These are chat archives for highfidelity/hifi

11th
Mar 2014
Leonardo Murillo
@murillodigital
Mar 11 2014 00:00
cool thing about this is, virtually trivial to have 1:1 going as well
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:00
@ZappoMan, are the Euler angles in the scripting language supposed to return angles in degrees or radians?
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:01
they just wrap the C++… so it's whatever was there.
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:01
ok thanks
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:01
I don't know what code uses them in JS… let me check.
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:01
that means they were in degrees
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:01
but I don't think any JS assumes it's in degrees...
let me check
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:02
I saw the code, but wasn't sure if you had done some degrees-->radians conversions elsewhere
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:02
this code clearly assumes degrees…
function fixEulerAngles(eulers) {
    var rVal = { x: 0, y: 0, z: eulers.z };
    if (eulers.x >= 90.0) {
        rVal.x = 180.0 - eulers.x;
        rVal.y = eulers.y - 180.0;
    } else if (eulers.x <= -90.0) {
        rVal.x = 180.0 - eulers.x;
        rVal.y = eulers.y - 180.0;
    }
    return rVal;
}
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:02
yup, degrees
ok, I'm converting most of the C++ code to use radians by default (because GLM is migrating to radians soon)
but I'll not change the JS behavior
Dimitar Dobrev
@ddobrev
Mar 11 2014 00:03
@birarda - would you prefer putting the convention changes in a separate commit to the group chat one or combined in a single commit?
Stephen Birarda
@birarda
Mar 11 2014 00:04
you can decide
both are fine
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:04
@AndrewMeadows - are you sure?
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:04
@ZappoMan, could you elaborate on that question?
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:04
are you sure you don't want to keep the C++ and JS in sync?
seems like all arguments for moving to radians would apply independent of the language choice of the code
I'd rather all the code be in the same units
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:05
well, I'm sure we should be moving to radians where possible with a few exceptions (human readable yaw/pitch/roll data mostly)
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:06
if we want to expose degreesToRadians() and radiansToDegrees() helpers for people who don't know how to convert… I can see that be somewhat useful.
I assume for example that this code, doesn't care what the units are…
        // add some rotation...
        var deltaRotation = getAndResetGrabRotation();
        var GRAB_CONTROLLER_TURN_SCALING = 0.5;
        var euler = Vec3.multiply(Quat.safeEulerAngles(deltaRotation), GRAB_CONTROLLER_TURN_SCALING);

        //  Adjust body yaw by yaw from controller
        var orientation = Quat.multiply(Quat.angleAxis(-euler.y, {x:0, y: 1, z:0}), MyAvatar.orientation);
and this…

        //  Adjust head pitch from controller
        MyAvatar.headPitch = MyAvatar.headPitch - euler.x;
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:07
conversion from one to another is just a multiplier. I think helper functions are overengineering
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:08
so why don't we just make everything in radians?
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:08
those examples are using Quat.safeEulerAngles so they are using degrees
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:09
today true… but if you changed it to return radians… theyd still work
so long as all the other functions assumed radians
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:09
I know of at least two exceptions to the "always use radians" rule:
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:09
MyAvatar.headPitch = radiansNotDegrees
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:10
(1) the data stored in FBX files are using degrees, so we extract them and do some math with degrees in FBXReader.cpp
(2) Human readable euler angles are often stored in degrees
such as fields that you might change in a UI
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:10
where does #2 happen?
we have vertical FOV in the prefs dialog box.
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:11
For example, if you were manually setting the pitch of an object, via a slider or edit box.
Dimitar Dobrev
@ddobrev
Mar 11 2014 00:11
@birarda - done, just about #include "ui_chatWindow.h":
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:11
right, I'm going to keep FOV in degrees
Dimitar Dobrev
@ddobrev
Mar 11 2014 00:11
I've done it like:
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:11
we don't have that in our system today… I'm saying.
Dimitar Dobrev
@ddobrev
Mar 11 2014 00:11

include "ChatWindow.h"

include "ui_chatWindow.h"

OK?
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:11
woah
I heard that.
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:12
whao is right
whaaaoooo
Stephen Birarda
@birarda
Mar 11 2014 00:12
haha
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:12
woooah
Stephen Birarda
@birarda
Mar 11 2014 00:12
thanks markdown
Dimitar Dobrev
@ddobrev
Mar 11 2014 00:12
Oh, did I do that? I copied code directly from Creator.
