com/wikimedia/operations-puppet/blob/21c72942dd7bf25dbe0759d2f867082e966bfb45/modules/puppetmaster/files/puppet-merge#L41 [15:44:06] or, your way :) [15:44:15] ottomata, I was trying that, but it seems to have failed for weird reasons. [15:44:22] I think I'm escaping wrong in my command. [15:45:26] You need to double-escape when you have quotes in the grep string too. [15:45:26] ah [15:45:26] aye [15:45:26] this returns 0 for me, btw [15:45:26] (grep weeewooo /etc/passwd; true); echo $? [15:45:34] Maybe there's a "weeewoo" in there [15:45:36] haha [15:45:37] nope [15:45:38] ;) [15:45:40] Analytics-EventLogging, Analytics-Kanban, Wikimedia-Search: Estimate maximum throughput of Schema:Search (capacity) {oryx} - https://phabricator.wikimedia.org/T89019#1228033 (Nuria) a:Nuria>kevinator [15:45:43] Analytics-EventLogging, Analytics-Kanban: Backfill EventLogging april holes. - https://phabricator.wikimedia.org/T96874#1228034 (JAllemandou) NEW a:JAllemandou [15:46:08] Analytics-Kanban, Analytics-Wikimetrics, Community-Wikimetrics, Patch-For-Review: Utf-8 names on json reports appear as unicode code points: "\u0623\u0645\u064a\u0646" - https://phabricator.wikimedia.org/T93023#1228042 (Nuria) a:Nuria>None [15:46:31] Analytics-EventLogging, Analytics-Kanban: EL alarms should be included just in the tugsten host - https://phabricator.wikimedia.org/T89469#1228044 (Nuria) Open>Resolved [15:47:18] Analytics-EventLogging, Analytics-Kanban, Patch-For-Review: Troubleshoot EventLogging missing chunks of data {oryx} - https://phabricator.wikimedia.org/T96082#1228047 (kevinator) backfilling task for missing data: T96874 [15:51:11] Analytics-EventLogging, Analytics-Kanban: Backfill EventLogging april holes. - https://phabricator.wikimedia.org/T96874#1228054 (JAllemandou) http://etherpad.wikimedia.org/p/el_utils [15:58:05] joal: can you ask your friend why he added -local_nodemanager_url=http://127.0.0.1:8042/http://127.0.0.1:8042/http://127.0.0.1:8042/http://127.0.0.1:8042/http://127.0.0.1:8042/http://127.0.0.1:8042/ [15:58:05] ? [15:58:07] ack! [15:58:13] -local_nodemanager_url=http://127.0.0.1:8042 [15:58:16] that line? [15:58:24] Sure [15:59:01] hmm [15:59:08] actually it will probably wait tomorrow [15:59:20] He is a morning person, so it won't be online anymore [15:59:29] s/it/he/ [16:00:16] ottomata: about short names [16:00:28] I don't mind having long ones since we are using parquet [16:00:48] Would we be using a row-oriented format, shorter = less disk used [16:01:45] compressed i don't care :) [16:01:47] hah [16:01:55] huhub [16:01:58] true [16:02:17] So, how do you want me to change that ? [16:02:35] but i mean, in general that is just my opinion by default, i think the meaning that is lost by using abbreviations everywhere is not worth the disk space gained :) [16:02:46] umm, whatcha mean? [16:02:48] I can agree with that :0 [16:02:58] I'm just trying to understand why we use an enum rather than just string constants [16:03:08] if we are going to return strings in most uses anyway [16:03:10] hm, do you want me to change the core function, removing the enum ? [16:03:18] yeah [16:03:21] original code [16:03:23] yeah [16:03:31] Ok, so I'll patch that as well [16:03:49] yea, i'm for it [16:03:52] ok ! [16:08:18] ottomata: -local_nodemanager_url (The URL of the local Yarn Node Manager's HTTP [16:08:21] interface, used to detect if the Node Manager fails) type: string [16:08:34] But it might have disappeared in new versions [16:08:53] that option is part of our version [16:09:03] but i don't have any indication that we need it [16:09:08] i'll set it though [16:09:17] :) [16:17:14] (CR) Nuria: [C: 1] "I am actually in the camp that prefers enums to string constants." [analytics/refinery/source] - https://gerrit.wikimedia.org/r/205833 (owner: Joal) [16:18:07] nuria: i'm fine with the enum in general, but if we are returning a string in our main use case for the method [16:18:09] what's the point? [16:18:42] huhu, I'll for this thing to be settled before doing any more change :) [16:19:10] ottomata: I think value