Tim Foster wrote:
> This looks good to me. We've still got a lot of places in the code that
> refer to 5.11. As I understand it, the majority of strings are only
> used in places where we're creating a pkg.fmri.PkgFmri or a
> pkg.version.Version and the build string isn't also declared in the fmri
> string.
>
> Would it make sense to use a global variable there instead of '5.11', or
> are we happy to leave it as is, and expect that s12 release engineers
> will use a pkglint check to flag build_release values == [5, 11] for any
> packages that are being published?
I've whipped together (by which I mean it mostly works) another patch that
simply changes the Version constructor to default build_version to "5.12",
and removes most of the hard-coded "5.11" instances in the code to simply
allow the constructor to set the default. I can't say I ever agreed with
the requirement that it be explicitly passed in. This reduces the places
it's hardcoded to four. But I'd rather do that as a separate change,
unless folks think it's something that should happen right now. And I'm
still not sure whether it's worth carrying this thing around all the time,
though I don't think it hurts much.
> Also, a nit - should
>
> > + nawk '$$1 ~ /^s12b[0-9]*/ {exit} END {print substr($$1,
>
> actually be:
>
> > + nawk '$$1 ~ /^s12b[0-9]+/ {exit} END {print substr($$1,
>
> s12b(one or more characters in the range 0-9), not s12(zero or more..)
Cool, thanks.
Danek
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
