This version (2017/05/27 13:44) is a draft.
Approvals: 0/1

[09:13:37] <purplefox> temporalfox: pmlopes cescoffier morning folks

[09:13:45] <pmlopes> good mornin

[09:13:49] <pmlopes> morning

[09:17:48] <purplefox> temporalfox: cescoffier pmlopes: let's do a quick bluejeans meeting this morning, when's a good time?

[09:17:57] <purplefox> I'd prefer asap

[09:18:06] <cescoffier> you said 10 AM my time no ?

[09:18:23] <cescoffier> but I'm available if you want to do it earlier

[09:19:21] <pmlopes> let me just reboot and i am ready

[09:29:49] <purplefox> cescoffier: pmlopes temporalfox we may as well wait 30 mins and do it as planned

[09:30:17] <pmlopes> purplefox: sure

[09:53:57] <cescoffier> pmlopes purplefox temporalfox - don't worry about the maven service factory build failure, I'm fighting against SSL certificates

[10:02:02] <temporalfox> hello everyone

[10:02:12] <cescoffier> purplefox: did you send a BJ invitation ?

[10:02:15] <cescoffier> hi temporalfox

[10:02:30] <temporalfox> just in time for the BJ :-)

[10:04:39] <purplefox> haven't sent an invite but anyone can do this :)

[10:05:22] <cescoffier> are you sure, don't we need a special acount for this ?

[10:06:09] <purplefox> just go to

[10:06:20] <purplefox> and log in with the same credentials as the vpn

[10:06:31] <temporalfox> do you have a meeting id ?

[10:06:32] <purplefox> i'm having some trouble right now getting a meeting to start

[10:06:43] <temporalfox> ok

[10:07:26] <purplefox> trouble with bj is they use a NSAPI plugin which is not supported in later versions of chrome on linux

[10:07:31] <purplefox> and will be dropped in all chrome soon

[10:07:40] <purplefox> so i have to use firefox for this

[10:07:48] <purplefox> and it gives me all sorts of plugin warnings

[13:12:09] <pmlopes> purplefox, temporalfox, cescoffier i've closed the mongo issue with the millis and i think that we should keep the current implementation instead of mine parser because i think it is not wise to add a custom parser just to handle the mongodb client and rely on the JDK DateTime API for all date/time operations, this way we are consistent across clients, components.

[13:12:48] <temporalfox> pmlopes is it an issue reported in other frameworks that use date and time ? or even in openjdk tracker ?

[13:14:11] <pmlopes> no, the JDK implementation is not broken per se, it is that the ISO states that TZ format is HH:SS where the colon is optional, while the DateTime API was implemented with non-optional colon

[13:16:17] <pmlopes> if we consume dates from any webbrowser it always works fine because browsers implement the EcmaScript404 which states that dates are always encoded as: YYYY-MM-DDTHH:mm:ss.sssZ which is the format we use for dates read from mongo

[14:54:38] <temporalfox> pmlopes I'm looking at the configuration of JWT, I think we should use Vertx filesystem to handle config files

[14:54:59] <temporalfox> pmlopes that can resolve from classpath or filesystem and most importantly handle the specific vertx cwd property

[14:55:33] <temporalfox> also I think we could reuse some code from vertx core for reading keystores

[14:55:39] <temporalfox> (but later after 3.0)

