Hi,

our ancient c++ class "TimeStamp" (describing who did something and when
he did it) from the tools library has a special value for an invalid
TimeStamp:

#define TIMESTAMP_INVALID_DATETIME
( DateTime ( Date ( 1, 1, 1601 ), Time ( 0, 0, 0 ) ) )

The whole magic behind that is that 01-01-1601 is taken as an invalid
date. I don't know why this date was chosen but perhaps somebody else
can shed some light on this.

In our UNOIDL we don't have a TimeStamp struct and so usually this is
expressed as two attributes, a css.util.DateTime and a string. The IDL
documentation says that setting the Day and Month members of the Date to
0 marks the date as "void". This term looks pretty strange to me but I
think it means that this is an unspecified date and so it comes close to
what we currently express by using our "invalid" TimeStamp.

This creates some problems for refactoring code from using C++ classes
to UNO services. If the refactored code had an external API that
provides a TimeStamp (like our SfxDocumentInfo) we would need to
automatically convert the "void" css.util.DateTime to an invalid tools
DateTime struct internally. This has the drawback that each DateTime
object has to be checked for being "void".

Alternatively we could change the definition of
TIMESTAMP_INVALID_DATETIME to be compatible with the "void" DateTime
from the UNOIDL. This should create problems only in two cases:

- somebody treats Day=0 or Month=0 as valid or uses it to express
something else

- somebody explicitly uses Date(1,1,1601) for whatever reason

I would say that we shouldn't care for this (perhaps we could grep for
"1601" in the source code) and change TIMESTAMP_INVALID_DATETIME.

Opinions, anyone?

Ciao,
Mathias

-- 
Mathias Bauer (mba) - Project Lead OpenOffice.org Writer
OpenOffice.org Engineering at Sun: http://blogs.sun.com/GullFOSS
Please don't reply to "[EMAIL PROTECTED]".
I use it for the OOo lists and only rarely read other mails sent to it.

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Reply via email to