| |
| |
| |
|
Page: 1 2 3 4
Comments:
<0> jauncie: internal/right/left are not manipulators, they're constants.
<1> Rethguals, do you primarily use win32
<0> Yes.
<1> that's one area that I like about win32, you don't have to test all sorts of platform-specific parameters
<0> That always seemed like a hack anyway. The gcc folk should just make it easier to upgrade compilers :)
<1> well, stuff like.. does the OS have rand(), random(), or arc4random()
<1> does it have strsep, or just strtok()
<0> That shouldn't be an OS issue.
<1> can be ;p
<0> Well. C has rand and strtok, end of story :)
<1> strtok has issues ;p
<0> Of course it does.
<1> yeah so, scons saved my from having to write testing directives in M4. And also from having to use automake & libtool to portably create shared libraries. It saved me so much grief that I'm going to give some money to the guy that wrote it
<0> You say that like its a weird thing :)
<1> it is
<0> Oh :(
<1> I don't pay for software in the sense that I don't use any that costs money
<0> Ah. My perspective is that software which saves me time has value, and if I haven't paid for it (because its free), I'll donate.
<0> The main exception to that is mIRC, which I use but hate.
<1> true
<1> I think it depends on the context
<1> I don't really.. make money, yet, on free software.
<1> but if I were, I'd be sure to 'tithe'
<1> did
<1> in the sense that free software furthers my ability to sell software
<2> Rethguals: You could hardly say that mIRC saves you time, anyway, eh?
<0> Perhaps I need to refine my definition of value :)
<0> ...or perhaps not.
<1> what do you think of the gnu public license
<0> Petty. :)
<1> in what manner
<0> I don't see any need to enforce the "You can't use it unless I can!" aspect.
<1> well I think the situation that it's designed to protect is hypothetical company taking the 600,000 lines of code that comprises gcc/g++, putting their name on it and selling it flat out without giving money or code back to the project.
<0> I mean, if someone wants to take a private fork, they should be able to, but they'll lose out on the benefits of having any active community around it. That decreases the value of the software.
<1> situation in which it's designed to protect, rather
<0> jdavis: ...but if someone slapped a Visual Studio cl*** IDE on it, shouldn't they be able to sell it?
<1> existing as how cl.exe and msdev.exe exist?
<0> Yeah.
<1> that's ok then
<0> ...by the LGPL, but not by the GPL.
<1> cause then you're just basically bundling the compiler with a nother program
<1> another
<1> rather than creating a "derived work"
<2> No, if they form different executables, the GPL code and the commerical code, then it's fine.
<0> Er, ok. Well... suppose someone embedded some of the GCC code in the process of implementing the spiffy IDE.
<0> I'm basically arguing that simplying reselling gcc won't fool anyone who doesn't deserve to be fooled.
<2> Then it becomes leeching :-)
<0> The value in free software is really that you're not writing it. If you want to take on the roll of writing it, its value decreases. If you consider that you can still improve it enough to make a profit, that sounds cool to me. I hate the GPL trying to enforce its viral properties.
<1> I dunno how long you've been into computers
<1> but do you remember the arc/pkarc fiasco
<2> Well there's a difference between free software and open source software.
<0> I remember that there was a bit of hoo-ha about compression formats. Wasn't that patent infringement?
<2> If you were giving away the executable only, then no one (short of dis***embly) would be able to use your work other than how you made it.
<1> nah the guy that wrote arc released the source code but said that it was "expected" that people register.. and of course companies had to pay for site licenses and stuff.
<1> well phil katz comes along, takes the source to arc.. rewrites some of the core routines in ***embly.. making it faster.. and then when the guy that wrote arc asked him to release his source code, he told him to get lost.
<2> Giving away the source completely freely is giving away much more, so some programmers don't actually give it away, they GPL it.
<1> claiming that his word was 100% original.. later an expert witness testified that phil katz's version had Thom's (arc author) comment spelling mistakes on it.
<0> Seems perfectly fine to me, except that PK was a lunatic :)
<1> I just find it interesting
<1> the 'gnu mantra'
<0> How?
<2> So the GPL is specifically designed to enforce cooperation or nothing. The LGPL is to encourage use (in a development sense).
<1> well I don't think that if someone spends a long time, possibly years, writing a lot of code, that somehow corporate entities are entitled to use that code to net them more money without either funding the project in some way or releasing their project as free software as well.
<1> when I say use, I mean create a derived work
<1> ie, take the code, make some changes, build an executable.. and start selling it.
<0> I would say that they shouldn't release it then.
<2> There's no compulsion to give back the new code if the executable is never distributed.
<0> If you're not going to allow people to use it, you're just gloating.
<2> It's not gloating, it's opening up the project to collaboration. If someone wants to *give* their improvements then the original developer would be happy.
<2> Moreover, there's more portability in distributed source than executable.