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:12
@AndrewMeadows - so what about JS APIs… let's make them radians.
all input/output.
so that means properties like MyAvatar.headPitch would be in radians not degrees… that doesn't seem unreasonable...
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:13
@ZappoMan, ideally we would pick one as default, and allow users to overide it with some #define or something (sorta like how GLM used to do it). But I didn't want to worry about that at this stage.
Dimitar Dobrev
@ddobrev
Mar 11 2014 00:13
@birarda - if that's fine, I'm committing.
Stephen Birarda
@birarda
Mar 11 2014 00:13
@ddobrev does the ui file need to be below it for compile?
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:14
@AndrewMeadows - JS doesn't really have the concept of #defines
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:14
Yes but... we don't have to make that transition right now. I don't really want to break all of the JS scripts on the first checkin.
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:14
I just checked
Dimitar Dobrev
@ddobrev
Mar 11 2014 00:14
@birarda - no but they kind of go together.
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:14
those two pieces of code above are the only places
Stephen Birarda
@birarda
Mar 11 2014 00:14
we want ChatWindow.h to be the last
so if you want to stick them together I think that’s alright
well
actually I think it should just be up witht he group above
alphabetiaclly
alphabetically*
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:15
Hrm... ok if so then it might not be too hard to switch the JS to radians. I'll consider it. But will certainly make that change as a separate commit.
Dimitar Dobrev
@ddobrev
Mar 11 2014 00:16
@birarda - I'd rather have them together but OK, I don't mind.
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 00:16
@AndrewMeadows - there's also some places that use Quat.fromPitchYawRoll() that assume degrees...
but I think we want to do this sooner rather than later… I'd prefer to keep the C++ and the JS in sync for units of measure for these classes...
I agree it can be two different commits, but let's not leave them divergent for long…. I think that will be too confusing for those of us who move between the two...
Dimitar Dobrev
@ddobrev
Mar 11 2014 00:18
@birarda - I've pushed.
Andrew Meadows
@AndrewMeadows
Mar 11 2014 00:25
@ddobrev, very cool that you got XMPP chat working
Stephen Birarda
@birarda
Mar 11 2014 00:26
okay @ddobrev - leo will have to merge that once he has all of the build boxes set up
but I left a comment so he knows he can when he’s ready
I sent out an email to the team to let them know that they will need qxmpp
Dimitar Dobrev
@ddobrev
Mar 11 2014 00:26
@AndrewMeadows - thanks. There is a little more work to do but it's usable.
Stephen Birarda
@birarda
Mar 11 2014 00:26
and will update the build guide now
Dimitar Dobrev
@ddobrev
Mar 11 2014 00:26
@birarda - thank you.
Is there anything else for now?
Stephen Birarda
@birarda
Mar 11 2014 00:28
nope!
Dimitar Dobrev
@ddobrev
Mar 11 2014 00:28
OK, I'm going then. Really late here. :)
Have a nice day, fellows!
Stephen Birarda
@birarda
Mar 11 2014 00:30
thanks for all of the hard work today!
Leonardo Murillo
@murillodigital
Mar 11 2014 15:45
@ddobrev around?
Dimitar Dobrev
@ddobrev
Mar 11 2014 15:45
@murillodigital - yep.
Leonardo Murillo
@murillodigital
Mar 11 2014 15:45
so it failed this time because the library is named differently when built in debug mode
qxmpp_d.lib
update cmake’s find module to look for that file name as well
pls
Dimitar Dobrev
@ddobrev
Mar 11 2014 15:46
OK.
Dimitar Dobrev
@ddobrev
Mar 11 2014 15:58
@murillodigital - it cannot find it still. I added 'qxmpp_d', without the '.lib'. CMake ignores extentions, I think, when looking for files.
extensions*
Or perhaps the build took a little longer to start, there's one running now.
Leonardo Murillo
@murillodigital
Mar 11 2014 16:01
yeah i think there was another build in the queue ahead of yours
lets wait for this one
Leonardo Murillo
@murillodigital
Mar 11 2014 16:10
success @ddobrev
Dimitar Dobrev
@ddobrev
Mar 11 2014 16:12
@murillodigital - great.
Stephen Birarda
@birarda
Mar 11 2014 16:31
@ddobrev @murillodigital so the find module only looks for the debug lib right?
for windows
ahh it looks in order
cool
Dimitar Dobrev
@ddobrev
Mar 11 2014 16:32
@birarda - don't merge yet.
I need a minute to check a small fix.
Stephen Birarda
@birarda
Mar 11 2014 16:32
no problem
Dimitar Dobrev
@ddobrev
Mar 11 2014 16:39
@birarda - it would take a little longer. I've found the fix but I'm minimising the amount of necessary code.
I'll let you know when it's good to go.
Stephen Birarda
@birarda
Mar 11 2014 16:41
sounds great!
Dimitar Dobrev
@ddobrev
Mar 11 2014 16:55
@birarda - pushed.
Remember that bug from last night about messages at the bottom being slightly cut off?
Stephen Birarda
@birarda
Mar 11 2014 16:55
yeah
you fixed?
am going to wait for build and then I can merge
Dimitar Dobrev
@ddobrev
Mar 11 2014 16:55
I believe it's fixed now, yes.
I've found a way to reproduce it with multi-line messages at the top. When the bottom was reached, those were cut off.
I think it was the same bug.
So it should be fixed.
Stephen Birarda
@birarda
Mar 11 2014 17:12
@ddobrev I’m going to wait to merge that until a couple of hours machines in office are done building qt5 from source
so that they can get buidling right away when that is done
that alright?
Dimitar Dobrev
@ddobrev
Mar 11 2014 17:14
@birarda - I cannot see new commits which means I can start working on those two last missing features right away, without the possibility of conflicts.
So it's OK.
Stephen Birarda
@birarda
Mar 11 2014 17:15
I’m not sure I understand what you mean
Dimitar Dobrev
@ddobrev
Mar 11 2014 17:16
@birarda - I'd rather have started the remaining work after merging, to avoid possible conflicts.
But I can see there are no new commits except mine, so I can start now anyway.
So I have no problems with you merging a little later.
Stephen Birarda
@birarda
Mar 11 2014 17:17
alright - should be within the hour or so
Dimitar Dobrev
@ddobrev
Mar 11 2014 17:29
@birarda - you said you were going to update the Build.md, right?
Stephen Birarda
@birarda
Mar 11 2014 17:29
yep
Dimitar Dobrev
@ddobrev
Mar 11 2014 17:30
Could you please let me know when you're about to do so, I'd like you to insert one more note.
Stephen Birarda
@birarda
Mar 11 2014 17:32
I have made some changes already
what would you like to add
it’s up in a pull request
Dimitar Dobrev
@ddobrev
Mar 11 2014 17:33
I use a CMake function to keep translations up to date.
QT5_CREATE_TRANSLATION
But it has a bug: when cleaning, it deleted the *.ts files.
It's been reported here:
but not fixed yet - apparently the maintainer fails to understand the problem.
So, you know, it'd be nice to have some note like "after cleaning, reset your ts files from Git".
There's also some patch for the problem in that bug report, the note may include it as an option.
Stephen Birarda
@birarda
Mar 11 2014 17:36
just a sec - am on something else but want to make sure I understand
Dimitar Dobrev
@ddobrev
Mar 11 2014 17:36
No problem, I'm here.
Stephen Birarda
@birarda
Mar 11 2014 17:44
okay @ddobrev what do we need to indicate exactly?
basically after ‘make clean'
the ts files are gone
Dimitar Dobrev
@ddobrev
Mar 11 2014 17:45
Correct.
Stephen Birarda
@birarda
Mar 11 2014 17:45
so you have to check them back out
Dimitar Dobrev
@ddobrev
Mar 11 2014 17:45
Right.
Stephen Birarda
@birarda
Mar 11 2014 17:45
is there any workaround we could add to our CMakeLists?
Dimitar Dobrev
@ddobrev
Mar 11 2014 17:46
I don't think so but then again, I'm not that familiar with CMake, as I said.
We could use the patch to update the Qt module itself.
Or we could add comments there requesting that they fix it for the next version, such as 5.2.2.
I've already written a comment myself.
Stephen Birarda
@birarda
Mar 11 2014 17:47
alright - I’ll add something to the build guide
there’s one more thing for your pull
since qxmpp is an external dependency and can easily be made optional
we should ifdef around the chat code based on the presence of the qxmpp library
same thing we do for oculus for example
and basically leave the menu option grayed out
now - that’s outside the scope of what Grayson asked for, so if you feel like that’s a lot of extra work then I am happy to do that myself after this gets merged
if you feel like it’s ten minutes of work and you can add that as one more commit to your pull, it would be nice to not break anybody’s build right now
up to you
Dimitar Dobrev
@ddobrev
Mar 11 2014 17:50
I think it wouldn't take me too long. Could you just please tell me what symbol I should use?

