Ros-diffs December 2007

ros-diffs@reactos.org
  • 22 participants
  • 605 discussions

[greatlrd] 31120: patch from Kamil Hornicek tykef at atlas dot cz (irc nick : Pigglesworth) Fixed Main_DirectDraw_GetDisplayMode for dx1, dx2 Fixed some compile warnings
by greatlrd@svn.reactos.org
17 years

[dchapyshev] 31119: - Update Russian translation
by dchapyshev@svn.reactos.org
17 years

[weiden] 31117: _findnext is already implemented. Fixes a crash during Abiword startup
by weiden@svn.reactos.org
17 years

[fireball] 31116: - Because we'll need to pass the ARC hardware information to the kernel, we need to use the usual FreeLDR trick of using a static buffer where to store the data instead of allocating memory. But, unlike our other data, this kind of information is variably-sized, and it's not possible to make arrays of arrays and start assuming upper bounds. Therefore, give us a large 16KB stash buffer, and implement a very simple array-based heap allocator so that all the ARC hardware memory wi
by fireball@svn.reactos.org
17 years

[fireball] 31115: - Further refactor the code by completing getting rid of the registry key parameters and work only with ARC configuration component nodes; the registry key is now hidden in an opaque structure from the rest of the code. - Fix some bugs in the ARC component node creation.
by fireball@svn.reactos.org
17 years

[fireball] 31114: - Start filling out ARC component data nodes for all the new refactored hardware node APIs (these are just dummies for now).
by fireball@svn.reactos.org
17 years

[fireball] 31113: - Refactor creation of component data under the HARDWARE\DESCRIPTION node to use separate routines in a new ARC file, so that registry routines are factored out of the process. This allows 1) Better portability with PPC branch (which will need to do the same kind of hardware work) 2) Ability to eventually get rid of the registry calls and build a true ARC hardware device tree instead (which the kernel will have to parse during boot). The idea is to remove registry writing suppo
by fireball@svn.reactos.org
17 years

[fireball] 31112: - Our NtCreateKey currently allows building trees (which is incorrect) if the parent key is a symbolic link (which does exist), but if the target doesn't exist (Since the check 'does parent exist' is done Before the symlink is converted to its target. One side-effect is that although we create the CurrentControlSet symlink to ControlSet001, we never create ControlSet001. We end up creating it later during the boot by creating a sub-key, by exposing the bug in NtCreateKey. Since
by fireball@svn.reactos.org
17 years

[jimtabor] 31111: Prevent crash in gdibatch.
by jimtabor@svn.reactos.org
17 years

[hpoussin] 31110: CurrentControlSet should already be created at this point.
by hpoussin@svn.reactos.org
17 years
Results per page: