On more thought, how about in addition to or in replacement of
PropList/PropSimpleList PropSet or PropSimpleSet?
Brock
On 06/29/10 05:59 PM, Brock Pytlik wrote:
doc/on-disk-format.txt:
84: nit Shouldn't this be 3.1 (and the following sections renumbered)
depot.py:
The exit ready option is being removed or at least undocumented?
pkg.depotd.1m.txt:
since we're changing almost all of this file anyway, can we change the
format slightly so that the offset isn't dependent on the length of
the longest option? I think something like what the grep man page (at
least on my system) does makes sense. It looks like:
<short option>, <long option>
<offset text description>
pkgrepo.1.txt:
Can pkgrepo set publisher-specific information?
server/depot.py:
1690-1736: I think this might be easier to maintain/understand if the
different sections were broken up into separate pieces, then combined
together.
Has imageconfig not been changed to use the new config.py module
simply because that's in a following phase, or is there a different
plan for imageconfig?
pkgrepo.py:
273,274: If we're going to examine all the sections and properties
anyway, why set the initial values to 7 and 8?
Thanks,
Brock
On 06/28/10 11:01 PM, Shawn Walker wrote:
Greetings,
The following webrev implements fixes and/or changes for the
following bugzilla entries:
8722 advanced repository metadata store needed
15528 repository administration and configuration tool
16264 client operations fail if non-existent file repository is
configured
16397 repository.__set_file_root references unbound EACCESS
16404 coverage reports miss class initialization and module setup
Summary of changes:
8183 lines changed: 6101 ins; 1673 del; 409 mod; 28029 unchg
webrev:
http://cr.opensolaris.org/~swalker/pkg-format/
Summary:
* added a new set of configuration classes to simplify configuration
metadata management, storage, and retrieval
* fixed overzealous URI validation that could cause client to fail
if a file repository was removed after it was configured for an
image
* fixed a traceback that could happen in an permissions error path
for repositories
* fixed coverage reporting to include class initialisation and module
setup (7x% to 9x% for the configuration classes)
* added on-disk format design document
* changed depot to read configuration data directly from SMF if
possible (in preparation for a further enhancement to implement
a restarter method that sends SIGUSR1 to the depot process for
svcadm refresh)
* greatly simplified test suite code for executing cmdline programs
Cheers,
-Shawn
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
_______________________________________________
pkg-discuss mailing list
[email protected]
http://mail.opensolaris.org/mailman/listinfo/pkg-discuss
