As usual Walter, the answer is to not push incremental changes to master but instead to develop them in a branch.
On Mon, Jun 27, 2016 at 8:10 PM, Walter Bright via dmd-internals <[email protected]> wrote: > Of course you have a great point. The trouble is, for me anyway, it's hard > to write the spec incrementally. It also tends to make for a crummy spec. I > find it easier and works better to take a more holistic view of spec > updates. > > For example, I am currently working on the list of issues tagged with > 'safe'. Many of them interact with each other. It's easier to work on them > more or less in parallel when I have the 'context' of how that part of the > code works in my head. The same goes for spec updates. > > But I don't want it overlooked, hence the shortcomings should all have > bugzilla entries. > > > On 6/27/2016 2:06 AM, Михаил Страшун via dmd-internals wrote: >> >> On 06/24/2016 09:14 AM, Walter Bright via dmd-internals wrote: >>> >>> All such issues should be files as bugzilla entries and marked with the >>> 'spec' keyword. >> >> >> This is fighting consequence and not a problem. Better approach would be >> to have a simple checklist for all dmd/phobos mergers to follow: >> >> - Does this PR have tests? >> - Does this PR have documentation? >> - Does language spec and/or dlang.org articles need to be updated? If >> yes, is there a PR for that? >> - Does ths PR need changelog entry? If yes, is it there? >> >> If there is a failure on any of these points, PR must not be merged, no >> exceptions or excused. Not even if it is written by you and desperately >> needed by Andrei - for a simple reason that being too slow will never do >> as much damage to the language as making bad quality changes. > > _______________________________________________ > dmd-internals mailing list > [email protected] > http://lists.puremagic.com/mailman/listinfo/dmd-internals _______________________________________________ dmd-internals mailing list [email protected] http://lists.puremagic.com/mailman/listinfo/dmd-internals
