IRC Networks
Irc Logs Stats
Start date: 2007-09-27 02:48:27
Last update: 2008-10-24 20:19:38
Channels: 41
Logged Lines: 6230436
Size: 1822.60 MB
Powered by
Channel Info
Network: freenodeChannel: #csharp |
Search in www.irclog.org
Log from #csharp at freenode 2006-05-27
[00:22]<eg->dynamic custom web apps are a different story
[00:22]<svgvgg_>maybe, maybe not, it's just to intergrate into the check the EMS, so it can check the enforcement of a policy
[00:22]<eg->ems?
[00:22]<svgvgg_>and react to "event's" dyncamically
[00:22]<svgvgg_>Enterprise Management System
[00:22]<eg->ok
[00:23]<svgvgg_>it's on sourceforge, or it will be i haven't actually put the files up yet
[00:23]<svgvgg_>my dissertation was a kind of mad 36 hour pro-plus and caffine run, admiteddly the code is a bit messy
[00:27]<svgvgg_>is there an easy way to get monodevelop to switch from it's normal path to using /usr/local anyone?
[00:33]<fyfuym-syfggag>any south africans here?
[00:34]<jgg_q_jgg>hi all
[00:34]<jgg_q_jgg>`abc? is a `namespace' but a `type' was expected
[00:34]<jgg_q_jgg>how can i fix that ?
[00:34]<jgg_q_jgg>in compiling a app give that :S
[00:35]<jgg_q_jgg>i think its C# language
[00:43]<dyguw>Hello
[00:43]<dyguw>In C# 3.0, will there be manual memory management? Like a 'delete' companion for 'new'.
[00:44]<kjr>uh
[00:44]<dyguw>I want manual memory management, I hate the GC.
[00:44]<kjr>so don't use a managed language...
[00:45]<lnvnzxu>then use c++/cli
[00:45]<dyguw>Well, I want to use .NET :)
[00:45]<lnvnzxu>eventually, it will enable mixing of managed and native heaps
[00:45]<dyguw>I see no need for a GC, it just makes the programmer lazy.
[00:46]<kjr>peterhu: can't you do that already?
[00:46]<kjr>minux: unless you're writing a memory management app, why should you be managing memory
[00:46]<kjr>why add that to the list of shit a programmer needs to take care of
[00:47]<dyguw>Because I want to decide when and where my objects are removed.
[00:47]<kjr>why?
[00:47]<wyfaj`>ah, still someone who's stuck in the past ;)
[00:47]<kjr>sup wilco`
[00:47]<dyguw>Because I've been doing that for the past 10 fscking years, I am used to do it. This GC thing makes me confused.
[00:47]<wyfaj`>hi
[00:48]<kjr>wilco`: see, we're already in agreement ;)
[00:48]<kjr>wilco`: didn't even take 7 years of Las Vegas
[00:48]<lnvnzxu>transparent mixing in the same type, like a managed type that encapsulates native types
[00:48]<wyfaj`>yeah, I'm trying to find some ground for a disagreement
[00:48]<wyfaj`>there's no fun in agreements.
[00:48]<kjr>wilco`: TOO LATE!
[00:48]<lnvnzxu>c++/cli does not yet support that
[00:48]<wyfaj`>no, it's not too late
[00:48]<lnvnzxu>it requires a shim between the two heaps
[00:48]<wyfaj`>see, it's already starting.
[00:49]<kjr>peterhu: wait, so mixing unsafe contexts with safe contexts is really just wrapped?
[00:49]<lnvnzxu>sorry, unsafe?
[00:49]<dyguw>I am used to do manual memory management, hence it doesn't take me any longer to write code for it. (You GC people say the programmer has to think about memory management, it comes naturally for me, nothing I think of)
[00:49]<lnvnzxu>minux, then perhaps the technology is not for you then
[00:49]<kjr>peterhu: the unsafe context... doesn't that require manual malloc/free (new/delete... whatever)
[00:50]<wyfaj`>minux: trust me, you may think you own it, but at some point even you are going to make mistakes.
