Seperate remote_path into remote_folder and remote_script. #3462
Conversation
|
Failure on appveyor seems unrelated. @markpeek @rickard-von-essen @cbednarski does this look OK to you? :) |
|
|
||
| - `remote_file` (string) - The filename the uploaded script will have on the machine. | ||
| This defaults to 'script_nnn.sh'. | ||
|
|
||
| - `remote_path` (string) - The filename where the script will be uploaded |
There was a problem hiding this comment.
remote_path now has the same documentation as remote_file, can you include some language to distinguish them ("full path to the file") and also what happens if all three are specified / recommend which ones should be used in which cases?
There was a problem hiding this comment.
Thanks for the feedback @cbednarski I will update the documentation accordingly.
In the event all three or set the remote_path will overwrite both remote_file and remote_folder. I only kept remote_path in order to not break peoples packer.json files.
|
@imduffy15 Thanks for the PR, and for updating the docs! The change LGTM but would you be able to add a test that verifies the change? I think it's sufficient to provide two configuration blocks that demonstrate each case, parse them, and validate that the expected path comes out. You can move this logic to a new function if that makes it easier to test. |
|
Here's an example test that does config parsing: https://github.com/mitchellh/packer/blob/master/post-processor/compress/post-processor_test.go#L153-L182 |
Commonly /tmp is set as noexec, as a result packer scripts fail to run. In order to get around this one can set a remote_path, whoever, remote path requires full filename and path. By making remote_path a combination of remote_folder/remote_script we can change remote_folder and keep the default script_nnn.sh Signed-off-by: Ian Duffy <ian@ianduffy.ie>
|
Updated with test cases to ensure the following:
|
|
Looks great! Thanks for adding the tests and docs. :) |
|
Why is the default directory |
|
In terms of the linux file system and folder structure I feel /tmp is a In my situtation, /tmp was mounted on the system with noexec for security I do not believe / would make a sane default, / is generally just On 20 May 2016 at 18:53, Sander Maijers notifications@github.com wrote:
|
|
Assuming Linux, take your pick based on FHS. |
…#3462) Commonly /tmp is set as noexec, as a result packer scripts fail to run. In order to get around this one can set a remote_path, whoever, remote path requires full filename and path. By making remote_path a combination of remote_folder/remote_script we can change remote_folder and keep the default script_nnn.sh Signed-off-by: Ian Duffy <ian@ianduffy.ie>
ghost
locked and limited conversation to collaborators
Commonly /tmp is set as noexec, as a result packer scripts
fail to run. In order to get around this one can set a
remote_path, whoever, remote path requires full filename
and path.
By making remote_path a combination of
remote_folder/remote_script we can change remote_folder
and keep the default script_nnn.sh