| |
| |
| |
|
Page: 1 2 3 4 5 6 7 8 9
Comments:
<0> I think it was a mistake, but given how strongly the C freaks were pushing "same capabilities as C files" I thnk we got off pretty clean
<0> but we're still too bound to C
<0> e.g. why the hell do we have to ADD << hexfloat << to the library
<0> it should have been there to start with (for consistancy)
<0> why did they waste 3 flags to give us default, octal, decimal, hexadecimal
<0> when we could have had "base256" which would have forever gotten rid of the "network order" bull****
<1> We should have << std::base(unsigned char) <<
<1> or rather std::ios_base::base(unsigned char)
<0> well, with only 3 bits available, it would be tricky to hold it
<0> but I would have extended it so frigging fast it would have made the committee's head spin
<1> It would require more than 3 bit, obviously.
<1> But who the **** cares if a stream object needs one more byte.
<0> but with 3 bits you could have gotten binary, quartic, octal, decimal, hex, base64, base256, and the default
<1> (and saves at least two bit)
<2> error: expected primary-expression before '>' token at this line while (n<>0) {
<1> ^Sorin^: != not <>
<0> yes, that' snot legal code
<0> <> is for templates
<1> we're not #pascal
<0> we're ~Pascal
<0> _m_ you might have a problem with an arbitrary base (you'd need to define the digit alphabet (perhaps)
<0> )
<0> I can't see it being _so_ expensive to construct an fstream that they chose to have a default one, and open()
<3> hmm, i've heard a lot about creating gui's using QT but i cant seem to find a free version of the libraries etc. Does one exist?
<1> vawjrwrk: yes. However, at least bases from 2 to 36 should be supported.
<1> nnp: yes, such version exist (well, "free" as in "GPL")
<3> yea, thats pretty much what i was looking for. I'll look harder
<0> "GPL free" some people call "enslaved"
<4> GPL I call someone's ideology.
<0> you don't by any chance mean Richard Stallman's do you?
<4> the probability is very high.
<3> you sound like a magic 8 ball :P
<4> I try
<5> heh
<0> _m_ Josuttis doesn't mention getenv()
<6> nnp, Qt4 is GPL'd for all platforms afaik
<3> cheers
<4> there is a precompiled mingw version of qt4 on trolltech's website
<4> if you buy their license, it becomes dual licensed!
<0> ah, in cstdlib
<7> what does this mean:
<7> /home/cristi8/tmp/ccAnZFbN.o:(.eh_frame+0x12): undefined reference to `__gxx_personality_v0'
<7> collect2: ld returned 1 exit status
<0> I think that's some install problem with gcc
<0> what did you type to get that error?
<7> gcc -o lyrics lyrics.cpp -lmysqlclient
<0> well, it should be g++ not gcc
<1> calc g++link
<8> g++link = If you get undefined symbol errors when linking C++ programs, make sure you are using "g++" and not "gcc" to link. "gcc" will compile C++ code, but will not link the C++ libraries for you.
<7> :) it worked
<9> use gpp not gcc
<7> thanks.. i usually use C language..
<0> gpp????
<9> g++
<6> gpp was that it was called in DJGPP:)
<1> cristi8: you don't, apparently.
<6> s/that/what
<1> err, I overlooked the "usually".
<0> cristi8 why aren't you using make?
<7> i don't know how
<6> make foo # where you have foo.cc
<7> i'll learn it when i get tired of compiling by hand
<1> make wouldn't find -lmysqlclient without a makefile.
<6> yeah
<6> true
<7> _m_: this is one of my first programs in C++. i *usually* work in C, now i'm trying C++ for a little project
<0> _m_ of course it would, if you had an environment variable LDLIBS=-lmysqlclient
<10> jjeeeeeeeeeeeet
<10> ergee
<1> true
<0> and if you're gonna be screwing w/ mysqlclient it would make sense
<1> However, that would not be very convenient.
<0> then oyu'd just have to type make lyrics
<1> which isn't very difficult.
<1> heh, I misread that, too.
<1> ("make lyrics" == Makefile)
<0> nope
<0> you don't need the actual Makefile
<0> that's the point
<1> You need it to avoid the h***le of setting environment variables.
<0> the default rules are put together quite nicely
<1> At one day, you'll work one more than one project. Setting environment variables would not scale well.
<0> it's a shame that *nix never figured out that there should be a way to automatically set env varibles when you cd
<1> Run has a script that does that.
<1> I think it's the wrong approach for software development, anyway.
<0> one would hope it exists, it's one of the more useful things.... you can do it with a .bashrc i think if you're using bash
<0> but it's been a while
<1> Unless you're able to check your environment variables into the version control system, too.
<0> fdo you check your compiler into version control?
<1> I use tcsh, but it would be possible with that, too.
<1> Actually, I do.
<0> we're implementing that here now also.. I've been doing it personally for quite a while
<1> Most people will give you strange looks for doign that.
<0> it's a shame that the committee hasn't demanded something in the core language like MS's pragma comment(lib, "......")
<0> it's been needed for a lot longer than C++ has existed
<1> But nothing feels as great as checking out a several years old archive, typing make, and seeing it just works. You can't do that well without checking the compiler in, too.
<0> _m_ agreed
<1> *shrug*, I never felt I need #pragma comment(lib, ...)
<0> how do YOU find the right name for the boost library?
<1> It feels like mixing program internals with the build environment.
<1> I *know* the name.
<0> but it's ****ing insanity that the SOURCE cannot tell the linker to search a library
<1> I do have a copy of the compiler and a copy of boost in my projects.
<0> _m_ then you are a very rare person
<1> (or in some 'environment' project shared by several projects)
<1> It's not like there are infinitely many boost libs.
<7> oh.. and one more thing.. how can i make a cl*** visible only in the current file ?
<0> switching between debug builds and release builds requires different boost libs
<0> cristi8 namespace { your definition}
<1> I don't use boost debug libs, unless I'm debugging Boost stuff.
<0> you don't have a choice in MS
<1> (and even then it's likely I don't use debug libs, either)
<0> options have to match
<0> especially w/ the dll / static stuff
<1> Yeah, you don't have that problem on Unix. You can freely mix debug and non-debug stuff.
<1> dll/static isn't much of a problem in my world, either.
<1> And dlls don't have their own heap here.
<0> ok, find YOU don't need it...which is why it's NOT in the standard
<0> like #pragma once
<0> which was in the Apple compilers before C++ existed
<1> #pragma once would be useful for everyone.
<0> that it's not useful to all is no reason not to put it in the standard
<1> I don't recall seeing #pragma once on Apple's compilers.
<0> that's the 1st place I saw it
<1> Yes, but being useful for everyone is a very good reason to put stuff into the standard :)
<0> it's not there
<1> The committee would probably reject such a proposal. #pragma is supposed to be entirely implementation defined.
<0> of course the 3 perople in the world who find it "good" to include the same file more than once doing some #defines in between are the ones that caused us to even need it to start wityh
<11> vawjrwrk: #pragma isn't part of the standard as having undefined behaviour?
<1> '#once' would probably have a better chance.
<6> are there actually any guaranteed #pragma's? I always though the point was to offer compiler specific options
<0> I meant the feature
<1> jottinger: it *is* part of the standard.
<0> <sigh>
<1> jottinger: it is implementation defined, not undefined.
<11> _m_: ah, K
<11> well, yeah... IANALL
<0> and we NEED a directive to tell what "language/alphabet" is being used for "abcde" not this crappy L"abcde" for _some_ long character set
<12> moin
<12> ah unicode talk
<12> ?
<1> no, not at all.
<0> mp. bitdhing about stuff not in C++
<1> vawjrwrk: I'm not sure we need that.
<12> i would love a *all "strings" become charset X
<1> However, I use L"" extremely rarely.
<0> _m_ ok _need_ is perhaps too strong, but it's ****ing stupid to require L"" just because we're all moving to a new standard character set
<0> _T is even worse
<1> _T is a nightmare.
<12> in most projects i have simply a char * -> std::wstring thingy
Return to
#c++
or
Go to some related
logs:
#AllNiteCafe
shana hill
tartgirl
javascript check a checkbox inside iframe
#c
#AllNiteCafe
#chatzone
#c++
#squid
#AllNiteCafe
|
|