[00:50]<-- svgvsdyzgjvr xrs>/dev/null")
[00:50]<kjr>minux: so implement the Idisposable interface
[00:50]<wyfaj`>if you don't believe me, I think you're, um, naive :)
[00:50]<dyguw>wilco`: I believe you :)
[00:50]<kjr>or write a pre-processor that just tells the compiler to ignore your destructor
[00:50]<wyfaj`>minux: obviously this may not be true for cases where you write trivial, little applications
[00:50]<wyfaj`>anyway.... are there any sideeffects of the GC that bother you?
[00:51]<lnvnzxu>kog, still not sure what you mean; C++/CLI has gc types and native types, gc types cannot contain native types (yet)
[00:51]<kjr>wilco`: yes, but loosely coupling management, we can isolate a point of failure for all programmers at once ;)
[00:51]<dyguw>Yes, I hate having objects lying around unused but still occupying memory space.
[00:51]<lnvnzxu>so you can't have: class Foo { }; ref class Bar { Foo f; }
[00:51]<wyfaj`>minux: you hate it because?
[00:51]<kjr>peterhu: I was talking specifically about C#
[00:52]<kjr>peterhu: C# has "unsafe" context - non-managed sections
[00:52]<dyguw>wilco`: I like bytes, I dislike using more bytes than neccessary.
[00:52]<dyguw>necessary*
[00:52]<kjr>peterhu: I've never used one, so I haven't looked too hard
[00:52]<wyfaj`>minux: if the object is no longer reachable and the allocated memory is actually required by other applications, it'll be freed anyway
[00:52]<lnvnzxu>kog, C# has no capability to inherently use native code, other than what the .NET platform provides it (P/Invoke, COM Interop)
[00:52]<kjr>minux: so, how much time have you spent writing assembler. pony up
[00:52]<lnvnzxu>unsafe just means "use pointers", it's still IL
[00:52]<kjr>peterhu: I thought you had to free stuff like that
[00:53]<dyguw>Kog: I've made my very own little language and a "compiler" that converts it to ASM.
[00:53]<kjr>peterhu: I'm not particularly familiar w/ Interoping native stuff either... I've done it
[00:53]<kjr>minux: cheater
[00:53]<lnvnzxu>not unless you P/Invoked something that did an allocation on something other than the GC heap
[00:53]<wyfaj`>unsafe = managed but unverifiable code
[00:53]<kjr>wilco`: ah so
[00:53]<dyguw>Kog: Well, to learn ASM of course.
[00:53]<kjr>peterhu: well, so I was dealing with named pipes... and I was doing interop for it obviously heh
[00:53]<dyguw>For the GNU GAC compiler to be precise.
[00:53]<kjr>peterhu: and I know that you can call "close" on the file object, which wraps the native bit
[00:54]<wyfaj`>minux: get Rotor 2.0 and implement a GC that is a non-GC ;)
[00:54]<kjr>peterhu: but I'm not sure who exactly owned that... I guess the VM does
[00:54]<wyfaj`>and add support for explicit cleanup of resources
[00:54]<dyguw>peterhu: The technology isn't for me, I'd code my apps in pure C/C++ if there were Windows Forms 2.0 bindings.
[00:54]<lnvnzxu>this is in contrast to a language like c++/cli, which has inherent support for native types, which you allocate on a native heap (CRT, what have you) like any other native C++ program
[00:54]<wyfaj`>where in this case with resources i mean objects as well
[00:54]<kjr>peterhu: well, so I wasn't quite sure if C# does anything native
[00:55]<kjr>peterhu: "managed" C++ is such a hybrid monstrosity
[00:55]<kjr>it's not really C++ anymore
[00:55]<lnvnzxu>"managed" C++ is deprecated and no longer exists
[00:55]<lnvnzxu>c++/cli is an ECMA standard
[00:55]<lnvnzxu>becoming ISO
[00:55]<wyfaj`>minux: you can do your winforms stuff in C# and write the rest in C++/CLI
[00:55]<lnvnzxu>and the language is fantastic
[00:55]<kjr>peterhu: ok, so I mislabeled, sorry... c++/cli
[00:55]<lnvnzxu>the IDE support just isn't quite there yet
[00:55]<kjr>peterhu: yeah, but I hear it's not really quite C++ anymore
[00:55]<kjr>peterhu: like it disowns multiple inheritence (punny eh?)
[00:56]<lnvnzxu>well, for reference types, yes, but that's a limitation of .NET itself







