-
Notifications
You must be signed in to change notification settings - Fork 2.3k
New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Newest preexec version disables ignorespace and ignoreboth #1844
Comments
@davidpfarrell @NoahGorny could you take a look at this? |
I left a comment on the linked issue. For now, I think we need to remove references to the function and claim that we don't support |
@davidpfarrell I'll make a PR while waiting for @NoahGorny's input. Thank you very much for reviewing this! |
@gaelicWizard I know you've since taken an interest in this issue. I want to finally close it out. I didn't really understand what you were trying to suggest for a solution. FWIW, I'm not looking to get into a drawn out solution, I just want something that works now, and can be improved later. @NoahGorny This relates to the other issue where I brought up patching vendored libs. The paths forward that I considered are:
Considering the overhead of multiple repos for a small use case, I'm leaning towards the first option. I look forward to reading everyone's opinions. cc: @davidpfarrell |
My rcaloras/bash-preexec#119 I think is the closest to "best of both worlds" as it allows all relying parties to get what they want in different use cases. I don't have anything to add beyond what I submitted there. |
Expected Behavior
When
HISTCONTROL
is set toignorespace
orignoreboth
, I expect commands with a preceding space to not be stored in history. This is a common expectation to work around situations where it's just easiest to enter a secure string as an argument to a command, and you don't want it stored in history.Edit: before this inspires a bunch of comments about how this is bad practice, how the secrets are leaked in
ps
, etc; none of that is relevant. This is a behavior that the user expects to be supported and because they are unable to opt-out of our use of preexec, this should be addressed.Current Behavior
The updated preexec library edits
HISTCONTROL
with the __bp_adjust_histcontrol function.Possible Solution
See rcaloras/bash-preexec#115
Steps to Reproduce
The text was updated successfully, but these errors were encountered: