* [feature] store env variables in History records
WIP: remove `HistoryWithoutDelete`, add some docstrings, tests
* Create History objects through builders.
Assure in compile-time that all required fields
are set for the given construction scenario
* (from #882) split Cmd::run into subfns
* Update `History` doc
* remove rmp-serde from history
* update warning
---------
Co-authored-by: Conrad Ludgate <conrad.ludgate@truelayer.com>
* wip
* Start testing
* Store host IDs, not hostnames
Why? Hostnames can change a lot, and therefore host filtering can be
funky. Really, all we want is a unique ID per machine + do not care what
it might be.
* Mostly just write a fuckload of tests
* Add a v0 kv store I can push to
* Appending works
* Add next() and iterate, test the pointer chain
* Fix sig
* Make clippy happy and thaw the ICE
* Fix tests'
* Fix tests
* typed builder and cleaner db trait
---------
Co-authored-by: Conrad Ludgate <conrad.ludgate@truelayer.com>
* Add `delete_at` is null condition during search
Since entries are searched everytime a delete process ends, the
--delete-it-all command will enter an infinite loop if searching the
whole history.
* Remove command blanking
Command blanking may violate the `unique(timestamp, cwd, command)`
condition.
* Overwrite command with random string when deleting
* Add rand dependency to client crate
---------
Co-authored-by: Ellie Huxtable <ellie@elliehuxtable.com>
Typo in config key `filter_mode_shell_up_key_binding` that's fixed by
this commit, and while at it fix a minor typo in a comment.
Co-authored-by: Per Modin <per@wgtwo.com>
* Put input chunk at the top in inline mode
* Invert the search results if bar is at top
* fix styling on reversed rendering
* add setting
* settings
---------
Co-authored-by: Patrick Decat <pdecat@gmail.com>
Co-authored-by: Conrad Ludgate <conrad.ludgate@truelayer.com>
* Stop running triggers on history delete
* Move to account management dir
* Alter trigger function to only run for inserts
* wip
* Add atuin account subcommands, and re-org delete
* Clarify docs
* Delete silly dupe migration
* Um where did this come from
* Oops, insert only plz
* At least patch this on the server side so we don't loop forever
* Postgres doesn't support <microsecond precision
Use millis - almost everything should support them and they are still
faster than a human can reasonably spam a button.
* Try different workspace structure
Move main crate (atuin) to be on the same level with other crates in
this workspace
* extract common dependencies to the workspace definition
* fix base64 v0.21 deprecation warning
* questionable: update deps & fix chrono deprecations
possible panic sites are unchanged, they're just more visible now
* Revert "questionable: update deps & fix chrono deprecations"
This reverts commit 993e60f8de.
Add `-r/--reverse` flag to `atuin search` to allow searching by oldest
results first.
Example to find the oldest `cargo` command:
```
atuin search --limit 1 --reverse cargo
```
* Use XDG data directory for fish import
On MacOS, atuin was looking for fish history under "$HOME/Library/Application Support".
Now atuin honors XDG_DATA_HOME, if set, and otherwise uses "$HOME/.local/share".
* cargo fmt
---------
Co-authored-by: Charles Gould <charles@gould.dev>
* Updated client config docs
- Example `config.toml` now includes all the newest options
- `settings.rs`, `config.toml`, and `config.md` now have uniform option order
* Remove trailing space
This flag allows the user to continue searching at an offset. This is
useful for building tools that use atuin to search for previous
commands and return only one result.
```
atuin search --limit 1
atuin search --limit 1 --offset 1
atuin search --limit 1 --offset 2
```
* feat: add *Nushell* support
* refactor: use `sh` to swap `STDOUT` and `STDERR` instead of using a temporary file
* feat: include both keybindings, with the current REPL buffer passed to *Atuin*'s
* feat: don't record commands run by keybindings
* Delete all instances of a command
Our search command will de-dupe results by default. But... This isn't
great for deleting! You don't want to run it over-and-over-and-over
until all commands are deleted.
Loop the query, and keep on deleting what it returns until they are all
gone.
* Optimize delete upload
It was running a request for every element, on every sync lol
Only push a delete if needed
Future: push all deletes in one request
* Make search_mode a part of SearchState
* Allow changing search mode using ctrl+s
* Tweak state reset for switched_search_mode
* Improve search_mode display in interactive mode
* Incorporate review suggestion
* Tweak language
* Fix Clippy and format
* Support old msgpack
I forgot it isn't backwards compatible... This should fix any sync
issues resulting from the deletion PR
* Update atuin-client/src/encryption.rs
Co-authored-by: Conrad Ludgate <conradludgate@gmail.com>
* Bye bye unwrap
---------
Co-authored-by: Conrad Ludgate <conradludgate@gmail.com>
* Drop events. I'd still like to do them, but differently
* Start adding delete api stuff
* Set mailmap
* Delete delete delete
* Fix tests
* Make clippy happy
* skim-demo
* skim some more
* Weight first word match higher (#712)
* some improvements
* make skim opt-in
---------
Co-authored-by: Frank Hamand <frankhamand@gmail.com>
* Fix before/after combined with limit
Mixing filters done in Rust with filters done in SQL is _no bueno_.
Been meaning to do this for a while anyways. Search params are getting a
bit fat but oh well!
* Make an excuse for a big function sig
* Do options map_or not if
* Fix tests
* feat: add common default keybindings
* feat: add `WORD_SEPARATORS` to config as `word_chars`, as this is what *Zsh* calls it
* feat: add option for *Emacs* word jumping
* feat: scroll with `PageUp` and `PageDown`, cf #374
* Only show help if terminal strictly has more than one line
* There is no border around the input line in compact mode
* Add command preview
* Dynamic preview height
* Avoid extra allocations
Co-authored-by: Conrad Ludgate <oon@conradludgate.com>
* Address clippy error
* Merge normal and compact views code
* Add show_preview setting, disabled by default
* Convert `bool` to `int` with `if` for legibility
---------
Co-authored-by: Conrad Ludgate <oon@conradludgate.com>
* Rework Bash import
Closes#745
- Imported history is now ordered correctly
- Timestamps (when `HISTTIMEFORMAT` is set) are handled correctly
* Timestamp tests test for strict sorting
