IRC Logs

Network: freenode Channel: #reactos

	Enter your search terms Submit search form
	Web www.irclog.org

[00:01]<fyzn2rff>hm, reactos now tells during 2nd stage setup that it can't open dinput.dll
[00:01]<fyzn2rff>dsound, dinput8
[00:01]<fyzn2rff>wth?
[00:01]<gznzvljzm>what
[00:01]<fyzn2rff>I will do a clean build now
[00:01]<gznzvljzm>have u format the hd
[00:01]<gznzvljzm>?
[00:01]<fyzn2rff>sure
[00:02]<fyzn2rff>I always format
[00:03]<gznzvljzm>werid
[00:03]<fyzn2rff>let me do a clean build
[00:03]<gznzvljzm>I am build a last time
[00:04]<fyzn2rff>probably it helps
[00:04]<gznzvljzm>see if I have not miss any regress
[00:08]<-- dvxn|syzzzyus xzs>http://www.bagdadsoftware.de")
[00:11]<fyzn2rff>well, I will sleep now
[00:12]<ddr>Does anybody know if ROS's PnP/USB support is good enough to get the FTDI USB<->Serial chip working?
[00:13]<gznzvljzm>about USB u should talk with fireball he just vanced to the bed
[00:13]<gznzvljzm>PnP u should talk with hpoussin
[00:13]<gznzvljzm>none of them are here
[00:13]<ddr>Oh well.
[00:14]<ddr>Thanks. ROS sure has come a long way by the looks of it.
[00:14]<gznzvljzm>yeah
[00:14]<ddr>You guys have done an amazing job.
[00:14]<gznzvljzm>we are working getting it futhers
[00:15]<ddr>Also, has anybody thought about taking the NTFS implementation from Linux and building a wrapper to make it an IFS driver?
[00:15]<ddr>That may be a good goal for port 0.3.x...
[00:15]<ddr>s/port/post/
[00:16]<gznzvljzm>I think of it time to time
[00:16]<gznzvljzm>when I count the negtive with the positve sides
[00:16]<gznzvljzm>the negtives wins
[00:16]<gznzvljzm>like it is bad idea
[00:17]<ddr>Just out of curiosity, why?
[00:17]<gznzvljzm>and u will lost alot speed and other stuff
[00:17]<ddr>Do you mean from the translation layer or just that NTFS has more overhead than FAT?
[00:17]<gznzvljzm>and windows and linux filesystem stack is alot diffent
[00:17]<gznzvljzm>the translations layers
[00:18]<ddr>I don't know much about the Linux kernel (at least, not these days). I am familiar with the BSD kernel architectures thoughh.
[00:18]<ddr>Well, it could be a gradual thing. I don't mean keep Linux file system semantics, but mimmick the Linux bio stuff in terms of IRP's until the code can slowly be changed. Thats one idea.
[00:18]<ddr>Or do you think ROS should have a from-scratch NTFS driver?
[00:19]<gznzvljzm>we have partly ntfs drv in ros
[00:19]<ddr>Really?
[00:19]<ddr>Wow. Didn't realize that.
[00:19]<gznzvljzm>I do not how well it is working
[00:19]<gznzvljzm>or it works at all
[00:19]<gznzvljzm>myg : our usb stack is from linux
[00:20]<gznzvljzm>our tcp/ip stack is from bsd
[00:20]<ddr>Yeah, I know.
[00:20]<gznzvljzm>it was alot work getting them inplace
[00:20]<ddr>True. I think the BSD stack was a good idea though.
[00:20]<gznzvljzm>but the work are not finish yet
[00:21]<gznzvljzm>Arty have done great work with tcp/ip stack
[00:21]<gznzvljzm>and fireball with usb stack
[00:21]<ddr>If we wanted to do NTFS like MS Windows does somebody would need to implement the LFS API's in the kernel.
[00:21]<zzvd>is head booting for anyone else?
[00:22]<gznzvljzm>it boting now
[00:22]<zzvd>hmm i'll make clean
[00:22]<ddr>In NT there is a generic API for journalling (a bit like Linux's jbd) stuff actually in NTOSKRNL.
[00:22]<ddr>Not sure if those API's are documented.
[00:22]<gznzvljzm>arty u must make a clean build getting it boot again
[00:22]<zzvd>ya
[00:22]<gznzvljzm>u can try the iso achive
[00:22]<gznzvljzm>to be sure it does
[00:22]<zzvd>oh don't know where it is
[00:23]<gznzvljzm>http://svn.reactos.org/iso
[00:23]<gznzvljzm>there is the iso archive
[00:23]<gznzvljzm>myg : we can look at captive drv as well
[00:23]<gznzvljzm>how he done some stuff
[00:24]<gznzvljzm>some stuff is right
[00:24]<gznzvljzm>some other part bit wrong
[00:24]<ddr>GL: Thats a good idea. What is the licensing of the captive stuff.... can it be integrated (legally) in ROS?
[00:24]<gznzvljzm>it comes from ros ntokrnl
[00:24]<zzvd>imo the main thing that keeps us from merging captive is heavy reliance on glib
[00:24]<gznzvljzm>with havy modifcations
[00:24]<gznzvljzm>arty not only
[00:25]<ddr>Hmmm. So (never having looked at the captive stuff) I take it the code they implemented does not use the Rtl* stuff?
[00:25]<gznzvljzm>he implement whole adv* into ntosnkrln and ntdll
[00:25]<gznzvljzm>it use rtl stuff
[00:25]<ddr>So what does it do with glib?
[00:25]<gznzvljzm>he implement all api he need into ntoskrnl
[00:26]<zzvd>memory allocation, string handling etc
[00:26]<zzvd>he used glib pervasively
[00:26]<gznzvljzm>yes he invent own cc
[00:26]<gznzvljzm>with glib
[00:26]<zzvd>which is ok considering he was going for using it in userland on linux
[00:26]<ddr>hmmm. Well, I should go look at it. that may be a good project for me if I can find the time.... de glib-ize the code.
[00:27]<ddr>Don't know if I have the time for getting it into the ROS kernel...
[00:27]<gznzvljzm>I found one fualt in ros ntoskrnl thx to cabitve source revivew and I fix it
[00:28]<ddr>I suppose that would be another interesting goal. Getting MS's NTFS.SYS running under ROS.
[00:28]<zzvd>myg: i think the first part of that goal is getting some open source fs drivers that work on windows working unmodified on linux
[00:28]<zzvd>on reactos sorry
[00:29]<zzvd>if that happens, I think we'll be much closer
[00:29]<ddr>I don't know of too many..... There was a closed-source port of XFS a while ago. Don't know what happened to it.
[00:29]<ddr>I know that ClearCase's MVFS is not a good place to start. :)
[00:30]<zzvd>myg: there are a few; ext2, cramfs, afs
[00:30]<ddr>arty: do any of them work under ROS?
[00:30]<gznzvljzm>bfs if i rember right
[00:30]<gznzvljzm>not yet
[00:31]<gznzvljzm>for we do not have a cache mangers
[00:31]<gznzvljzm>to start with

Back to #reactos
Or go to some related logs:
#ubuntu
#gentoo
#php
#cisco
#reactos