Hi Slava, Hi everyone, Last week I learned that coding blind against platforms I can't test on isn't going to work. With that in mind I'm thinking my next step will be to make win32 factor completely reentrant whilst keeping the other platforms single threaded (i.e. not touching their platform specific bits). My native platform is linux 64 so I'm hoping coding for win32 (via wine) will keep me honest in this regard.
Cheers, Phil Phil Dawes wrote: > I noticed Slava's tweet that master was working with win32 so have > rebased my changes on top of this: > > http://github.com/phildawes/factor/tree/reentrantvm-checkpoint > > (Apologies if anybody merged this branch before) > > BTW, having all the functions in one class provides a nice overview of > the functionality in the vm: > http://github.com/phildawes/factor/blob/598a8bdf77c2eeae631f1eb406164ba2cfb1d171/vm/vm.hpp > > Cheers, > > Phil > > > Phil Dawes wrote: >> Hi Slava, Hi factor list, >> >> Last week I ran into a bunch of nasty problems backporting my vm changes >> to the clean-winnt-x86-32 branch, so I've taken a new tack: >> >> 1) I'm now developing on top of the clean-winnt-x86-32 branch, since >> this works for both win32 and linux64, and then forward porting to master. >> >> 2) I'm aiming to create a checkpoint branch that can be merged into the >> main factor trunk to act as a stable base to implement the fun hairy stuff. >> >> To achieve (2) I've cherry-picked and cleaned up a branch of patches (98 >> of them!) that implement just the platform independent changes and do >> not touch any factor or os-specific C/asm files. This branch bootstraps >> and works on win32 and linux and has a pretty good chance of working on >> all the other platforms. >> >> http://github.com/phildawes/factor/tree/reentrantvm-checkpoint >> >> The patchset basically moves all the c code into a singleton vm class >> (apart from the global variables referenced directly by factor compiled >> machinecode) and provides forwarding functions for those referenced by >> platform-specific code. There's two new files: vm.hpp and >> inlineimpls.hpp. The former contains the monster vm struct and the >> latter contains the inline functions that depend on the vm struct (moved >> there for dependency reasons). The inlineimpls.hpp should disappear >> later once the remaining globals are removed and the header files re-jigged. >> >> Slava: could you take a look at this branch and see if it's fit for >> merging into trunk, and if not let me know what needs to be done? >> >> Also could somebody with a Mac (x86 or ppc) download this branch and >> confirm that bootstrapping works ok? >> >> Many thanks, >> >> Phil >> >> >> ------------------------------------------------------------------------------ >> Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day >> trial. Simplify your report design, integration and deployment - and focus >> on >> what you do best, core application coding. Discover what's new with >> Crystal Reports now. http://p.sf.net/sfu/bobj-july >> _______________________________________________ >> Factor-talk mailing list >> [email protected] >> https://lists.sourceforge.net/lists/listinfo/factor-talk >> > > > > ------------------------------------------------------------------------------ > Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day > trial. Simplify your report design, integration and deployment - and focus on > what you do best, core application coding. Discover what's new with > Crystal Reports now. http://p.sf.net/sfu/bobj-july > _______________________________________________ > Factor-talk mailing list > [email protected] > https://lists.sourceforge.net/lists/listinfo/factor-talk > ------------------------------------------------------------------------------ Let Crystal Reports handle the reporting - Free Crystal Reports 2008 30-Day trial. Simplify your report design, integration and deployment - and focus on what you do best, core application coding. Discover what's new with Crystal Reports now. http://p.sf.net/sfu/bobj-july _______________________________________________ Factor-talk mailing list [email protected] https://lists.sourceforge.net/lists/listinfo/factor-talk
