I really don't like this patch. It is clearly to fix an issue with the MSVC compiler. I would feel the same way about any other compiler as well. We should not have to jump through hoops to support every broken compiler. I still see no reason to change the policy to support MSVC or any other compiler for that matter.
On 5/28/2016 7:26 PM, Cirilo Bernardo wrote: > This may be a little off-topic, but the POSIX compliance of MSVC > has improved quite a bit and cross-platform support is no longer > the nightmare that it was 10 years ago. Even 4 years ago I would > have said supporting MSVC would be crazy, if not because of the > cross-platform nuisances then because it was not a tool accessible > to the general public. For the past few years Microsoft have been > making MSVS/MSVC available free of charge to individuals and > small businesses so I would now consider the tool as available to > the general public, and the POSIX compliance has been at a point > where there aren't too many cross-platform issues other than in > the GUI, which for us is abstracted through wxWidgets. So I think > we should reconsider MSVC support at some stage. I for one would > prefer an MSVC build to an MSys/MinGW build. For me MSys/MinGW > has always been a tremendous nuisance and not something I would > ever install if I had an alternative, and I believe the MS tools are > good enough and available enough at the moment. > > - Cirilo > > On Sun, May 29, 2016 at 3:01 AM, Chris Pavlina <[email protected] > <mailto:[email protected]>> wrote: > > I'm not going to commit this one without approval, since we've > already stated > we're not going to support MSVC. > > On Fri, May 27, 2016 at 04:32:54PM +0200, Simon Richter wrote: > > > > The combination of > > - template resolution > > - optimization > > - structured exception handling > > - chained operator= > > > > appears to trigger a bug in the compiler. > > > > As this code is nonsensical anyway (numeric_limits<>::min() > returns the > > smallest positive value for floating point types, but the largest > negative > > value for integral types), it should probably be rewritten. > > > > This change just avoids the compiler bug, though. > > --- > > include/math/box2.h | 8 ++++++-- > > 1 file changed, 6 insertions(+), 2 deletions(-) > > > > > _______________________________________________ > Mailing list: https://launchpad.net/~kicad-developers > Post to : [email protected] > <mailto:[email protected]> > Unsubscribe : https://launchpad.net/~kicad-developers > More help : https://help.launchpad.net/ListHelp > > > > > _______________________________________________ > Mailing list: https://launchpad.net/~kicad-developers > Post to : [email protected] > Unsubscribe : https://launchpad.net/~kicad-developers > More help : https://help.launchpad.net/ListHelp > _______________________________________________ Mailing list: https://launchpad.net/~kicad-developers Post to : [email protected] Unsubscribe : https://launchpad.net/~kicad-developers More help : https://help.launchpad.net/ListHelp
