IRC Log Viewer » #firebreath » 2010-10-05

IRC Nick Time (GMT-7) Message
taxilian 09:10 crap. crum. where is Georg when I need him =]
kalev 10:10 there's some amazing activity on this irc channel, I cannot keep up any more :)
taxilian 10:10 lol
it's been getting fun =]
kalev 10:10 looks like you've gotten the project really going now
nice work
taxilian 10:10 now if cygmatic will show up… I need his help to fix a problem with building on linux =]
oh, it's picking up the pace a bit, yeah
how are things with you? anything new on that job?
kalev 10:10 I talked to a Red Hat guy and he said that the HR is still getting organized
not sure what that means, I'm really never had to sell myself like that before
but in the mean time I'm pretty busy as my old job
taxilian 10:10 fair 'nuff
hope all goes well
kalev 10:10 thanks
cygmatic 11:10 taxilian: still need help with that stuff
taxilian 11:10 yes
yes I do
it's in dev; can you pull it down?
I'm hoping it's easy
cygmatic 11:10 will do
taxilian 11:10 (for you =])
getting "no matching function for call to…" on those templates
also we have the interdependency issue so I had to move the BrowserHost function templated implementations into the other file :-(
I spent about 45 minutes on it, but now I have to finish my statistics homework :-(
taxilian 12:10 cygmatic: anything? still looking at it? looking at it later? just wondering
cygmatic 12:10 taxilian, looking... major problem fixed though
(don't try to pass temporaries to functions expecting non-const references)
taxilian 12:10 was I doing that? oops
cygmatic 12:10 also, for future stuff: you can return expressions of void type from void functions, but not non-void expressions
taxilian 12:10 yeah, I realized that partway through
just didn't go back and clean up like I should have
cygmatic 12:10 i don't like commiting just partially working stuff, so... :)
taxilian 12:10 lol. I didn't realize it was partially working until late last night
the build server wasn't automatically building
at least it was in dev :-/
should have done it in another clone
actually, I commit partially working stuff all the time; I just don't push it usually until it works =]
cygmatic 12:10 pushed now
taxilian 12:10 all working?
cygmatic 12:10 for me it builds
euh, no
taxilian 12:10 very niceish
cygmatic 12:10 one thing remaining somewhere else
taxilian 12:10 ahh
I'll wait to fire off test builds on the server, then
cygmatic 12:10 the libs build but not the examples/unit tests
taxilian 12:10 I need to just start developing on Mac; when it builds on mac, it generally builds on windows as well
hmm. yeah, those are probably my fault
to explain quickly, since I don't know how to fix them
I made PluginEventSink a cast-able object type
with ptr_cast
maybe I didn't do that correctly?
cygmatic 12:10 fbtestplugin was a reference vs. const ref
second i'm looking at
taxilian 12:10 I have a dream that someday I'll be able to fix things like this myself =]
until then, I think I'll just try to avoid being the one to implement them :-P
but that should be the last really tricky one for awhile
cygmatic 12:10 my ptr_cast only checked for downcasts, not for upcasts
didn't think of that as you don't need the dynamic cast there
i.e. shared_ptr<Base>(derivedPtr) is fine
taxilian 12:10 oh, really? didn't realize that
I thought you always had to use cast to go either direction
cygmatic 12:10 only downcasts, up-casts can be checked at compile-time
and an incomplete type problem... the types used in is_base_of<> may not be incomplete
you may want to solve that include problem more nicely
for plugineventsink.h
taxilian 12:10 which problem?
cygmatic 12:10 for is_base_of<T,U>, both T & U can't be incomplete at the point of instanciation
which was the case in a unit-test with PluginEventSink
taxilian 12:10 ahh
I can't even figure out why it needs PluginEventSink there...
I see
cygmatic 13:10 hmm? ^^
taxilian 13:10 cygmatic: my vote is that we just remove the check and let them do ptr_cast however they want
cygmatic 13:10 and blame the user if he relied on it using dynamic_pointer_cast?
taxilian 13:10 yeah
I just don't see an easy way to solve this other than that
and that way we can make use of shared_ptr however we want
I don't see us changing away from boost at this point
it would require too much change and I can't imagine a good enough reason
cygmatic 13:10 shall i strip it or will you?
taxilian 13:10 if you could, that would be great
I am trying not to get too distracted from my homework right now =]
it was due today… now I'm trying to make sure I'm ready for the test, since I'm not going to get it turned in on time
hmm. in fact, I better go eat lunch; I'll be back in a bit
thx for your help
cygmatic 13:10 you're welcome & good luck
taxilian 14:10 cygmatic: hey, you fixed it!
cygmatic 15:10 yep, all fine on hudson now, just the xp build server was offline
taxilian 15:10 yeah; still trying to figure out how to fix that
the slave daemon keeps having errors and stopping
taxilian 15:10 cygmatic: I'm getting an error compiling on mac with "-read_only_relocs suppress" being needed; have you ever seen that?
I only saw it when I built from outside of the plugin root: -read_only_relocs suppress
../firebreath-dev/ ../firebreath-dev/examples buildex
taxilian 15:10 ooh
prepmac does not work correctly when run on something outside of the plugin dir
yeah, that'd do it
cygmatic 15:10 will try to reproduce later
cygmatic 16:10 i can't reproduce neither that nor travis' problem