Zack Weinberg [EMAIL PROTECTED] writes:
On 4/28/07, Stephen Leake [EMAIL PROTECTED] wrote:
Here is a possible fix for umask in key_store.cc on MinGW:
I don't mean to rain on your parade,
No problem; I was just trying to get things to compile, and getting
back into writing C++.
but I really don't think key_store.cc should have to know that
umask is the relevant concept, especially since the way to do this
on Windows is so very different. If we have a
write_data_userprivate(), we can use 3-argument open() on Unix, and
we can [theoretically - I stared at MSDN for some time this
afternoon without being able to figure out how] do the appropriate
dance on Windows too.
That is a much better solution.
I actually coded this up quite some time ago but forgot to finish
debugging it. I've pushed it now. It works on Unix; I hope the
Windows code works too, but I don't actually have the ability to
test it - even to compile it.
Actually, you can trigger the buildbot:
http://www.venge.net/mtn-wiki/BuildBot
I've checked it out, and it does compile. I don't think I'll bother
staring at MSDN to make it actually work.
I ran 'make check', and everything passed or failed as expected.
--
-- Stephe
___
Monotone-devel mailing list
Monotone-devel@nongnu.org
http://lists.nongnu.org/mailman/listinfo/monotone-devel
