mirror of
https://github.com/martinvonz/jj.git
synced 2025-01-19 19:08:08 +00:00
config: update comment about new config file creation
Spotted while updating get_new_config_file_path() callers. "jj config path" isn't a command to create new config file.
This commit is contained in:
parent
ecf25a1d9b
commit
e1cd1a00e6
1 changed files with 7 additions and 5 deletions
|
@ -267,7 +267,12 @@ impl ConfigEnv {
|
|||
match &self.user_config_path {
|
||||
ConfigPath::Existing(path) => Ok(Some(path)),
|
||||
ConfigPath::New(path) => {
|
||||
// TODO: should we update self.user_config_path to Existing?
|
||||
// TODO: Maybe we shouldn't create new file here. Not all
|
||||
// callers need an empty file. For example, "jj config path"
|
||||
// should be a readonly operation. "jj config set" doesn't have
|
||||
// to create an empty file to be overwritten. Since it's unclear
|
||||
// who and when to update ConfigPath::New(_) to ::Existing(_),
|
||||
// it's probably better to not cache the path existence.
|
||||
create_config_file(path)?;
|
||||
Ok(Some(path))
|
||||
}
|
||||
|
@ -308,10 +313,7 @@ impl ConfigEnv {
|
|||
pub fn new_repo_config_path(&self) -> Option<&Path> {
|
||||
match &self.repo_config_path {
|
||||
ConfigPath::Existing(path) => Some(path),
|
||||
ConfigPath::New(path) => {
|
||||
// TODO: should we create new file?
|
||||
Some(path)
|
||||
}
|
||||
ConfigPath::New(path) => Some(path),
|
||||
ConfigPath::Unavailable => None,
|
||||
}
|
||||
}
|
||||
|
|
Loading…
Reference in a new issue