This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

[v3] time facets

This removes some unnecessary inheritance in the locale facets. 

It also changes the non-standard member functions of time_get to be
consistent with the other facets.

There is no need to add a __timepunct_cache here, as __timepunct itself
is not a datum that uses virtual functions. Therefore, I've omitted this
part of the patch. 

If the LWG changes this at some point, I believe it can be added without
API changes to time_get or time_put. Of course, if the LWG changes or
clarifies this, changing time_get and time_put without changing their
API's will not really be super important....

In general, I think the GNU implementation and design of the time and
date facets to be credible and reasonable, resulting in usage consistent
with the other locale facets, although the time and date facets seem to
vary widely between implementations. Clarifying elements of the date and
time facets is something I will bring up within the LWG after this

tested x86/linux
tested powerpc-apple-darwin7.2.0

Attachment: p.20040323-1.patch
Description: Text document

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]