* add v to version if it does not have a leading v
Signed-off-by: kpenfound <kyle@dagger.io>
* test version without the v
Signed-off-by: kpenfound <kyle@dagger.io>
---------
Signed-off-by: kpenfound <kyle@dagger.io>
* avoid overwriting existing dagger install unless force-install is true
Signed-off-by: kpenfound <kyle@dagger.io>
* add force-install to version tests
Signed-off-by: kpenfound <kyle@dagger.io>
* only skip install if we requested the same version that we already had
Signed-off-by: kpenfound <kyle@dagger.io>
* fix dagger check logic
Signed-off-by: kpenfound <kyle@dagger.io>
* remove incorrect and unused default for engine-stop
Signed-off-by: kpenfound <kyle@dagger.io>
* add tests for commit and latest
Signed-off-by: kpenfound <kyle@dagger.io>
* wrong commit in commit test
Signed-off-by: kpenfound <kyle@dagger.io>
---------
Signed-off-by: kpenfound <kyle@dagger.io>
* add options for shell
Signed-off-by: kpenfound <kyle@dagger.io>
assemble dagger command in new step with output
Signed-off-by: kpenfound <kyle@dagger.io>
fixing some inline bash
Signed-off-by: kpenfound <kyle@dagger.io>
wrap conditional inputs in quotes
Signed-off-by: kpenfound <kyle@dagger.io>
use cat for heredoc script file
Signed-off-by: kpenfound <kyle@dagger.io>
missing space in heredoc
Signed-off-by: kpenfound <kyle@dagger.io>
add tab ignore and dagger shebang
Signed-off-by: kpenfound <kyle@dagger.io>
use toJson and jq to safely write the shell input to a file
Signed-off-by: kpenfound <kyle@dagger.io>
put DAGGER_COMMAND in a variable
Signed-off-by: kpenfound <kyle@dagger.io>
missing semicolon for the dagger command
Signed-off-by: kpenfound <kyle@dagger.io>
safely evaluate if shell is set
Signed-off-by: kpenfound <kyle@dagger.io>
extra > in shell file write
Signed-off-by: kpenfound <kyle@dagger.io>
add some logging
Signed-off-by: kpenfound <kyle@dagger.io>
strip extra newline from jq when shell is not set
Signed-off-by: kpenfound <kyle@dagger.io>
debugging
Signed-off-by: kpenfound <kyle@dagger.io>
toJSON and jq will send null instead of empty string with no input
Signed-off-by: kpenfound <kyle@dagger.io>
safely handle all inputs
Signed-off-by: kpenfound <kyle@dagger.io>
did i fix the wrong thing?
Signed-off-by: kpenfound <kyle@dagger.io>
change up the approach a bit for an easier diff
Signed-off-by: kpenfound <kyle@dagger.io>
fix shell test
Signed-off-by: kpenfound <kyle@dagger.io>
single quotes around shell input
Signed-off-by: kpenfound <kyle@dagger.io>
try with piped input
Signed-off-by: kpenfound <kyle@dagger.io>
try single quotes again
Signed-off-by: kpenfound <kyle@dagger.io>
debug
Signed-off-by: kpenfound <kyle@dagger.io>
debug
Signed-off-by: kpenfound <kyle@dagger.io>
fix shell test
Signed-off-by: kpenfound <kyle@dagger.io>
does shell emit a newline?
Signed-off-by: kpenfound <kyle@dagger.io>
remove the trailing newline if its in stdout
Signed-off-by: kpenfound <kyle@dagger.io>
break the trailing output test
Signed-off-by: kpenfound <kyle@dagger.io>
fix shell test
Signed-off-by: kpenfound <kyle@dagger.io>
missed !
Signed-off-by: kpenfound <kyle@dagger.io>
add comment about weird test
Signed-off-by: kpenfound <kyle@dagger.io>
* use echo -n in shell test
Signed-off-by: kpenfound <kyle@dagger.io>
---------
Signed-off-by: kpenfound <kyle@dagger.io>
* test: add versioning tests
Signed-off-by: Justin Chadwell <me@jedevc.com>
* feat: add output for action
To chain on from the github action, it's very useful to be able to get
the output from an action.
Signed-off-by: Justin Chadwell <me@jedevc.com>
---------
Signed-off-by: Justin Chadwell <me@jedevc.com>
