>>>>> On Sat, 22 Jun 2019 10:35:40 +0300, Eli Zaretskii said: >> From: Paul Eggert >> Date: Fri, 21 Jun 2019 17:25:43 -0700 >> Cc: emacs-devel@gnu.org >> >> On 6/21/19 5:21 PM, Glenn Morris wrote: >> > Was having src/macuvs.h in the repository causing a problem? >> > It has changed twice since 2014. >> >> Not a major problem, no. I made the change because Lars ran into >> something; perhaps he can chime in. Eli> So I think we should put it back, and we should this time have a Eli> comment explaining why it is in Git although it's a generated file. Eli> May I ask that you, Paul, will in the future not rush into changing Eli> something, especially when we are not sure of its reasons, without Eli> waiting a few days for possible comments from people who do know the Eli> reasons? I think it will make the Git repository cleaner and will Eli> prevent unnecessary discussions and arguments. TIA. *grumble* This stuff has broken the build on macOS, probably because itʼs now building an initial emacs in a configuration thatʼs not been tested in a while, judging by the the contents of config.h: In file included from nsterm.m:49: ./nsterm.h:1322:20: error: typedef redefinition with different types ('NSUInteger' (aka 'unsigned long') vs 'enum NSWindowStyleMask') typedef NSUInteger NSWindowStyleMask; ^ /System/Library/Frameworks/AppKit.framework/Headers/NSWindow.h:35:32: note: previous definition is here typedef NS_OPTIONS(NSUInteger, NSWindowStyleMask) { ^ In file included from nsterm.m:49: ./nsterm.h:1326:6: error: enumeration previously declared with fixed underlying type enum NSWindowTabbingMode ^ (and src/macuvs.h obviously does not get regenerated) Robert