fluid-work IRC Logs-2013-03-08

Owned by Fluid IRC Log Bot
Last updated: Mar 08, 2013Version comment
4 min read

[10:38:29 CST(-0600)] <jessm> fluid-everyone: sorry to miss standup

[10:38:50 CST(-0600)] <jessm> i've been working on the Leonardo grant contract and deliverables and am going to catch up on PGA use cases today

[14:02:01 CST(-0600)] <cindyli> hi, Bosmon

[15:15:22 CST(-0600)] <Bosmon> cindyli !

[15:16:05 CST(-0600)] <cindyli> Bosmon: all enactors are done, would you mind code review those enactors?

[15:16:18 CST(-0600)] <Bosmon> cindyli - sure!

[15:17:43 CST(-0600)] <cindyli> i'm half-way refactoring uiEnhancer, which is in the same branch. that part, not done yet

[15:18:04 CST(-0600)] <cindyli> Bosmon: the comparison of that branch with infusion master - https://github.com/cindyli/infusion/compare/fluid-project:master...FLUID-4908

[15:18:11 CST(-0600)] <cindyli> thanks

[15:18:23 CST(-0600)] <Bosmon> thanks

[15:20:45 CST(-0600)] <Bosmon> cindyli - I think the implementation could be simplified by just having all ants use the standard path "value" for their setting

[15:21:43 CST(-0600)] <Bosmon> And you can then ditch the "getModelValueByPath" machinery etc.

[15:22:13 CST(-0600)] <cindyli> Bosmon: in that case, all ants need to have a model path that points to the model key that the ant responds to

[15:22:27 CST(-0600)] <cindyli> oh i thought that's where we need it

[15:23:34 CST(-0600)] <Bosmon> cindyli - I think after many years its time we started to grow up and allow each ant to have its own private bit of model : P

[15:23:42 CST(-0600)] <Bosmon> We've suffered with this crazy "global model" system for long enough....

[15:24:22 CST(-0600)] <cindyli> right. then the question is how to link the private model to a portion of the entire model

[15:24:25 CST(-0600)] <Bosmon> Yes

[15:24:40 CST(-0600)] <Bosmon> We should make a simple "changeApplier relay" system of the kind we have in the videoPlayer integrations

[15:24:46 CST(-0600)] <Bosmon> It needn't be very sophisticated for now

[15:24:56 CST(-0600)] <Bosmon> but at least it will get all of this annoying model gearing code out of the ants

[15:25:02 CST(-0600)] <cindyli> you mean indirectRead and write?

[15:25:06 CST(-0600)] <Bosmon> No, the other one (smile)

[15:25:14 CST(-0600)] <Bosmon> indirectRead and Write is actually similar to THIS system

[15:25:21 CST(-0600)] <cindyli> ya

[15:25:21 CST(-0600)] <Bosmon> In that it relies on a "global shared model"

[15:25:32 CST(-0600)] <cindyli> that's what i'm doing (smile)

[15:25:35 CST(-0600)] <Bosmon> I mean the OTHER system, that we use to connect UIOptions model to the videoPlayer's model

[15:25:51 CST(-0600)] <cindyli> i will take a look on that

[15:26:10 CST(-0600)] <Bosmon> The thing called "fluid.modelRelay"

[15:26:16 CST(-0600)] <Bosmon> We don't need anything as sophisticated as that

[15:26:23 CST(-0600)] <Bosmon> Since we don't need to take care of any timing issues

[15:26:27 CST(-0600)] <Bosmon> Or any "lensing"

[15:26:45 CST(-0600)] <Bosmon> But I think we should at least set up things in something like the way we mean to go on

[15:26:51 CST(-0600)] <Bosmon> The "global model" system isn't really tenable in the long term

[15:27:33 CST(-0600)] <cindyli> cool. great information. will find out

[15:27:36 CST(-0600)] <Bosmon> You can write a very simple "model relay" component that takes two models, two appliers, and two relative paths as arguments

[15:27:43 CST(-0600)] <Bosmon> And then sets up relay of ChangeApplier events from one to the other

[15:27:54 CST(-0600)] <Bosmon> This way all of this code stays outside the components

[15:28:12 CST(-0600)] <Bosmon> All you need to make sure of is you don't relay the events circularly forever (smile)

[15:28:56 CST(-0600)] <cindyli> haha

[15:29:13 CST(-0600)] <Bosmon> http://issues.fluidproject.org/browse/FLUID-3674

[15:29:22 CST(-0600)] <Bosmon> This open JIRA and the related ones describe some of the plans

[15:29:47 CST(-0600)] <Bosmon> Although when it was written, we still believed in the "Cautious ChangeApplier"

[15:29:51 CST(-0600)] <Bosmon> I think we don't believe in it any more

[15:30:59 CST(-0600)] <Bosmon> only 3.5 years old : P

[15:31:19 CST(-0600)] <Bosmon> No, 2.5.....

[15:31:39 CST(-0600)] <cindyli> sounds quite complex, i mean the jira

[15:31:53 CST(-0600)] <Bosmon> Yes

[15:32:10 CST(-0600)] <Bosmon> We just need something very simple for now... you can see in the comments of fluid.modelRelay that it supports (at least) 3 modes

[15:32:17 CST(-0600)] <Bosmon> Of which the simplest is: // i) a simple string representing a direct relay between changes to one path and another

[15:32:21 CST(-0600)] <Bosmon> And this is all you need to support ehre

[15:33:06 CST(-0600)] <cindyli> nice

[15:33:40 CST(-0600)] <cindyli> thanks

[22:29:46 CST(-0600)] <Bosmon> well done, avtar_ ..... working into the dark hours!

[22:30:02 CST(-0600)] <avtar_> lol

[22:31:29 CST(-0600)] <Bosmon> What blew up with the hardware?

[22:32:35 CST(-0600)] <avtar_> it originally was network related but when i started transferring vm images the kernel started to panic generating scsi bus errors

[22:32:46 CST(-0600)] <Bosmon> ouch, awful

[22:33:16 CST(-0600)] <Bosmon> SCSI seems so quaint these days (smile)

[22:34:10 CST(-0600)] <avtar_> /dev/cloud?

[22:34:26 CST(-0600)] <avtar_> that sounds more 2013'ish

[22:34:38 CST(-0600)] <Bosmon> haha

[22:34:43 CST(-0600)] <avtar_> time to go breathe non-data centre air

[22:34:50 CST(-0600)] <avtar_> bbiab

[22:34:53 CST(-0600)] <Bosmon> Well, when a consumer-grade device can push 6 GB/s you wonder why anyone would bother....

[22:34:56 CST(-0600)] <Bosmon> Good luck (smile)

[22:35:00 CST(-0600)] <Bosmon> sorry, Gb/s (smile)

[22:35:02 CST(-0600)] <avtar_> thanks