[14:57:04] <pmlopes> well jwt can load keystores from classpath or file it is a url, file:/home/user/file.jks for example [14:57:43] <pmlopes> where should i change the loading of key stores? in the examples? [14:57:52] <temporalfox> no I mean in the impl [14:58:34] <temporalfox> we are using in some places of vertx, the vertx filesystem [14:58:40] <temporalfox> like templates [14:59:26] <temporalfox> look at [14:59:34] <temporalfox> VertxInternal.resolveFile [14:59:41] <temporalfox> that resolves from FS or classpath [15:00:13] <temporalfox> in addition in takes in account a System property “vertx.cwd” for reconfiguring the current dir (used by examples) [15:00:40] <pmlopes> so if i do that i can remove the prefix where to look for files [15:00:42] <temporalfox> using this at least would allow to use vertx.cwd for loading files [15:00:44] <temporalfox> yes [15:01:03] <temporalfox> and if we have other places loading files, we could use it as well [15:02:46] <temporalfox> pmlopes later we can try to reuse code from vertx core for reading keystores, since it can also handle PKCS12 and PEM files [15:03:09] <temporalfox> (I'm not sure the core code is easy to reuse outside of core at the moment) [15:04:39] <temporalfox> I will spend time on this JWT refactoring now and let you know [15:04:47] <temporalfox> it seems straightforward to do [15:25:10] <temporalfox> pmlopes ahah you and I did the same thing [15:25:17] <temporalfox> anyway it was not big [15:25:32] <temporalfox> one thing we should do is change the json keystore configuration format [15:25:44] <pmlopes> i am updating the examples [15:25:57] <temporalfox> can you change it to { “keyStore” : { “path”: …, “password”: … } [15:26:07] <temporalfox> so it's the same than in vert.x core for keystores [15:26:14] <pmlopes> ok [15:26:30] <temporalfox> later we can introduce JWTOptions [15:26:38] <temporalfox> and reuse KeyStoreOptions [15:26:42] <temporalfox> which has JKSOptions [15:27:24] <temporalfox> PfxOptions and PemKeyCertOptions [15:27:31] <temporalfox> drop the “type” [15:27:41] <temporalfox> do we have other place we can use Vert.x for loading ? [15:27:53] <temporalfox> I'm thinking of the properties files for auth [15:28:20] <temporalfox> like [15:28:55] <temporalfox> ah in this case it's handled by shiro [16:05:31] <pmlopes> temporalfox: the type is important if i drop it i must assume the jks file is jceks because it is the only format where i can store mac's [16:06:03] <temporalfox> ok [16:06:28] <temporalfox> I will try the polyglot examples [17:53:54] <jtruelove> purplefox did you see my question on your feedback in this pull request → [18:11:22] <lorin1> Is there a way to get the size of the worker thread pool programmatically at run time? I'd like to be able to verify that my configuration script that sets the pool size is correct. [19:00:19] <LaSombra|home> Hello, good evening. I recently discovered Vert.x and would like to contribute. I read the contributing doc, and got everything set up, I believe. Now, I've found a bug I think I can tackle but I couldn't find the GitHub repository for the Vert.x core [19:00:52] <LaSombra|home> Would it be the eclipse/vert.x one? [19:01:30] <AlexLehm> yes [19:01:42] <AlexLehm> the core project uses eclipse bugzilla however [19:01:46] <LaSombra|home> Thanks AlexLehm [19:02:13] <LaSombra|home> AlexLehm, yeah, that's where I found the bug. [19:53:49] <aesteve> hi LaSombra|home LaSombra :) [19:54:29] <aesteve> that's a feature request I reported I think you linked earlier [19:55:38] <aesteve> if you want to contribute to vertx-core make sure you signed the Eclipse CLA, then sign your commit (with the “-s” option) before submitting your changes [20:14:28] <LaSombra|home> Hey aesteve :) I signed the CLA [20:15:35] <LaSombra|home> aesteve, now I need to look what I can do [20:16:21] <aesteve> I think this one is really easy, it's just a quality of life issue but not sure if it should be implemented officially or not [20:16:44] <aesteve> that's why I opened a Bugzilla and not patched it directly [20:36:41] <AlexLehm> purplefox, I have one item open for the mail client and that is the idle timeout on the pooled connection [20:38:02] <AlexLehm> I have kept that in, I can take it out if you prefer, but I kind of think this is useful [21:20:00] <LaSombra|home> aesteve, got it