ifdef what?

Stephen Birarda
@birarda
Mar 11 2014 17:50
define something in interface CmakeLists
Dimitar Dobrev
@ddobrev
Mar 11 2014 17:50
Ooops.
Stephen Birarda
@birarda
Mar 11 2014 17:50
HAVE_QXMPP
if QXMPP is found after the call to find package
you can see examples with libovr and sixense
Dimitar Dobrev
@ddobrev
Mar 11 2014 17:50
OK.
Stephen Birarda
@birarda
Mar 11 2014 17:50
and then look for that in the code
Stephen Birarda
@birarda
Mar 11 2014 18:00
@ddobrev you will also change find_package(qxmpp REQUIRED) to find_package(qxmpp)
so it doesn’t stop building if it doesn’t find it
Dimitar Dobrev
@ddobrev
Mar 11 2014 18:00
@birarda - that's OK, I got it.
I copied LibOVR and changed the names.
Dimitar Dobrev
@ddobrev
Mar 11 2014 18:23
Pushed.
Stephen Birarda
@birarda
Mar 11 2014 18:23
awesome - thanks
I’m going to remove qxmpp from my machine and test locally
couple of errors @ddobrev
XmppClient.h is not finding QXmppClient (evidently)
also, ChatWindow.h has nothing defined in the header wihtout QXMPP, which produces an error
Dimitar Dobrev
@ddobrev
Mar 11 2014 18:43
@birarda - Working on it.
@birarda - fixed the XmppClient but I don't get an error with ChatWindow.h.
Stephen Birarda
@birarda
Mar 11 2014 18:47
I just updated command line tools and Xcode this morning
so it’s possible that’s new with newever version of compiler
but it seems to not like an empty .h file
Dimitar Dobrev
@ddobrev
Mar 11 2014 18:47
@birarda - I see.
Stephen Birarda
@birarda
Mar 11 2014 18:47
let me show you exact error
ahh
it’s actually the automoc
the shell script that wants to run the moc on ChatWindow bails since there are “no relevant classes found"
so I suppose if we don’t have QXMPP just exclude it from interface srcs in the cmakelist?
Dimitar Dobrev
@ddobrev
Mar 11 2014 19:15
@birarda - I did it another way but the build is fine over here.
Stephen Birarda
@birarda
Mar 11 2014 19:16
clang —version is what?
Dimitar Dobrev
@ddobrev
Mar 11 2014 19:16
425.0.28
I mean that I've included the header back.
Stephen Birarda
@birarda
Mar 11 2014 19:17
yeah just guessing that clang is causing the difference
Dimitar Dobrev
@ddobrev
Mar 11 2014 19:17
But I didn't remove it from the sources, just #ifdef-ed the offending pieces.
So it should now work for you too.
Stephen Birarda
@birarda
Mar 11 2014 19:19
great - will try in a sec
Dimitar Dobrev
@ddobrev
Mar 11 2014 19:58
@birarda - so, any problems?
Stephen Birarda
@birarda
Mar 11 2014 19:59
sorry - stuck on something else right now
checking now
Stephen Birarda
@birarda
Mar 11 2014 20:07
awesome!
okay so the only annoying thing is the ts file
is there a way we can sort of hack that by making a copy that we pass to QT_CREATE_TRANSLATIONS
and that way it is copied and the copy is deleted
and nothing happens to the working dir
because otherwise that’s going to confuse some devs
Dimitar Dobrev
@ddobrev
Mar 11 2014 20:09
To avoid changing it when qxmpp is missing?
Stephen Birarda
@birarda
Mar 11 2014 20:09
no to avoid the ts file that is in the repo being deleted
Dimitar Dobrev
@ddobrev
Mar 11 2014 20:10
I cannot think of such a hack.
It's only deleted on explicit clean.
So I think it would be better to apply that fellow's patch to our Qt.
Or have them fix the issue by spamming them with comments.
Stephen Birarda
@birarda
Mar 11 2014 20:11
it’s being deleted for me on a fresh build of cmake
the issue is that somebody is going to miss it
and then git commit all
I’d rather a solution that doesn’t involve all of us patching our Qts
alternatively we could copy the patched qt5_create_translation macro to our repo
as a module
Dimitar Dobrev
@ddobrev
Mar 11 2014 20:13
Well, the other way would be to copy the function, rename it, apply the guy's patch and use that.
