IRC Log Viewer » #firebreath » 2012-12-07

IRC Nick Time (GMT-7) Message
diorcety 07:12 Hi
Why it is mandatory to return VariantList for list of object?
jshanab 07:12 I think because javascript can have a variable be a single item or a list???
jshanab_ 08:12 taxilian. Good morning FYI boost 1_46 works. Anything after that has issues(not all the same) for me.
taxilian 08:12 interesting
jshanab 09:12 I am moving on for now and I see no reason to even try to fix 1_47 - 1_50. It is that has_binary_operator macro clash that I will have to fix for the future. (>=1_51)
This is osx specific and from what i have seen will need to include the boost before cocoa.h. I tried that in a few places but it did not work :-(
taxilian 09:12 hmm. interesting
did you ever try FBTestPlugin with those versions?
jshanab_ 09:12 I did try it on a few. I could test more later for the unable to convert pointer error found in 1_47. The problem with the macro clash does not show up with the FBTestPlugin because it is not the use of boost. It is the use of boost in conjunction with cocoa and maybe even something I am doing.
ok,ok Probley something I am doing ;-)
BTW, Do you know if there is a way to tell Xcode to open the issue navigator with the tree initially closed?
taxilian 09:12 I've never used the issue navigator
ahh
interesting
I'm using boost 1.50 on ios
jshanab 09:12 oh. Every time i build the left hand pane in xcode 4.5.2 changes to it
taxilian 09:12 I probably just don't know which one it is
I don't always bother to learn the names of all the panes
jshanab_ 09:12 As soon as you learn it, they will change it
(or patent it and sue you for saying outloud)
taxilian 09:12 lol
no worse than anyone else, reallyi
I think I have most of the next week to do firebreath and plugin stuff; some of that needs to be on our plugin specifically, but hopefully I can track down a few issues that have been around for a bit
jshanab_ 09:12 Xcode 4.5.2 has some pretty major bugs that 4.3.3 did not have. I added two cpp files, ran prep and then reopened Xcode and hit build. It had added the cpp files to the project but did not even try to build them. I have to clean and sometimes delete derived data for them to be seen. crazy
taxilian 09:12 the whole derived data thing is kinda annoying
jshanab_ 09:12 It used to have "intellisense" data only, now it contains all the errors and build products so it can show you errors when you open (even if you deleted the build directory)
taxilian 09:12 hehe. makes it difficult to build things with a build server too because the output of a real build now goes into the deriveddata area
there has got to be a way around that, but I haven't yet found it
havne't really had time to try either, though
jshanab_ 09:12 Yeah. Same here. Gotta maintain the seperation between source control, temp products, and intellisense. MS dumps everything in folders that should be for source control, Xcode does it's thing. neither of them right.
taxilian 09:12 in a lot of ways that's one big win for cmake, IMO; I can keep my build stuff entirely seperate
jshanab_ 09:12 Oh no. The boost error is back now that I got furthor along in the build.
it is the "non-const lvalue referenced" in MethodConverter.h (on this line "BOOST_PP_REPEAT(50, _FB_METHOD_WRAPPER, BOOST_PP_EMPTY())" )
jshanab_ 10:12 A feature I would like to see in VS and Xcode is a "compiler stack trace". The above error is a good example, It does not trace back to the source line that called for it.
taxilian 10:12 hmm. that is really odd
actually the trace will show you where it is
but it's a little tricky to read
jshanab_ 10:12 It is latest FB but with system boost 1_46
trace?
taxilian 10:12 okay, I'm going to switch computers. be back in a bit, as soon as I manage to crawl downstairs (sprained my ankle, can't walk)
the compiler errors give you what you need to track it down
but it's hard
be back in a bit
jshanab_ 10:12 I get nothing. Ouch.
taxilian 10:12 hey, I made it!
jshanab_ 10:12 It looks like it is totally self contained and the only person calling this fn is dispatch_gen.py
Gotta watch out for those toddler landmines, toys with wheels that just fit under foot.
taxilian 10:12 lol. well, this was actually from playing ultimate frisbee
jshanab_ 10:12 ultimate frisbee. That is like football with frisbees right?
taxilian 10:12 more or less
okay… working on FireBreath. Where to start?
jshanab_ 10:12 They had baskets hanging on every lightpole for a while at my University so people could play frisbee golf. They finally took them down after a few more important people got hit in the head
I need to solve one of two boost issues and it looks like I am gonna need your help. Choices are 1_47 invalid lvalue or 1_52 macro collision in has_binary_operator :-)
taxilian 10:12 I need a way to reproduce it
if it doesn't happen on FBTestPlugin
tempted to try updating to 1.52
jshanab_ 10:12 I will build the FBTestPlugin with 1_46 to see if the lvalue one pops up. I would really like to move fwd not backward. Best way to reproduce would be to let you remotly acces my machine!
I can change boost versions pretty fast :-)
What is MethodConverter.h for?
taxilian 10:12 MethodConverter.h is part of the autoconversion stuff used by JSAPIAuto
do you have skype?
jshanab_ 10:12 no, not yet but I was thinking of a join me session
taxilian 10:12 could try it. skype is what I use most often, but whatever
voice chat would be helpful along with it; odn'tk now if join.me has that
jshanab_ 10:12 Should I set up for the newer internal boost?
taxilian 10:12 whatever you want to look at first
might be interesting to start with 1.52 if you have it handyish; I'd really like to get the newer boosts working
but if it's not even working on older, that might be a place to start
jshanab_ 10:12 I think I tried updateing the boost in the FB internal to the 1_52. Let me check. Two different problems though. 1_48 and onward has the macro collision that really needs to be fixed. Let me re-prep
ok prepped for internal boost 1_50
building fresh. (deleted derived data and build directory before prep)
I don't have 1_52 on this machine, I think I re-pulled over it. Downloading it for later testing I will be glad to test that one on osx
taxilian 14:12 anyone else have a pet peeve firebreath issue they'd like me to particularly focus on? (I'm going through the jira tasks, of course, but having a hard time being motivated today)
diorcety 17:12 hi
why the return value for array have to be explicitly variantlist?