Sinking Into Oblivion

Here you can discuss ReactOS related topics.

Moderator: Moderator Team

zefklop
Developer
Posts: 114
Joined: Sat Feb 11, 2006 8:47 pm

Re: Sinking Into Oblivion

Post by zefklop » Wed Jan 12, 2011 2:38 pm

Z98 wrote:Until ROS stops tripping over itself with respect to memory management and the like, we're in this pattern of basically fixing broken things as we run into them instead of planning on adding 'new' stuff. Or more specifically, attempts to add new stuff keeps exposing major bugs or other missing components and so work is diverted to deal with those first.
This is a perfectly valid way to do things. There a few methods to implement things.

I'd like to implement A on top of B, but B is crappy. Two choices :
- work around B crappiness by implementing A with hacks
- uncrapify B before implementing A
- implement A without hacks and fix what needs to be fixed in B in the process
There is no "standard" good choice to make. It depends on how critical is A and how good I know B so I can fix it.

Take explorer as an example. It's critical (we need a shell) and sits on top of shell32.dll. But shell32.dll is crappy -> our explorer is a grotesque hack. Beside is explorer-new, which is well implemented and awaits for a good shell32.

zed260
Posts: 105
Joined: Thu Oct 02, 2008 2:10 am
Location: cleveland tn
Contact:

Re: Sinking Into Oblivion

Post by zed260 » Thu Jan 13, 2011 5:15 am

@zeflop uncrapify B before implementing A is longer term the best option longer term


whichever option makes most sense in terms of least man hours over long term implmenting hacks would likely lead to even more man hours being spent longer term undoing those hacks


ultimately we need more devs which means we need corprate sponership in some form

vicmarcal
Test Team
Posts: 2732
Joined: Mon Jul 07, 2008 12:35 pm

Re: Sinking Into Oblivion

Post by vicmarcal » Thu Jan 13, 2011 2:06 pm

zed260 wrote:@zeflop uncrapify B before implementing A is longer term the best option longer term


whichever option makes most sense in terms of least man hours over long term implmenting hacks would likely lead to even more man hours being spent longer term undoing those hacks


ultimately we need more devs which means we need corprate sponership in some form
And during the last year we are doing so, we have rewriten almost all the Memory code from Common Cache to Heap Manager going through Memory Manager. We are rewritting the whole message system and the whole Gdi32...All these rewrites has introduced some regressions which stopped our release schedule. We are now in rewritting stage too, so the schedule is not followed until these major rewrites are almost finished.
And now the results are impressive, so soon you will enjoy them.
Image

jcalvarez
Posts: 27
Joined: Thu Apr 02, 2009 10:41 am

Re: Sinking Into Oblivion

Post by jcalvarez » Thu Jan 13, 2011 4:25 pm

I do not really understand why people are so upset about the lack of releases. Creating a windows clone is not a trivial task, look at Wine, I played with it in 1996, under FreeBSD 1.1.5, and it was horrible. They had their first official release 12 years later, 2008, and guess what, it is a very successful project.

The fact is, ReactOS code has not reached the critical mass yet, when most of the changes are in the upper layers, and the core is fairly stable. When that happens, people will see an exponential growth of "visible" progress and faster releases.

Just be patient, the guys are doing it for fun. That is commendable.

Regards,

Jose

### Memory allocation error, cannot start command.com....windets! ###

Post Reply

Who is online

Users browsing this forum: Bing [Bot] and 3 guests