Stephen Birarda
@birarda
Mar 11 2014 20:13
hah - yep, we just said the same thing
let’s do that
Dimitar Dobrev
@ddobrev
Mar 11 2014 20:13
Yes, so the same.
Stephen Birarda
@birarda
Mar 11 2014 20:13
patched_qt5_create_translation
or whatever
in the modules folder
Dimitar Dobrev
@ddobrev
Mar 11 2014 20:15
OK, I'll try it.
Stephen Birarda
@birarda
Mar 11 2014 21:23
@ddobrev how is that going?
Dimitar Dobrev
@ddobrev
Mar 11 2014 21:23
@birarda - had serious trouble with the patch and CMake in general. I never should've taken it.
Building now, to see if it works at all.
Stephen Birarda
@birarda
Mar 11 2014 21:24
hmm
did the patch need some other qt pieces?
if the patch isn’t possible I would say we can also have cmake copy the ts file to a duplicate we can hand off to create transliations
translations*
Dimitar Dobrev
@ddobrev
Mar 11 2014 21:25
This is what the patch essentially does.
But it messes the file names up.
Stephen Birarda
@birarda
Mar 11 2014 21:37
well, let me know if I can help at all
Leonardo Murillo
@murillodigital
Mar 11 2014 22:05
@ddobrev those fails are false, they have to do with GitHub’s current technical difficulties
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:05
@murillodigital - what fails, did the build fail?
Leonardo Murillo
@murillodigital
Mar 11 2014 22:06
yeah i think it was coming from you
but its not the build failing, just unable to checkout the commit
since github’s busted
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:07
@murillodigital - yeah, I cannot push over here and then tried their page. They blew up.
@birarda - please try now, it should be fixed.
Stephen Birarda
@birarda
Mar 11 2014 22:11
building now
want to jump in and test chat with me @ddobrev ?
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:12
OK.
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 22:21
oh hey! it worked this time
hello people I haven't spoken to in a about ~8 months I think?
Stephen Birarda
@birarda
Mar 11 2014 22:21
welcome back Geenz
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 22:21
welcome back Geenz!
Stephen Birarda
@birarda
Mar 11 2014 22:22
I think your return has broken github
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 22:22
D:
didn't d oit
it was that weird octocat thing's fault, I swear
Grayson Stebbins
@problem
Mar 11 2014 22:24
:P
Stephen Birarda
@birarda
Mar 11 2014 22:27
@ddobrev mind explaining that create vs add translation issue to me here?
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:27
@birarda - not at all but about the window first.
You can see me at the bottom.
They've promised to fix it for Qt 5.2.2.
The same happens with the metavoxel editor, by the way.
Thanks for your comment at the translation issue. That unfortunate namesake of yours hasn't even bothered to reply. Let's hope we can spam him to death. :)
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 22:32
fork damn you fork!
Stephen Birarda
@birarda
Mar 11 2014 22:32
alright - good to know about the dialog
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:32
@birarda - I saw it today myself. I copied text from Firefox to test the messages and saw that the window stayed on top of the browser.
So I sent them a little something.
Now about the create vs add translation:
You know about ts and qm files, right?
Stephen Birarda
@birarda
Mar 11 2014 22:34
nope - haven’t played with them at all
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:34
OK.
In short, Qt translates the following way: it runs ldupdate on your code and extracts everything from tr()/QCoreApplication::translate calls and *.ui files.
It generates *.ts files for it, one per language.
Stephen Birarda
@birarda
Mar 11 2014 22:36
so which file needs to be checked into the repository
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:36
Then you edit that with Qt Linguist to fill your translation in, run ldrelease on the saved .ts and get a binary .qm file that is used at run-time.
Right, ts needs to be stored because it contains developers' work.
Stephen Birarda
@birarda
Mar 11 2014 22:37
and qt5_create_translation is called on the ts file?
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:38
It is passed these files and all source dirs.
Any ts that hasn't been created, is.
The rest are updated.
Stephen Birarda
@birarda
Mar 11 2014 22:38
in this case the ts file already exists - correct
?*
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:38
So, the difference between those create and add is that create starts from sources and finishes with qm-s, while add starts with ts - that is, you need to have them already.
Stephen Birarda
@birarda
Mar 11 2014 22:39
right - but for our use case we already have the ts don’t we?
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:39
We have it because I had it generated by create.
We didn't ahve it before that.
have*
Stephen Birarda
@birarda
Mar 11 2014 22:39
right but we don’t need to call create every run
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:40
We do.
Despite its name, it also updates ts-es.
Stephen Birarda
@birarda
Mar 11 2014 22:40
in case a translation is added to the ts file?
sorry - to the source and the ts file
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:41
create adds a new entry for the thing in the source and keeps whatever translations are already in.
Stephen Birarda
@birarda
Mar 11 2014 22:42
alright
sounds good
test completed
just need to you pull from upstrea master
upstream*
and push that to your branch
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:42
Yes, it's brilliant, save for that bug.
OK, merging.
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 22:43
any of you guys using a brew'd version of Qt? Or are you just grabbing it from the Qt website?
Stephen Birarda
@birarda
Mar 11 2014 22:43
we use a patched version
that removes a pesky wifi scanning bug
you can tap our formula repo
brew tap highfidelity/hombrew-formulas
and then brew install highfidelity/formulas/qt5
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 22:44
oh, awesome, thanks!
Stephen Birarda
@birarda
Mar 11 2014 22:44
that’s going to take a while
it builds from source
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 22:44
oh goody
I remember how long that takes
guess I'll grab some grub and try to de-stress after today's events while I wait
brew says that tap's invalid
Stephen Birarda
@birarda
Mar 11 2014 22:47
oh
brew tap highfidelity/homebrew-formulas
typo
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 22:48
there we go!
also this new chat thing's nice
Stephen Birarda
@birarda
Mar 11 2014 22:48
yeah - helpful for sure
geenz have a look at BUILD.mb for the other deps you’ll need
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 22:48
yeah, I saw; thankfully I keep a copy of glm and zlib on hand
though for glm, have you guys considered just doing a submodule of that? I know g-truc keeps a github repo of his stuff
and he branches the releases as well
Stephen Birarda
@birarda
Mar 11 2014 22:50
hadn’t really thought about it but I think as much as possible we’d like to keep external deps outside the repo
submodules also introduces a bit of a learning curve for beginner git users who don’t know how to use em
more difficult than homebrew imo
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 22:51
eh, depends on the SCM client; of all of the ones I've used thus far, source tree handles it in the most elegant (and "automagic") way
Stephen Birarda
@birarda
Mar 11 2014 22:51
ironically a bunch of our beginner git users are command line users
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 22:52
....
lolwut
that's kinda funny actually
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:54
@birarda - let me ask you, do you have an idea about sorting users by time online?
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 22:54
and now I wait for Qt to do a fresh compile
Stephen Birarda
@birarda
Mar 11 2014 22:54
@ddobrev if there’s nothing in xmpp that gives us that
I think @problem would be okay not doing that
otherwise we’d have to have some sort of bridge where we keep track of time online
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:55
I've been unable to find anything, at least in their API.
@problem - may I ask what time online did you have in mind: online in the chat or online in the world?
Stephen Birarda
@birarda
Mar 11 2014 22:56
we need to separate the two concepts - we want people to be able to be plugged in with a chat client and not necesarily in interface
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:56
I see.
So XMPP would have to have something.
Stephen Birarda
@birarda
Mar 11 2014 22:57
yep
they have last user login but I believe that’s for admins only
Dimitar Dobrev
@ddobrev
Mar 11 2014 22:58
By the way, if you want, you could add your two cents to the tool window on top bug as well - https://bugreports.qt-project.org/browse/QTBUG-36082 .
Stephen Birarda
@birarda
Mar 11 2014 22:58
if they’ve already said they’re fixing it I’m alright with that one
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 22:59
sorry to keep jumping into the chat here, but is there a change log somewhere that isn't the commit log that I can use to get up to speed on what's been going on for the past 8 months?
Stephen Birarda
@birarda
Mar 11 2014 22:59
hmmm
I don’t think we have that succintly documented anywhere
there have been a ton of changes
@ddobrev if we really needed that feature perhaps it is an extension we could write into our server
but I don’t think it’s that important
Dimitar Dobrev
@ddobrev
Mar 11 2014 23:04
@birarda OK. @problem seems to be away and it's a little late here. Could you please discuss the issue with him and send me a mail about it afterwards?
Stephen Birarda
@birarda
Mar 11 2014 23:04
not a problem, don’t worry about that for now
let’s just get that pull up to date and merge it in
Dimitar Dobrev
@ddobrev
Mar 11 2014 23:05
I thought it had been already merged.
