~mfoord/juju-core/tests-replset : changes

~mfoord/juju-core/tests-replset » Changes from revision 1010.3.1

From Revision 1010.3.1 to 992

Rev	Summary	Authors	Date
1010.3.1	checkpoint: tests pass; major refactoring to avoid import cy... checkpoint: tests pass; major refactoring to avoid import cycle completed	Benji York	11 years ago
1010	Add ServiceDestroyUnits to the API Add ServiceDestroyUnits to the API R=fwereade, dimitern, rog CC= https://codereview.appspot.com/7554046	Benji York	11 years ago
1009	state/watcher: fix bug in UnwatchCollection state/watcher: fix bug in UnwatchCollection An outstanding request being sent to a collection watch would not be cancelled by UnwatchCollection. R=dimitern, dfc, jameinel, fwereade CC= https://codereview.appspot.com/7610044	Roger Peppe	11 years ago
1008	charm: stream SHA256 of charm charm: stream SHA256 of charm There's no need to have the whole thing in memory at once. R=dimitern, bac, jameinel, dfc CC= https://codereview.appspot.com/7417051	Roger Peppe	11 years ago
1007	state: stricter txn assertions on service destroy state: stricter txn assertions on service destroy As originally formulated, a changed unit count would not be detected if the relation count was unchanged; this engendered lp:1152717. We now check rather more strictly. Several service destruction tests were rearranged and renamed for clarity. R=dimitern, rog CC= https://codereview.appspot.com/7591044	William Reade	11 years ago
1006	state: unit assignment respects machine series state: unit assignment respects machine series That is: it is no longer possible to assign a unit to a machine with the wrong series, and machines created for the use of a given unit now take that unit's series. A "Series" field was added to the unit document to make it possible to check this without way-too-many roundtrips. R=dimitern, dfc, rog CC= https://codereview.appspot.com/7715046	William Reade	11 years ago
1005	cmd/juju: add shared juju version command cmd/juju: add shared juju version command In Juju charm testing it would be great if we could tell the version of juju we are using. The version command was only available on jujud, so I have moved it into a shared location and added it juju as well. R=thumper, dimitern CC= https://codereview.appspot.com/7794045	Dave Cheney	11 years ago
1004	Make tests pass on raring Make tests pass on raring Raring ships with a later version of git (1.8.x). This newer version requires user and email to be known before committing and on some raring setups, the info could not be found. So I added a test fixture to ensure the info was available when the tests are run. R=dfc, thumper CC= https://codereview.appspot.com/7677044	Ian Booth	11 years ago
1003	Add a --format flag for jujuc commands missing it Add a --format flag for jujuc commands missing it Fixes lp:1129130. Also marks the --format flag as deprecated for commands that don't produce output. R=dfc, dimitern CC= https://codereview.appspot.com/7795045	Tim Penhey	11 years ago
1002	environs/config: ensure ~/.juju exists environs/config: ensure ~/.juju exists When doing WriteEnviron we would create ~/.juju/environments.yaml but we weren't creating the containing dir. (Which meant actually running 'juju init -w' wouldn't work for new users.) This adds an unconditional MkdirAll() before writing environments.yaml, but that should be ok, as it is a no-op if the dir already exists. R=dfc, dimitern, rog CC= https://codereview.appspot.com/7759044	John A Meinel	11 years ago
1001	state: add allWatcher.handle method state: add allWatcher.handle method This handles a request from a StateWatcher and adds the request to the set of outstanding requests as appropriate. R=dimitern, fwereade CC= https://codereview.appspot.com/7727045	Roger Peppe	11 years ago
1000	state: add allWatcher state: add allWatcher This is the data structure that will manage the shared state for all watching clients. Initially, we have a single method, "change", which mutates the state in response to a watcher-for change. For the record, this is the allWatcher inner loop as planned: for { select { case <-aw.tomb.Dying(): return tomb.ErrDying case change := <-in: if err := aw.changed(change); err != nil { return err } case req := <-aw.request: aw.handle(req) } aw.respond() } R=dimitern, fwereade CC= https://codereview.appspot.com/7727044	Roger Peppe	11 years ago
999	various: use path instead of filepath for Windows various: use path instead of filepath for Windows cleanup one of the imports that had conflicts with variable names. R=rog, dimitern CC= https://codereview.appspot.com/7813043	John A Meinel	11 years ago
998	various: use path instead of filepath for Windows various: use path instead of filepath for Windows (yes, that sounds backwards, but it is correct) There are quite a few places where code uses 'filepath' to generate a path string. However it is generating the path on the local client but then sending that path to the remote location (such as cloudinit). This means that when running on Windows, it was telling the new instance to create things like '\var\lib\juju\bootstrap\agent.conf' which didn't work so well :). With this patch, and the patch to goyaml, and a small hack to force the series to "precise" and the Arch to "amd64", I'm able to run "juju bootstrap" and "juju status" on Windows. The test suite doesn't run (because we don't have mongod) and I haven't tested deploying any actual charms (the next place I expect to be a problem). But this seems a reasonable start. The test suite passes on Linux still, so the change seems safe enough. There are still places that use 'filepath' in the code, but I tried to audit that they were only using those paths locally. Also, Windows does need '\' style paths for commands that you execute, though you can use '/' for any files you need to open, etc. R=dimitern, dfc CC= https://codereview.appspot.com/7813043	John A Meinel	11 years ago
997	Refactor the cmd/juju tests. Refactor the cmd/juju tests. Remove the local makeFakeHome function from the cmd/juju package, and use the one defined in testing/environs. Also refactor the help function in the testing package to allow explicit creation of the fake home without any environments.yaml set. R=dfc, jameinel, TheMue CC= https://codereview.appspot.com/7744044	Tim Penhey	11 years ago
996	Fix the other test that checks smart output. My fault for n... Fix the other test that checks smart output. My fault for not running the full test suite earlier.	Tim Penhey	11 years ago
995	Update the smart formatter to match pyjuju Update the smart formatter to match pyjuju The smart formatter has been updated to return "True" and "False" for boolean fields rather than "true" and "false". The tests have been updated too. R=dfc, jameinel CC= https://codereview.appspot.com/7764045	Tim Penhey	11 years ago
994	Add ServiceDestroy to the API Add ServiceDestroy to the API Implemented destroy service as a statecmd and added it to the API. R=dfc, dimitern CC= https://codereview.appspot.com/7671046	Matthew Scott	11 years ago
993	Add Resolved to the API Add Resolved to the API R=dimitern, fwereade, rog CC= https://codereview.appspot.com/7460047	Benji York	11 years ago
992	trivial/Timeit: Add a helper for timing tests trivial/Timeit: Add a helper for timing tests While we have '-cpuprofile' available, it doesn't give useful information when that time is spent inside time.Sleep(). (We don't spend any CPU time while sleeping. :) Since we don't have a tracing profiler, this gives a rough approximation of that. You have to manually insert trace points, but it is still helpful. I used this in investigating some test performance, and felt it would be useful to keep around for future performance investigation. R=dimitern, TheMue CC= https://codereview.appspot.com/7537044	John A Meinel	11 years ago

Older »