comes from easier to read code if the enum is itself a class (so values are well defined) [16:19:36] ottomata: so something like: [16:19:40] https://www.irccloud.com/pastebin/0OrIR1bv [16:19:41] why is Classname.Value easier than just VALUE [16:19:41] ? [16:20:29] it seams just as readable to me nuria [16:20:42] ultimately both are just constants [16:20:53] code readability is the same [16:21:02] its just the underlying data that is stored that is different [16:21:05] in one case it is an int [16:21:09] in another it is a string [16:21:38] Analytics-EventLogging: Send graphite metrics for Schema. as well - https://phabricator.wikimedia.org/T95780#1228121 (Tgr) >>! In T95780#1200229, @Nuria wrote: > Sorry but strongly disagree, comment above implies using EL as an Error Log, and EL is not an error log system, it's a tier-2 system... [16:21:46] if we were using just the thing in java code to do stuf [16:21:49] i'd say enum is nice [16:22:01] but, we are making this result useful outside of java java [16:22:06] so, we return a string [16:22:23] just not sure what the point of storing it in memory for a little while as an int is [16:22:36] storing it as an int makes joal have to write a switch statement to return the proper string [16:22:55] ottomata: but if you define an enum like : [16:22:59] which makes code less readable, in my opinion [16:23:40] https://www.irccloud.com/pastebin/VbdLWYkq [16:23:54] there is your string to int translation, case is not needed right? [16:24:09] ottomata: we were using impala 2.0.0 from cdh 5.2.0 [16:24:12] now, ottomata not ahill to die on [16:24:41] hmmm? [16:24:47] nuria: not sure i understand [16:25:30] *not a hill to die on, so i am fine with just using strings [16:25:32] how does that avoid a case? [16:25:55] aye, ya, but i just want to understand [16:26:03] if we can use enum, and can still do [16:26:23] return classifyReferer(referer) -> String [16:26:27] then i'm cool [16:27:56] oh [16:27:57] joal [16:27:57] ottomata: so the "case" is not needed if the enum has both the string and int value [16:28:04] there is enum .name() method? [16:28:12] that's what i'm reading now [16:28:14] yes, it exists [16:28:45] oh cool then maybe just do [16:28:56] public enum RefererClassification { [16:28:56] UNKNOWN ("unknown"), [16:28:56] ... [16:29:02] and then in the UDF just do [16:29:14] return classifyReferer(referer).name() [16:29:15] ? [16:29:34] possible :) [16:29:41] I don't mind doing one or the other [16:29:49] either that or string constants is fine with me [16:29:53] We just need to make a decision [16:29:54] and it sounds like nuria likes the enums [16:30:00] :) [16:30:00] so, go with enums with string values [16:30:10] i just want to avoid caseing the string values really [16:30:19] It's just that we've used strings and not enums everywhere else :) [16:30:20] that way there is a real constant string value associated with the result [16:30:22] yeah [16:30:23] OR [16:30:24] use strings! [16:30:27] i don't care now :) [16:30:33] huhuhu ) [16:30:34] does make things simpler [16:30:37] ottomata: hahah [16:30:43] ay ay [16:30:49] From almost there, to a good knotbag :) [16:30:54] haha [16:31:22] joal, up to you at that point, as long as the string value is a constant that is associated with the return of classifyReferer, I am fine [16:31:26] ottomata: also i like returning long names like "unkown" rather than an abbreviation, yes. [16:31:29] ok [16:31:52] i just don't want the switch statment in the UDF [16:31:59] the string value should be part of the method return [16:32:01] nuria, ottomata : I'll go for strings from the core, to be coherent with the coed base [16:32:05] that way users of the method can't mix it up [16:32:09] ok cool, fine with me [16:32:25] nuria: ? [16:32:47] time to make some lluuuunch, bbiab [16:34:54] nuria: HAve the fact of winning this battle against ottomata means you prefer me going for enums ? [16:35:16] joal: nah [16:35:20] :) [16:35:32] not at all [16:35:37] you ok for strings then I guess :) [16:35:41] yes [16:35:46] Cool ! [16:35:56]