[pkg-discuss] Re: code review request: pkg dehydrate/rehydrate
- From: Yiteng Zhang <
- Cc: Bart Smaalders <
>, Erik Trauschke <
>, Shawn Walker <
>, Danek Duvall <
- Subject: [pkg-discuss] Re: code review request: pkg dehydrate/rehydrate
- Date: Thu, 19 Jun 2014 11:55:03 -0700
As discussed in the pkg meeting, I implemented most of the idea we
It will be fast to have a review on this webrev if you take a look on
1) "dehydrated" property with the value of dehydrated publishers will be set
on the image, and it will be deleted if all publishers are rehydrated.
2) Any pkg operation are allowed after pkg dehydrate/rehydrate and tested.
3) dehydrate will *not* check the "dehydrated" property, so dehydrate
go through all the specified and configured publishers even they were
"dehydrated". Then dehydrate will check the existence of the files that
to remove. If they exist, propose them to remove. If not, do nothing.
4) rehydrate will check the "dehydrated" property. If a publisher has
been marked as
"dehydrated", it will install the removed files from this publisher. If
not, it does nothing
about this publisher. Then rehydrate will *not* check the existence of
the files that
are supposed to install, so rehydrate will always install them.
5) For pkg fix, I implemented an approach for printing stuff like this:
defined some abstract
methods in ProgressTrackerBackend and ProgressTrackerFrontend, and
methods in their subclasses(ProgressTracker, CommandLineProgressTracker and
FancyUNIXProgressTracker). I think it is consistent with the way we do
for pkg verify.
on-12clone webrev: https://ips.java.net/webrev/yitezhan/on12-clone-hydrate/