October 29, 2003 Re: Longhorn | ||||
---|---|---|---|---|
| ||||
Posted in reply to davepermen | "davepermen" <davepermen_member@pathlink.com> wrote in message news:bnov1j$104l$1@digitaldaemon.com... > In article <bnot23$tgg$1@digitaldaemon.com>, Achilleas Margaritis says... > > > >If D is a "systems" language, why does it have garbage collection (that can't be turned off) ? It does not seem that I can make an O/S using D... > > > you can turn on and off the gc, no problem. don't bitch on D if you don't know > d. > > import gc; > gc.disable(); > gc.enable(); Also, you can plug in your own gc. Or shut it down and overload the new operator and implement it using C's malloc, or even implement it using assembler. The choice is yours. It can all be done with D. Lars Ivar Igesund |
October 29, 2003 Re: Longhorn | ||||
---|---|---|---|---|
| ||||
Posted in reply to Christian Kaiser | "Christian Kaiser" <chk@online.de> wrote in message news:bnotsj$uhi$1@digitaldaemon.com... > You should not underestimate the efficiency of C# and the CLR. Comparisons (in windows::developer network) show that it can be much more efficient than > current c or C++ - depending on the quality of the libraries etc. Well, the guys at MicroSoft says the darndest things ;) Lars Ivar Igesund who won't pretend to know much about the managed code stuff |
October 29, 2003 Re: Longhorn | ||||
---|---|---|---|---|
| ||||
Posted in reply to Christian Kaiser | Hey! That was my study. :) And I did not read the results to say that C# was good enough for the implementation of operating systems. If that's what they're suggesting, I find that both scary and ridiculous, and I'll be spending more time on my Linux box "Christian Kaiser" <chk@online.de> wrote in message news:bnotsj$uhi$1@digitaldaemon.com... > You should not underestimate the efficiency of C# and the CLR. Comparisons (in windows::developer network) show that it can be much more efficient than > current c or C++ - depending on the quality of the libraries etc. > > Ch. > > "Lars Ivar Igesund" <larsivi@stud.ntnu.no> schrieb im Newsbeitrag news:bnoo2j$m3f$1@digitaldaemon.com... > > A managed version of D is probably doable with enough resources. > > His point I guess, is that managed code don't sound like it's any good for > > the efficiency necessary for a systems language. I haven't read anything > > about Longhorn yet, so I should probably not comment any further. But > > I can't let go; It would probably be excessively stupid by MS to drop > > support for their C API with 10+ years of legacy application code. > > > > Lars Ivar Igesund > > > > "Albin Pucnik" <albin.pucnik@guest.arnes.si> wrote in message news:bnonf7$l90$1@digitaldaemon.com... > > > So there is no place for D in the managed world of Longhorn? > > > > > > Bine > > > > > > J. Daniel Smith wrote: > > > > "completely managed code" isn't true, but Longhorn clearly increases > the > > > > focus on managed code even more. > > > > > > > > D has been designed as a "systems" language (a "replacement" for > > traditional > > > > C/C++). MSIL/CLR/.NET has been suggested before, but seems to be > > against > > > > the "systems" nature of D. > > > > > > > > Dan > > > > > > > > > |
October 29, 2003 Re: Longhorn | ||||
---|---|---|---|---|
| ||||
Posted in reply to davepermen | I've not used it, so don't know, but I presume the disable() / enable() are reference counted, and only when the rc is 0 is it on? If that's not how it works, it should be "davepermen" <davepermen_member@pathlink.com> wrote in message news:bnov1j$104l$1@digitaldaemon.com... > In article <bnot23$tgg$1@digitaldaemon.com>, Achilleas Margaritis says... > > > >If D is a "systems" language, why does it have garbage collection (that can't be turned off) ? It does not seem that I can make an O/S using D... > > > you can turn on and off the gc, no problem. don't bitch on D if you don't