Hi Nigel,
Thank for the patch! It looks promisin. However could you please clean your patch? 80% of it's content are just suppressing tabs and space... It is important that patches contain the minimum modifications otherwise it takes ages to check..

On Sun, Mar 16, 2008 at 9:09 PM, Nigel Kerr <nigel.kerr@gmail.com> wrote:
good folk,

I just submitted
, I would appreciate it if a committer could let me know when or if it
can be applied.  See the patch description for more gorey details.

In particular, though, I'd note that

1. CTest now enabled via this patch: a test target is added, and all
tests added are run.  "tests" are just executables, that exit with
non-zero on any failure.  There is one test there now, a somewhat
primitive test versus my date/time functions.  This has been useful to
me in getting things correct, and any sort of non-gui computation
could probably benefit from such automated tests of normal and
boundary conditions.  (However: having just pressed the submit button,
i realize the test i've got for whether Qt understands EST/EDT time
won't work in any other timezone...argh...).

2. timezone detection works on the following principles:
a. we don't care what the name of the local timezone is, just the GMT offset
b. we let Qt figure all that out, we rely on QDateTime's .toUTC() to
provide the offset.
c. (as a consequence of b) if Qt doesn't know details about the shift
for a point in the past or future, we fall back to shift at the

I've seen the code in question for #2 work on a mac and a windows xp
machine with Qt 4.3.4: selecting various dates in the last year, I can
get reliably the right offset-from-GMT.

i'm pretty sure the date-time dialog box isn't in its final
appearance, but it is serviceable now.


This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2008.
Stellarium-pubdevel mailing list