<1> more eyes to spot bugs
<0> Hm, the point of opening it to collaboration is a good one that I hadn't considered.
<0> I guess I just don't consider source code very valuable.
<0> Of itself.
<1> as a teaching tool?
<0> Huh?
<1> what about as a tool for learning
<1> source code, that is
<2> Among the situations that GPL prevents is: Company takes open source code of OpenProggy. Company makes small tweaks to make compatible program ClosedProggy with a few extra features. Some users pay for ClosedProggy, some use OpenProggy. ClosedProggy v2.0 is released with extra features but without backwards compatibility. ClosedProggy users upgrade. OpenProggy users then need ClosedProggy to interoperate with those users
<0> I can't hold many open source projects up as a good tool for learning, and the audience is somewhat limited.
<0> wth: It doesn't prevent that at all. The new one does, but that's controversial.
<0> All you do is slap a key on it.
<1> Rethguals, why is that. Do you think that open source correlates with poor quality?
<2> It does if the commercial version uses code from the open source version, since the company is compelled to release the changes, and then those can be compiled without cost by the users who weren't paying.
<0> jdavis: No. I think that like all software, the quality is very variable.
<0> I believe that it's neither better nor worse than closed-source.
<2> But I think the primary thing is to allow a project where collaboration can come from any source, without the risk of someone else making money out of the effort that has been given for free.
<1> you know Id software released the source to quake3 as free software
<1> and I thought it was good
<1> It was interesting to page through
<1> I appreciated them doing that
<0> Yeah. They released it after it ceased to have value to them :)
<0> Arguably, they could have released it under the GPL originally and still licensed it for huge sums of money.
<1> I dunno
<1> Right up until they released it they were getting a $250,000 security against an 8% royality for it
<0> There was one company that licensed it quite close to the end of its life, yes... but really it had limited value once the Doom 3 engine could be licensed.
<2> For the compiler example: If you wanted to use the code from Borland C++ in the design of your new IDE, you'd have to approach them for a license to use the code, and you'd either be denied or pay for it. The same applies with the g++ code. It just so happens they make it available under another license that makes it easy for it to be developed by multiple people and compiled on lots of platforms.
<0> Heh, as if you'd want code from BCC :)
<2> Maybe Kai/Intel C++ would be more plausible :-)
<1> I liked borland back in the day
<0> Kai bought Kai so that they could discontinue it :p
<0> Intel did, rather.
<2> I do think that many people who release their code with GPL would be just as happy with a less restrictive license, but just use GPL because it's easiest.
<1> Borland C++ back in the day.. 93 or so I think
<1> maybe it was borland c++ 4
<2> 93? I was still on BASIC in 93.
<1> heh
<1> I never "got" basic
<2> Maybe AMOS, or so.
<0> Heh, I was on C in 93.
<0> Just.
<2> I think that was 96 for me.
<1> I didn't like C very much
<1> in a sense
<0> I still don't.
<1> I liked the power, coming from simpler languages
<1> but it was too easy to smash a buffer
<1> or free a pointer twice
<1> or some other stupidity
<3> still easy to do that in C++, heh
<2> You typically have fewer char[] lying around, though.
<1> true
<1> snprintf is one of the big problems
<1> a lot of C programmers today still don't know how snprintf behaves if it runs out of space
<1> They ***ume that it'll at least null terminate, which it doesn't
<1> worse it returns how much it 'would have written if the buffer were infinite'
<2> At least it won't write all over the whole code :-)
<1> I used to loathe that so much
<1> program crashes at random places.. debugger doesn't help because the stack frame pointer is clobbered
<1> due do some stupid bug like an off by one
<1> havn't had to deal with that for years thankfully
<0> jdavis, the standard version of snprintf does null terminate.
<0> It was non-standard before C99.
<1> I didn
<1> 't know that
<2> The pre-C99 gnu version gave -1.
<1> But it's definitely a good change
<0> The snprintf function is equivalent to fprintf, except that the output is written into
<0> an array (specified by argument s) rather than to a stream. If n is zero, nothing is written,
<0> and s may be a null pointer. Otherwise, output characters beyond the n-1st are
<0> discarded rather than being written to the array, and a null character is written at the end
<0> of the characters actually written into the array.
<2> Of course, you should just check the return value anyway.
<0> 7.19.6.5... I guess it's now one of those things you need autoconf for :)
<1> Rethguals, that's a good idea
<1> well that's just it
<1> yeah you'd probably autoconf that ;p
<1> since there are probably more than a few platforms out there that are not compliant
<2> Or just say if you haven't gotten a C99 compliant compiler by now, you deserve the buffer overflow... ;-)
<1> I don't know much about c99
<2> But yes, that's exactly what autoconf is for.
Return to
#c++
or
Go to some related
logs:
saima khan
#php
raspberry beret mp3
clink.tcl help
#india
#india
scoolastic
lara veronin
sana geldim sana zuzu
#india
|
|