On 02/21/12 09:06, Chris Quenelle wrote:
On Tuesday February 21, at 8:24AM, Shawn Walker wrote:
On 02/20/12 22:13, Chris Quenelle wrote:
On Monday February 20, at 7:39PM, Shawn Walker wrote:
On 02/20/12 15:09, Danek Duvall wrote:
Shawn Walker wrote:
I'd also like to suggest we have pkg.vendor to easily identify the
company that produced the package easily (e.g. set name=pkg.vendor
value=Oracle).
Eh, I'm not sold on that.
Alternative suggestions welcomed. Liane thought I should add this or an
equivalent as part of this proposal.
It came up recently with appcert(1) where it would have been nice to be able to
reliably identify Oracle-provided packages based on metadata.
In what context does appcert care if a package is supplied by Oracle?
Is it inferring a level of support or stability based on the information?
Oracle produces a diverse set of software all told.
One of the primary goals of appcert is to determine whether you are using
private interfaces in library files provided by Oracle and report
warnings/errors if you are. (Private being determined by how those symbols are
declared in the linker data.)
Hence, being able to know that a package was supplied by Oracle subjects it to
that restriction.
Likewise, it allows binaries provided by Oracle to link to private symbols
provided by Oracle.
Just to make sure I understand. Appcert worries about private interfaces in
all Oracle software, but not
private interfaces in non-Oracle software. I guess that makes sense.
Not sure it's justifies a special piece of metadata by itself. If so, then an attribute
with "appcert"
in the name might make more sense.
Identifying the provider of a package in a generic way seems useful
beyond appcert purely for reporting and/or identification purposes.
-Shawn
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
