re: Your comment "the cost is big". FWIW: One item that I recall was very complex was the message loop implementation, which handles both native Windows events and coordinates inter-thread Task processing.. It was quite difficult to create a task processing system that integrated with the Windows Message Loop, and subtle differences in Win 2K made processing of the Windows messages both cumbersome (a lot of extra code) and slow (multiple scans of the Windows Message Queue were needed in Win 2K). I think that after we reduced the complexity by abandoning Win 2k we were able to make progress which lead to eventual message loop vs message pump refactoring (to support linux and mac). It might be possible to add back in that Win 2k complexity now (almost as another port).... but I will warn you, it was a VERY delicate operation, and performance critical to the app in general. In addition to generic slowdowns, great care must be taken to not starve any of the various task queues (or Windows messages). I'm sure there were many other issues, but the above is one I recall.
Caveat coder, Jim On Wed, Jun 24, 2009 at 10:50 PM, pi <[email protected]> wrote: > > I presume that Chromium decided to support Windows 2000 when the > project started in 2006. The reasons may be: > > (1) The profit is big. There were 6% Windows 2000 users in 2006. > > (2) The cost is small. There should not be too many differences > between Windows 2000 (5.0) and Windows XP (5.1). > > Afterward, Chromium decided to cut out Windows 2000 when the project > grew up in 2008. The reasons may be: > > (1) The profit is small. There were only 2% Windows 2000 users in > 2008. Surely, there would be fewer users in future. Furthermore, most > of these remaining users were in corporate environments that were > locked-down against using chrome as a third party program. > > (2) The cost is big. Certain of functions need to be implemented > cumbersomely for compatibility with Windows 2000. Moreover, some > undocumented features of Windows 2000 lead to extra failures. For > example, when initializing an impersonated thread of a restricted > sandbox process, nt!ZwMapViewOfSection succeeds on Windows XP, but > fails as 0xc0000022 STATUS_ACCESS_DENIED on Windows 2000. > > Is it right? > > cpu wrote: > > Yes, the real reason is that there is an ongoing cost of keep that > > version working including extra QA cycles for each release. In terms > > of supporting a windows version with very few users we should focus > > our efforts on Win7. > > > > But you are welcome to keep an external fork. If there is any > > consolation, this was argued at length a year ago. > > > --~--~---------~--~----~------------~-------~--~----~ Chromium Developers mailing list: [email protected] View archives, change email options, or unsubscribe: http://groups.google.com/group/chromium-dev -~----------~----~----~----~------~----~------~--~---
