Avoiding excessive compiling after a rebase

John Jefferies john.jefferies at globalgraphics.com
Thu Jun 16 12:29:46 CDT 2011

On 16/06/2011 17:15, Kevin Bullock wrote:
> On Jun 15, 2011, at 7:10 PM, John Jefferies wrote:
>> For this purpose, I don't really care about nanosecond timestamps, the nearest minute would do. I'll see what I can come up with on a spare Friday afternoon.
> Whether _you_ care about nanosecond resolution or not is unfortunately irrelevant... if hg 'restores' a timestamp on a file a few nanoseconds off, you'll still have to rebuild it.

Hmmm. I've never come across a build system that behaves like that, so 
I'd be interested in learning about one that does.

I can safely say that my build system only relies on the relationship 
between timestamps of targets and dependencies and not on their absolute 
values, e.g. it only considers whether a header file is newer than it's 
.obj dependencies. So restoring the timestamp to the nearest minute 
won't usually provoke a rebuild. [Of course, timestamps should be 
restored as accurately as possible, they just don't need to be for the 
purpose of avoiding excessive rebuilding for me.]


More information about the Mercurial mailing list