On 08/28/12 19:04, Alan Coopersmith wrote:
On 08/28/12 02:21 PM, Shawn Walker wrote:
Greetings,
The following webrev contains changes for the following RFE:
7140597 facets are ORed even between different top-level facet
categories
https://cr.opensolaris.org/action/browse/pkg/swalker/pkg-facet/webrev/
In your man page, the example manifest snippet uses foo.html, but the text
beneath describes "api.html"
That's what I get for rewriting something multiple times. I was unhappy
with each example case I wrote.
Also I don't understand that example:
file path=usr/share/doc/html/foo/foo.html facet.doc=all facet.doc.html=all
In the above example, api.html will only be installed if both the 'doc'
and 'html' facets are \fBtrue\fR in the image. This allows administrators
to exclude only HTML-format documentation without excluding plain text
documentation.
Wouldn't the existing system handle that case by using simply:
file path=usr/share/doc/html/foo/foo.html facet.doc.html=true
and then setting facet.doc.html=false ? Or having only html docs
by setting facet.doc*=false facet.doc.html=true ? I don't see the benefit of
setting all for a subset facet here.
There's a subtle difference to be fair. If an action is tagged with
'facet.doc.html=true', then setting 'facet.doc' will have no affect on
the image. Currently, only setting 'facet.doc.html' would have an effect.
Put differently, if you tag the action with both 'facet.doc' and
'facet.doc.html', a user could set 'doc=false' and 'doc.html=true' and
get only HTML documentation. That may not make much sense.
I think the only reason we use the '.' notation is to make relationships
and wildcards more sane. That is 'facet.doc*=true' or 'facet.doc*=false'.
You're right that the benefit of forcing the user to set both 'doc=true'
and 'doc.html=true' or 'doc*=true' is questionable.
The examples in the bug report make more sense to me:
facet.doc vs facet.locale
facet.devel vs facet.locale
facet.doc vs facet.devel
Perhaps that's what the api.html reference was going for?
file path=usr/share/doc/html/foo/api.html facet.doc=all facet.devel=all
It's what I had originally before I rewrote the text several times.
Here's my revised version:
file path=usr/share/doc/foo/foo.txt facet.doc=all
facet.locale.en_GB=true facet.locale.en_US=true
file path=usr/share/doc/foo/api.txt facet.doc=all
facet.devel=all
In the above example, foo.txt will only be installed if the
'doc' facet and at least one of the 'locale' facets is true
in the image. This allows administrators to exclude docu-
mentation, but still enable or disable support for specific
locales. In addition, api.txt will only be installed if
both the 'doc' and 'devel' facets are true in the image.
Also, pkg(1) always uses capitalized True& False for change-facet arguments,
not the lower-case versions that pkg(5) used - making the two consistent would
be nice.
I will leave that to Alta to decide on; there are many other cases of
true and false in pkg(5) that I'm not modifying here, so it seems better
to be consistent.
-Shawn
_______________________________________________
pkg-discuss mailing list
pkg-discuss@opensolaris.org
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
