11:25 p.m.
repo: baserock:baserock/morph
branch: danielfirth/RT257
sha1: dd41de8ff3d6729253dbea546f39a859123ed531
7:32 p.m.
On Tue, Oct 29, 2013 at 02:25:35PM +0000, Dan Firth wrote:
+Implementation sections for deployment
+======================================
+
+ IMPLEMENTS WHEN attempting to deploy the (system|cluster) (\S+) in branch (\S+)
+ cd "$DATADIR/workspace/$MATCH_3"
+ attempt_morph deploy "$MATCH_2"
+
Implementations sections for reading error messages
===================================================
- IMPLEMENTS THEN the (init|build|checkout|branch) error message includes the string
"(.*)"
+ IMPLEMENTS THEN the (init|build|checkout|branch|deploy) error message includes the
string "(.*)"
grep "$MATCH_2" "$DATADIR/result-$MATCH_1"
IMPLEMENTS for test file and directory handling
The implementation itself looks sound, but as per discussions yesterday,
we want more generic implementations, so I'm voting -1.
7:56 p.m.
On 30/10/13 10:32, Richard Maw wrote:
The implementation itself looks sound, but as per discussions yesterday,
we want more generic implementations, so I'm voting -1.
As I remember, we agreed that we would do the error hashing approach at
a separate time. It's a non-trivial amount of work to tack onto a bugfix.
Can we please get the stuff that's in flight in and refactor all of this
error handling in a separate ticket?
9:08 p.m.
On 29/10/13 14:25, Dan Firth wrote:
repo: baserock:baserock/morph
branch: danielfirth/RT257
sha1: dd41de8ff3d6729253dbea546f39a859123ed531
In my opinion it's okay (+1). The tests pass, and is clear. But maybe we
should try to follow the Paul suggestions for the new IMPLEMENTS
Could be better:
IMPLEMENTS WHEN the user attemps to deploy the (system|cluster) (\S+) in
branch (\S+)
instead of:
IMPLEMENTS WHEN attempting to deploy the (system|cluster) (\S+) in
branch (\S+)
Here is the example Paul gave to me:
Here's my attempt:
SCENARIO re-running 'morph branch' fails, original branch untouched
GIVEN a workspace
AND a git server
WHEN user creates system branch foo
THEN system branch foo is checked out
WHEN user modifies system branch foo
AND user creates system branch foo
THEN morph fails because foo exists
AND modified system branch foo is still checked out
3617
days inactive
3618
days old
4 comments
3 participants
participants (3)
-
Dan Firth -
Pedro Alvarez -
Richard Maw