diff options
| author | Tomasz Maczukin <tomasz@maczukin.pl> | 2019-06-11 23:31:44 +0200 |
|---|---|---|
| committer | Tomasz Maczukin <tomasz@maczukin.pl> | 2019-06-12 08:44:17 +0200 |
| commit | f4b5f930ff7252468c69cf4decbbb05d44ba3b71 (patch) | |
| tree | 25facd3466acfa641ffc427b6b633f701b10bae9 | |
| parent | 9bf141cf718ab9ca770c3bd6b7e2c6719cbe9da7 (diff) | |
| download | gitlab-ce-docs/make-it-clear-about-after-script-context.tar.gz | |
Leave clear notices about after_script execution contextdocs/make-it-clear-about-after-script-context
We frequently see new issues mentioning that something defined
in `before_script` or `script` doesn't work with `after_script`.
Our documentation mentions that first two are concatenated and
executed together, why the third one is executed in a separated
shell context. But it looks like it's not clear enough.
Since this misunderstanding of how scripts are being executed
is repeated frequently, let's make it more clear in docs.
| -rw-r--r-- | doc/ci/variables/where_variables_can_be_used.md | 8 | ||||
| -rw-r--r-- | doc/ci/yaml/README.md | 18 |
2 files changed, 22 insertions, 4 deletions
diff --git a/doc/ci/variables/where_variables_can_be_used.md b/doc/ci/variables/where_variables_can_be_used.md index 8009b1d5e8a..7d3f39a8c19 100644 --- a/doc/ci/variables/where_variables_can_be_used.md +++ b/doc/ci/variables/where_variables_can_be_used.md @@ -89,6 +89,14 @@ Supported: - In `script`, it will work in the following lines of `script`. - In `after_script`, it will work in following lines of `after_script`. +Please notice the specific case of `after_script` scripts, that can: + +- Only use variables defined before the script within the same `after_script` section. +- Not use variables defined in `before_script` and `script`. + +Both restrictions are caused by the fact, that `after_script` script is executed in a +[separated shell context](https://docs.gitlab.com/ee/ci/yaml/README.html#before_script-and-after_script). + ## Persisted variables NOTE: **Note:** diff --git a/doc/ci/yaml/README.md b/doc/ci/yaml/README.md index fa4b0378f61..cb7a93e0cc7 100644 --- a/doc/ci/yaml/README.md +++ b/doc/ci/yaml/README.md @@ -215,10 +215,20 @@ This can be an array or a multi-line string. `after_script` is used to define the command that will be run after all jobs, including failed ones. This has to be an array or a multi-line string. -The `before_script` and the main `script` are concatenated and run in a single context/container. -The `after_script` is run separately. The current working directory is set back to -default. Depending on the executor, changes done outside of the working tree might -not be visible, e.g. software installed in the `before_script`. +Script specified in `before_script` is: + +- Concatenated with script specified in the main `script`. Job-level `before_script` definition + override global-level `before_script` definition when concatenated with `script` definition. +- Executed together with main `script` script as one script in a single shell context. + +Script specified in `after_script`: + +- Have a current working directory set back to the default. +- Is executed in a shell context separated from `before_script` and `script` scripts. +- Because of separated context, cannot see changes done by scripts defined in `before_script` or `script` scripts: + - in shell - for example, command aliases and variables exported in `script` script, + - outside of the working tree (depending on the Runner executor) - for example, software installed + by a `before_script` or `script` script. It's possible to overwrite the globally defined `before_script` and `after_script` if you set it per-job: |