[tz] NetBSD and Android bionic APIs for tz values

Alan Barrett apb at cequrux.com
Mon Aug 25 10:53:33 UTC 2014


On Mon, 25 Aug 2014, Christos Zoulas wrote:
>On Aug 25, 12:32am, eggert at cs.ucla.edu (Paul Eggert) wrote:
>| In doing so it struck me that the proposal for new functions tm_strftime
>| and tm_strftime_l is overkill, and that it's better to simply fix
>| strftime and strftime_l to do the right thing with %z and %Z, as that's
>| what glibc does.  I understand there's an argument that POSIX doesn't
>| allow this behavior, but it's not clear to me that the argument is
>| correct, and anyway NetBSD would be in good company when it's compatible
>| with tzcode and with GNU/Linux here.  So attached is a a revised patch
>| that simply removes strftime_z/strftime_lz and fixes
>| strftime/strftime_l.  The rest of the patch is the same as before.

Thanks.  For tzcode, I think that makes sense.

>Thanks Paul, looks fine to me. Since we can't really bump libc (and
>the function removal can break things; even the const removal changes
>the signature for c++, but I guess that's not that important) the
>choices are:
>	- remove the functions and if older code used them the code
>	  will stop working
>	- keep the functions, but put them in compat, removing them
>	  from headers.
>
>I guess I'll bring it up and see what others have to say.

Yes, we should discuss that in a NetBSD mailing list.

--apb (Alan Barrett)


More information about the tz mailing list