Talk:Version Status

Notice Board

NT-style booting done at r43273.

Sources

<Fireball> FAT32 - needs a rewrite of a file system driver (fastfat), test it against windows 2003
<Fireball> NTFS - completely missing (not to count a very simple stubbed IFS driver). It would make sense to use NTFS
 3g from Linux and develop an own driver, also should be developer against Windows 2003
<Fireball> *developed
<Fireball> Ext2 - support in the kernel is missing for a good/known to work driver
<Fireball> VXD - not planned at all, it's a Win9x thing
<Fireball> WDM - hm, this deserves more explanation, let's postpone it to later
<Fireball> then
<Fireball> GDI -> Timo
<Fireball> GDI+ - completely shared with Wine
<Fireball> OpenGL -> Pigglesworth
<Fireball> DirectX -> Pigglesworth
<Fireball> Sound section -> janderwald
<Fireball> Bootloader - in a rather good shape, needs further development in terms of adding more features and fixing existing
<Fireball> Kernel - deserves its own talk, with different modules, etc
<Fireball> Mm and Ob to be part of Kernel subsection
---
<Fireball> Bootloader - we switched to the NT-compatible booting method a few days ago. Old bootmethod is deprecated and will be deleted from codebase after the 0.3.11 release
<Fireball> Memory manager - in a process of a soft rewrite, resulting in a functional equivalent of NT's (2003's) memory manager, without made up ROS-specific implementations and assumptions
<Fireball> Object manager - nearly perfect
<Fireball> functionally equivalent to 2003's Ob, tested code, works very good, experienced only bugfixes for the last two years I think
<Fireball> Configuration Manager - needs bugfixing, underlying cmlib needs improvements
<Fireball> Common Cache - stays as it is for now, adapting to the changing memory manager. There exist three rewrites: No-Cc (a version of the kernel which does no caching at all, very slow); cache-manager-rewrite, started long time ago, but still occasionally maintained in a working condition; arty-cc by Art Yerkes, totally new Cc with a good, separated from Mm architecture, WIP


<janderwald> portcls - 70 %
<janderwald> sysaudio 80 %
<janderwald> wdmaud 60 %
<janderwald> kmixer 40 %

<Pigglesworth>
opengl32.dll itself:
- almost complete, works fine in windows (xp), can load various ICD drivers provided by video card manufacturers
- missing only few wgl* APIs, will be fixed soon
- default software rendering is done through mesa3d icd driver
--> actual status of opengl support depends on the ICD driver ie. not us
- basic opengl games work (quake3a)
directx
- we support directdraw, directx8 and directx9
- currently we use only Wine libraries thus status depends on wine
- because it relies on an additional opengl layer its performance is affected by the status of opengl
--> kinda works with mesa but mesa have currently problems supporting shaders, don't know what the situation with real HW is, have never tested
- basic ddraw/direct3d demos work, few games work too

<Pigglesworth> [Win32k] the best is to ask jimtabor or Phycius.. I know james worked on the class rewrite, timo
 fixed xlate problems, but what is the overall status I don't know.. menus needs fixing, font rendering too, the real dx implementation is mostly missing

<Physicus> support for User mode printer drivers: unimplemented
<Physicus> text/font system: needs complete rewrite
<Physicus> DIB code: mostly implemented, could need optimisation in some places
<Physicus> mouse pointer: missing alpha support
<Physicus> region support: implemented, only minor interface improvements needed
<Physicus> pens: missing support for geometric pens
<Physicus> brushes: missing support for driver managed hatch brushes
<Physicus> handle manager: implemented, needs improvement
<Physicus> gdi handle manager that is
<Physicus> user handle manager: needs to be fixed to work more like windows
<Physicus> driver loading: needs rewrite (WIP by Timo)
<Physicus> bitmap code: especially DIB sections are broken like sh**, needs lots of fixing
<Physicus> for all user related code you should ask jimtabor, he is currently working in that area

<janderwald> portcls - is the class driver for the audio driver subsystem
<janderwald> midi support is missing
<janderwald> and testing for port driver implementation for IPortWavePci & IPortWaveRT
<janderwald> sysaudio - handles access to the audio port drivers by manageing the device interfaces
<SuperDog> you mean hardware testing is needed to design those two functions IPort* ?
<janderwald> todo: needs to map multiple audiostream to a virtual audio pin in kmixer
<janderwald> yes
<janderwald> wdmaud - is kernel mode component of wdmaud.drv 
<janderwald> implements the wave / mixer / midi api
<janderwald> needs midi api's implemented and wave in support + mixer support
<janderwald> portcls / sysaudio also need testing & bugfixing for wave in (audio recording)
<janderwald> kmixer - performs mixing / conversions of streams
<janderwald> currently very poorly implemented
<janderwald> needs kernel fpu float saving support
<janderwald> before more work can be done