know > d. > > import gc; > gc.disable(); > gc.enable(); > > and for most stuff, even system stuff, gc is not something hurting at all. a > potential memory leak is much more hurting. think of some service you set up > that runs 24/7 and happens to leak memory on say a daily basis.. have fun debugging while waiting several days each time to actually measure that you > leaked.. > > even microsoft is not sure yet how to bugfix exchange 2000 to not leak anymore > and eat up ram more and more. > > system level stuff has to be good performing, sure. but it definitely have to be > save. thats priority and THE reason why the gc is around everything. savety. > > |
October 29, 2003 Re: Longhorn | ||||
---|---|---|---|---|
| ||||
Posted in reply to Albin Pucnik | "Albin Pucnik" <albin.pucnik@guest.arnes.si> wrote in message news:bno9i3$2cd$1@digitaldaemon.com... > Hi! > It looks like Microsoft is moving ( with its new > operating system codename Longhorn) to completely > managed code( see reports from their PDC). > What does that mean for D? Will D compilers translate > code to MSIL(microsoft intermediate language) to be > excecuted by their virtual machine (CLR)? I haven't studied MSIL enough, but I think it's possible to do a D implementation for it. To start such a project, you'd need little more than the open source front end to D. |
November 01, 2003 Re: Longhorn | ||||
---|---|---|---|---|
| ||||
Posted in reply to Matthew Wilson | Hehe i read that too. I was kind of mad after that, I hate VM's ( dont take away my arguments! ) C "Matthew Wilson" <matthew-hat@-stlsoft-dot.-org> wrote in message news:bnp7ev$1ct2$1@digitaldaemon.com... > Hey! That was my study. :) > > And I did not read the results to say that C# was good enough for the implementation of operating systems. If that's what they're suggesting, I find that both scary and ridiculous, and I'll be spending more time on my Linux box > > "Christian Kaiser" <chk@online.de> wrote in message news:bnotsj$uhi$1@digitaldaemon.com... > > You should not underestimate the efficiency of C# and the CLR. Comparisons > > (in windows::developer network) show that it can be much more efficient > than > > current c or C++ - depending on the quality of the libraries etc. > > > > Ch. > > > > "Lars Ivar Igesund" <larsivi@stud.ntnu.no> schrieb im Newsbeitrag news:bnoo2j$m3f$1@digitaldaemon.com... > > > A managed version of D is probably doable with enough resources. > > > His point I guess, is that managed code don't sound like it's any good > for > > > the efficiency necessary for a systems language. I haven't read anything > > > about Longhorn yet, so I should probably not comment any further. But I can't let go; It would probably be excessively stupid by MS to drop support for their C API with 10+ years of legacy application code. > > > > > > Lars Ivar Igesund > > > > > > "Albin Pucnik" <albin.pucnik@guest.arnes.si> wrote in message news:bnonf7$l90$1@digitaldaemon.com... > > > > So there is no place for D in the managed world of Longhorn? > > > > > > > > Bine > > > > > > > > J. Daniel Smith wrote: > > > > > "completely managed code" isn't true, but Longhorn clearly increases > > the > > > > > focus on managed code even more. > > > > > > > > > > D has been designed as a "systems" language (a "replacement" for > > > traditional > > > > > C/C++). MSIL/CLR/.NET has been suggested before, but seems to be > > > against > > > > > the "systems" nature of D. > > > > > > > > > > Dan > > > > > > > > > > > > > > > > |
November 10, 2003 Re: Longhorn | ||||
---|---|---|---|---|
| ||||
Posted in reply to Albin Pucnik | http://download.microsoft.com/download/e/d/1/ed1ca94f-7330-4602-a053-703684d b74bb/20031028LHORNDB.exe Scary scary scary scary scary. And don't beware the .EXE extension, its just a wrapped .ZIP file holding a .WMV file. -- Chris S. |
Copyright © 1999-2021 by the D Language Foundation