Stephen Birarda
@birarda
Mar 11 2014 23:05
@Geenz feel free to ask anything specific if you can think of anything
I can’t merge it since it’s not up to date with HEAD
Dimitar Dobrev
@ddobrev
Mar 11 2014 23:06
Oh, I see.
Didn't get that earlier. Just a minute.
Stephen Birarda
@birarda
Mar 11 2014 23:06
sorry, I may have been unclear
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:07
just don't want to interrupt too much here while I try to get caught up is all
so let's see here, last thing I remember is I submitted a patch to handle sky calculations per-pixel instead of per-vertex, I think that got merged in
oh, and rigged voxel avatars had just recently became a thing
Stephen Birarda
@birarda
Mar 11 2014 23:08
hah
well
avatars are different now
we use fst models
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:08
fst models? that's a new one
Stephen Birarda
@birarda
Mar 11 2014 23:08
basically an abstraction around a mesh for your body and face
fst is a faceshift specific type but we’re leveraging so that your “model” can have multiple levels of detail
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:09
ah
Stephen Birarda
@birarda
Mar 11 2014 23:09
but effectively an fbx mesh
you’ll get a default one when you run the client
and there’s a directory full of ones you can try out
we’re now using assigned servers @Geenz
so before you had a target for each one
instead now the domain-server assigns an assignment-client something to do
which can include audio-mixer, avatar-mixer, voxel-server, particle-server, metavoxel-server
and some others
so you have three major targets now
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:11
if it ever comes up in the future, I'd actually kind of recommend the GLTF format; it falls more under the category of "hardware optimized" in that it's pre-indexed ready for pushing to index and vertex buffers so it may not 100% have everything you need, but the guys at Khronos actually released open source libraries for the format
Stephen Birarda
@birarda
Mar 11 2014 23:11
interface, domain-server, assignment-client
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:11
hm
well that's neat
Stephen Birarda
@birarda
Mar 11 2014 23:12
we’re just about to add chat to the client today
we’ve got a working windows build now too
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:12
particle-server and metavoxel-server are new, I believe I remember voxel-server, and I definitely remember the avatar and audio mixers
Dimitar Dobrev
@ddobrev
Mar 11 2014 23:12
@birarda - is OctreeStats and VoxelStats one and the same?
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:12
oh awesome
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 23:12
@ddobrev - yes
Stephen Birarda
@birarda
Mar 11 2014 23:13
@ddobrev there was a pull request earlier today to remove VoxelStats I believe
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 23:13
assuming you're talking about VoxelSceneStats… that's gone… it was just a wrapper for OctreeSceneStats
Stephen Birarda
@birarda
Mar 11 2014 23:13
@stojce you here?
Stojce Slavkovski
@stojce
Mar 11 2014 23:13
yes
Brad Hefta-Gaub
@ZappoMan
Mar 11 2014 23:13
removed as Stephen mentions in last couple days
Stephen Birarda
@birarda
Mar 11 2014 23:14
you shouldn’t post username to /api/v1/places
that’s implicit with whoever owns the access token, no?
Stojce Slavkovski
@stojce
Mar 11 2014 23:14
k, I guess
Stephen Birarda
@birarda
Mar 11 2014 23:14
so let me go look at why you got the 500
what did you post with?
Stojce Slavkovski
@stojce
Mar 11 2014 23:15
tried with postman for chrome
Stephen Birarda
@birarda
Mar 11 2014 23:15
was content-type for sure json
Stojce Slavkovski
@stojce
Mar 11 2014 23:17
x-www-form-urlencoded
Stephen Birarda
@birarda
Mar 11 2014 23:18
you should be posting application/json
{
"access_token": "e34fb08f26a505278547e28ea3387dea4d45147b06f1cbf10deb03d5357256bd",
"address": {
"position": "0,0,0"
},
"name": "invite"
}
there’s an example of the json in the body
but maybe x-ww-form-urlencoded also works
can you fire it now and I can catch the 500?
ahh, I see
Stojce Slavkovski
@stojce
Mar 11 2014 23:21
posted, did you catch it?
