On Saturday, September 21, 2002, at 10:37 PM, Curtis L. Olson wrote:
> Jonathan Polley writes: >> I have a few questions brought about by some recent experiences, but >> what they boil down to is "Are we going to reestablish the minimun >> system requirements for FlightGear." >> >> First, due to deficiencies in the compiler, it is becoming more and >> more difficult to build FlightGear under MSVC 6.0. MSVC 6.0 cannot >> built the 3D cloud code so that it is debuggable (at least under my >> setup), but is CAN compile a non-debug version (this has happened >> with other modules, too, so it is not just a problem with the 3D >> clouds). These errors are not even compiler errors, but are library >> errors (go figure). There have also been edits to FlightGear simply >> because MSVC 6.0 can't handle certain C++ constructs. Do we want to >> continue to modify the code to support this compiler? If so, for >> how much longer? > > Does MS have something newer people can migrate to that works better? > I guess the MSVC developers will have to decide what can and can't be > done and make a decision about the best course of action. Generally, > building with more compilers and platforms is good for a project, but > it is annoying to have to make ugly code concessions to the stragglers > (not that MSVC falls into this category.) The main thrust of this question was to determine if version 7.0 (.NET) of MSVC should become the preferred platform. While it probably shouldn't happen now, at what point in time should this happen? I'm not sure at what point in time MSVC 5.0 became obsolete for building FlightGear (MSVC 6.0 has been around for a while). There may be a point in time where there is no active building being done under version 6.0. I expect to be upgrading to version 7.0 within the next couple months. >> Due to changes in OS libraries, Darrel's current MacOS X application >> runs under version 10.2 but not 10.1.X. I will be building a 10.1.X >> version this weekend, but I am hoping that 10.1 can be dropped >> within the next six months to a year. > > If people can reasonably upgrade to 10.2 and we clearly document that > FlightGear requires version 10.2, then I don't see a huge problem with > that. Since the Mac developers of which I am aware run version 10.2, there really isn't any ongoing builds taking place under that OS. While I do have one HDD that has 10.1 installed, I try not to boot from it (I will be doing this to rebuild Darrell's application for 10.1, but that's about it). I expect that most people who will be doing development for the Mac will be running under 10.2, and within the next six months I expect there to be very few 10.1 user's in the world (making this a moot point). >> Is there, or will there be, a plan to transition away from gcc 2.95 >> to 3.x? If so, any ideas as to the time frame? MacOS X 10.2 uses >> gcc 3.1.1 so I hope there isn't a requirement for gcc 3.2 for a >> while. > > It's my sense that gcc 3.x.x tends to be pickier than 2.95, so I'm not > aware of needing to do stupid/ugly code tricks to suppport 2.95. That > being the case, I don't feel the need to go out of my way to find > something to break the gcc-2.95 compiler, we'd probably break things > for other compilers in the proces. I wasn't sure if 2.95 caused any problems similar to what MSVC 6.0 was causing (i.e., forcing the use of non-standard C++ syntax). This was more of a long term question. With the upgrade of autoconf, I was wondering what other changes might be approaching. Jonathan Polley _______________________________________________ Flightgear-devel mailing list [EMAIL PROTECTED] http://mail.flightgear.org/mailman/listinfo/flightgear-devel
