| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9 10 11 12 13
Comments:
<0> wait
<1> I don't think he has access to lower versions
<2> gcc --version
<1> those things are hard to find since the year 2003
<0> its 3.2
<2> schools arent that good with upgrading.. i dont know where he is what he uses etc
<2> ***umption is the mother of all ****up anyway
<1> codeblocks is new =P
<2> i dont know it :P
<1> so anyway, the error seems to be solveable
<1> as you said ***umptions are bad, so I'd want to see the code before solving.
<0> the syntax im using is right?
<1> maybe?
<0> template <typename T> Application Singletron<T>::singletron= 0;
<1> sometimes it is
<2> it would be right if the compiler at that time knows about Application and Singletron
<0> Application is a decendent of Singletron. Singletron is the template
<1> Application inherits from Singletron ?
<0> yeah, im placing after defining the 2
<0> yes
<2> so Application is something like: cl*** Application;Singletron<type> { }; ?
<1> so you want to do struct foo{ foo(int){} }; template <typename T> struct bar: foo{ static foo baz; }; template <typename T> foo bar<T>::baz= 0; // ?
<0> cl*** Application : public Singleton<Application> {...}
<2> SSPecteR cant you paste the code you have on the pastesite?
<1> no he can't
<1> he has been asked 3 times and he didn't, so he can't
<1> obviously
<2> http://pastebin.com/
<2> well I am stepping out until he does
<2> :P
<0> it will take some time :(
<0> wait
<2> I am at work.. why would i wait :P
<1> I only wait for nice chicks
<0> the code is a little big... im removing everything unrelevant
<1> it seems like you have some kind of a crtp singleton
<1> since the member depends on the tmplate, maybe you need an extra typename there
<1> template <typename T> typename Application Singletron<T>::singletron= 0; // or something
<0> ok pasted :P
<0> http://pastebin.com/559159
<0> the error:
<0> expected init-declarator before < token
<0> (line 60)
<2> typename T* Singletron<T>::singletron= 0;
<3> singlet*r*on
<2> _m_ been there done that :P
<2> a namespace would have been nice
<0> oh, **** :P
<1> err
<0> i fixed it not it gives other error
<1> yes, what _m_ said
<1> template <typename T> T* Singleton<T>::singleton= 0;
<1> where the hell did you make up "singletron" anyway =P
<2> i just pasted line 60 :P
<3> SSPecteR: did you fix all occurrences?
<0> yeah
<0> error:
<0> conflicting declaration 'Application Singleton<T>::singleton'
<2> Apperantly not
<0> i put everything Singleton
<1> he meant occurances of tron
<2> oh
<1> I guess I'm invisible.
<3> line 60 uses the wrong type for the member variable.
<3> As Tamama already told you.
<0> i removed all occurences of 'tron'. im reading ya :P
<2> reading half :P
<1> Tamama, is the bill from a nigerian nobility?
<2> heh no
<2> some silly moved domains
<2> they think they still have them
<0> now its working :))))
<0> thanks cynic for pasting it fixed :)
<2> -.-
<0> thanks Tamama :)
<0> u guys save my life
<0> i still didnt understood what hapen in this line though
<0> template <typename T> T* Singleton<T>::singleton= 0;
<2> you need to think from singleton
<2> singleton has nothing to do with Application
<2> Singletons singleton member is a T*
<0> ive searched a lot in the net about initializing static variables in templates, but no luck
<0> this line initializes all types i use with Singleton?
<2> right
<0> thanks again man
<0> time to get bak to work :(
<0> bye
<2> lates
<2> *cough*
<4> why do i have to derive CWinApp in order to create an interface thread?
<5> critical updates completed, gotta reboot
<2> hm
<2> when sending udp packets in a blocking while loop, windows 2k gets 1/6th less packets out than linux-2.4
<6> add a tiny sleep
<2> how will that make windows faster?
<7> linux faster than windows?
<7> check the tcp buffer settings
<6> it will help it not loose packets
<2> I am sending
<2> i an not receiving
<6> still, it loses them
<2> nope
<2> since it is blocking
<2> windows simply does not saturate the network
<7> why UDP instead of TCP?
<2> irrelevant
<6> put a counter, just in case. I saw once that when you send more than 200 small packets at once, you don't recieve all of them
<2> I am not interested in receiving
<2> in fact
<2> there is nothing listening to where i send it
<6> ok then
<2> since after all, i am measuring output
<6> mr32_melb : UDP is better than TCP for certain tasks
<2> like real time video.. :P
<2> anyway... any ideas on how to get more output on windows?
<6> maybe tweaking the registry to increase some interanal buffers ?
<6> internal
<5> Tamama some of the interesting problems I've seen w/ the 'doze network has to do when there is no other end
<2> (sidestep: the funny thing in this all.. the windows machine is more than 3 times as fast :P)
<2> vawjr: yeah it has a timeout then on udp lol
<2> and it becomes SLOOOOOOOW
<6> try two threads. not likely to help, though
<2> exception: and what would i put in the threads?
<5> depending on how the "other end" isn't there...there are differences betwneen no ip address at all, and ip, but being refused
<6> the same sending loop
<2> exception: you mean have 2 threads fill the buffer?
<2> instead of 1?
<2> I dont think that will matter much
<6> it depends on their implementation - maybe they have a buffer for every thread
<2> as my app is mostly blocking
<2> the buffer is per socket
<5> I can cause my system at work to grind to a halt if I up the "concurrent compilations" to 4 when I'm using the network drive to hold the boost includes
<2> heh
<2> it works with network folders as well.. :P
<6> so send from several sockets
<5> other machines can access the same files I cannot
<2> open 'my computer' * grind *
<5> and occasionally I then get compiler errors "file not found"
<2> vawjr: i have that regularly too
<5> which is bull****.....ergo, ****ed up code in Microsoft's stack
<5> it's been this way for 20 years, I don't know why people are surprosed
<5> or surprised
<5> Microsoft does NOT know how to write multi-tasking stuff, period.
<7> vawjr: example?
<2> I just set the send buffer size of the socket to 128KB (no idea what it is by default)
<2> but 128KB should be enough
<2> mattered none
Return to
#c++
or
Go to some related
logs:
#linux
#mirc
struts action goes nowhere
#teens
24-disc John Gabel Entertainer
#linux
bona serra
#AllNiteCafe
chenobol
#linux
|
|