Stephen Birarda
@birarda
Mar 11 2014 23:21
Parameters: {"access_token"=>"b920fbc5d1b453c6b5b1262934f97393a27a1e4677c51a65e7a1a60a5d526e23", "name"=>"YAS", "position"=>"-123.704,111.453,75.2499", "orientation"=>"4.65985e-07,-66.8316,4.27653e-07", "domain"=>"test.example.com"}
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:21
I assume I shouldn't be getting any output after brew hits make on Qt
Stephen Birarda
@birarda
Mar 11 2014 23:21
you want position, orientation and domain to be underneath an address key
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:21
or is that a bad sign
Stephen Birarda
@birarda
Mar 11 2014 23:21
not for a while @Geenz
no that’s right
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:22
okay, cool
Stephen Birarda
@birarda
Mar 11 2014 23:22
until it tells you otherwise it’s alright
@stojce position, domain, orientation should be below a key ‘address'
my bad, I wrote that wrong on the worklist job
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:22
I remember Qt 4.x would give you a seemingly endless log
Stojce Slavkovski
@stojce
Mar 11 2014 23:25
@birarda is this right format?
{
"access_token": "b920fbc5d1b453c6b5b1262934f97393a27a1e4677c51a65e7a1a60a5d526e23",
"adress": {
"name":"test1",
"position":"1,1,1",
"orientation":"1,1,1,1",
"domain":"localhost"
}
}
Stephen Birarda
@birarda
Mar 11 2014 23:26
adress -> address
and name is not beneath address
{
"access_token": "e34fb08f26a505278547e28ea3387dea4d45147b06f1cbf10deb03d5357256bd",
"address": {
"position": "0,0,0"
},
"name": "invite"
}
that only has a position but that’s an example
Stojce Slavkovski
@stojce
Mar 11 2014 23:26
getting 401
Stephen Birarda
@birarda
Mar 11 2014 23:27
that access token won’t work - it’s from my local dev
use your correct one
Stojce Slavkovski
@stojce
Mar 11 2014 23:27
that’s mine token
in mine example ...
Stephen Birarda
@birarda
Mar 11 2014 23:28
sorry - I assumed you copied my whole example
Dimitar Dobrev
@ddobrev
Mar 11 2014 23:28
Fellows, is your project open source?
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:28
I need my password reset! And apparently worklist says no
Stephen Birarda
@birarda
Mar 11 2014 23:28
hit it again am watching log
@ddobrev we haven’t nailed down licensing yet
which is why you see the all rights reserved license
Dimitar Dobrev
@ddobrev
Mar 11 2014 23:29
@birarda I see. It's in a public repo, though, so I was wondering.
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:32
yay, Qt finished
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:44
so should qxmpp be trying to nab Qt4?
Stephen Birarda
@birarda
Mar 11 2014 23:44
nope
brew install highfidelity/formulas/qxmpp
use the tapped one
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:49
getting this:
==> Downloading https://qxmpp.googlecode.com/files/qxmpp-0.7.6.tar.gz
Already downloaded: /Library/Caches/Homebrew/qxmpp-0.7.6.tar.gz
==> qmake -config release PREFIX=/usr/local/Cellar/qxmpp/0.7.6
==> make
make: * No targets specified and no makefile found. Stop.
Stephen Birarda
@birarda
Mar 11 2014 23:49
well that’s a new one
can you do
qmake —version
and tell me what version of Qt it says it’s using?
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:51
ruhroh:
Geenzs-iMac:hifi geenz$ qmake -version
QMake version 2.01a
Using Qt version 4.8.4 in /Library/Frameworks
time to search and destroy!
Stephen Birarda
@birarda
Mar 11 2014 23:52
yeah
I should also fix that formula to prefer a qt5 qmake
I could*
Dimitar Dobrev
@ddobrev
Mar 11 2014 23:53
@birarda - pushed.
Stephen Birarda
@birarda
Mar 11 2014 23:53
awesome @ddobrev as soon as this builds it’s going in
sorry you’re up late again
actually @Geenz I guess it shouldn’t be our responsibility in the individual qxmpp formula to pick which qmake we want
we just want to make sure there is one and then use it
so you’ll need to kill that with fire (or at least get it out of your path) first
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:55
yeah, looks like I have an old version of the whole Qt package that I forgot about installed from ages ago
Stephen Birarda
@birarda
Mar 11 2014 23:56
it’s certainly possible you were working with us at a time when we weren’t even requiring qt 5
Jonathan "Geenz" Goodman
@Geenz
Mar 11 2014 23:58
yes actually, that'd be accurate
at that time if I recall correctly, Philip mentioned that Qt 5 was on the roadmap but no one was ready to begin updating things