This commit makes adding icons to file names something that the file name renderer does, rather than something that each individual view does. This is now possible thanks to the previous commit a1869f2, which moved the option to do this into the same module. The repeated code has been removed.
It happens to fix a bug where the width of each column was being incorrectly calculated for the grid-details view, making lines slightly too long for the terminal because the icon wasn't being taken into account.
All four of the view mode command-line argument parsers tested for the --icons option. Because it was common, the behaviour has been moved to the struct that handles file styles, meaning it can be parsed in one place.
This is a better place for it, as the icons are to do with the file name, not the view. It also means that the lines view has no options left for it, which is fitting.
This commit significantly refactors the way that options are parsed. It introduces the Theme type which contains both styling and extension configuration, converts the option-parsing process into a being a pure function, and removes some rather gnarly old code.
The main purpose of the refactoring is to fix GH-318, "Tests fail when not connected to a terminal". Even though exa was compiling fine on my machine and on Travis, it was failing for automated build scripts. This was because of what the option-parsing code was trying to accomplish: it wasn't just providing a struct of the user's settings, it was also checking the terminal, providing a View directly.
This has been changed so that the options module now _only_ looks at the command-line arguments and environment variables. Instead of returning a View, it returns the user's _preference_, and it's then up to the 'main' module to examine the terminal width and figure out if the view is doable, downgrading it if necessary.
The code that used to determine the view was horrible and I'm pleased it can be cut out. Also, the terminal width used to be in a lazy_static because it was queried multiple times, and now it's not in one because it's only queried once, which is a good sign for things going in the right direction.
There are also some naming and organisational changes around themes. The blanket terms "Colours" and "Styles" have been yeeted in favour of "Theme", which handles both extensions and UI colours. The FileStyle struct has been replaced with file_name::Options, making it similar to the views in how it has an Options struct and a Render struct.
Finally, eight unit tests have been removed because they turned out to be redundant (testing --colour and --color) after examining the tangled code, and the default theme has been put in its own file in preparation for more themes.
This was being passed around everywhere as a parameter, when it can exist just as nicely as a struct field. This means many functions can take one argument less.
This was meant to be a small change, but it spiralled into a big one.
The original intention was to separate OptionsResult and OptionsError. With these types separated, the Help and Version variants can only be returned from the Options::parse function, and the later option-parsing functions can only return success or errors.
Also, Misfire was a silly name.
As a side-effect of Options::parse returning OptionsResult instead of Result<Options, Misfire>, we could no longer use unwrap() or unwrap_err() to get the contents out. This commit makes OptionsResult into a value type, and Options::parse a pure function. It feels like it should be one, having its return value entirely dependent on its arguments, but it also loaded locales and time zones. These parts have been moved into lazy_static references, and the code still passes tests without much change.
OptionsResult isn't PartialEq yet, because the file colouring uses a Box internally.
I read through every file and applied a couple of rustfmt suggestions. The brace placement and alignment of items on similar lines has been made consistent, even if neither are rustfmt's default style (a file has been put in place to enforce this). Other changes are:
• Alphabetical imports and modules
• Comma placement at the end of match blocks
• Use newlines and indentation judiciously
• Spaces around associated types
• Spaces after negations (it makes it more clear imho)
• Comment formatting
• Use early-returns and Optional `?` where appropriate
This commit makes changes to the way variables are referenced:
• Make types Copy when possible
• Make methods take `self` instead of `&self` where possible (trivially_copy_pass_by_ref)
• Remove unnecessary borrowing (needless_ref)
• Remove unnecessary cloning (clone_on_copy)
• Remove `ref` from match arms where possible (new Rust match ergonomics)
This commit removes the env_logger dependency, replacing it with a simple implementation. Doing so removes like ten other transitive dependencies that no longer need to be included in the build.
It also gains the ability to enable trace-level logging. The users crate, which contains such logging statements as of the version I published a few days ago, has been upgraded to celebrate.
Also, change the log imports to globs. I'm only interested that a file doing logging, not what level it's logging at.
This was an unintended consequence of #653. The Files iterator stopped using IgnoreCache and started using GitCache, which would always populated when the `--git` option was passed, without checking whether files were meant to be ignored. This meant that passing `--git` started ignoring files even without `--git-ignore`.
The solution for now is to explicitly pass the flag around, which probably should be a better type than bool but isn't. This makes the git-ignoring-related extended tests pass.
This improves performance by a factor of at least 2 in large --tree workloads by avoiding the repeated creation/destruction of the pool and containing threads.
Cycling pools also encountered lots lock contention, which accounted for most of the time saved by reusing a single pool.
This is all a big commit because it took a lot more work than I thought it would! The commit basically moves Git repositories from being per-directory to living for the whole life of the program. This allows for several directories in the same repository to be listed in the same invocation; before, it would try to rediscover the repository each time! This is why two of the tests “broke”: it suddenly started working with --recurse.
The Dir type does now not use Git at all; because a Dir doesn’t have a Git, then a File doesn’t have one either, so the Git cache gets passed to the render functions which will put them in the Table to render them.
This commit adds a cache for Git repositories based on the path being queried.
Its only immediate effect is that when you query the same directory twice (such as /testcases/git /testcases/git), it won’t need to check that the second one is a Git directory the second time. So, a minuscule optimisation for something you’d never do anyway? Wrong! It’s going to let us combine multiple entries over the same repository later, letting us use --tree and --recurse, because now Git scanning is behind a factory.
The error in #178 was being hidden from output most of the time, and because exa isn’t a GUI program, there’s nowhere it can really dump log output like this. Now that users can opt in with the EXA_DEBUG variable, there is a place it can go.
Also stop the ERROR log level being printed by default.
exa now ignores errors when checking for extended attributes when the user didn’t explicitly demand that they be checked. If a file does have xattrs, it’ll still display the @ in the permissions column; errors will now just cause the @ to be hidden instead.
This changed a lot of the xtests, which were displaying the error message in a few situations. Those tests have gained @-suffixed companions so the actual error messages can still be tested.
Fixes#178 (finally)
This commit moves the definitions of Filter and DirAction from the options module to the fs module, but leaves the parts that actually have to do with option parsing alone.
Now, the options module shouldn’t define any types that get used elsewhere in the program: it only adds functionality to types that already exist.
This commit replaces the “two normal cases” of showing a link’s target or not with “one default and one special case” of preferring to hide them, displaying the link targets by setting a flag instead.
Doing this simplifies the file name constructor, which gets to remove an argument.
The new FileStyles value will contain all the fields necessary to “style” a file’s name. Right now this is only the Classify field, but there can be more later. The benefit of this is that when we add more, we won’t need to update all the places where file names are displayed.
This commit moves the Environment field from the Table to its Options, and properly gets rid of the name ‘columns’ from the last commit.
Having it in the Options is important, because it means it can be generated from some command-line options. Also, it reduces the number of arguments that need to be passed to Table::new; there would have been 4 with the inclusion of the Environment, but by moving some of the code into the function, we can avoid this (and any further arguments).
The views have been renamed to be the Optionses of their module; now the options for the Table — Columns — has followed suit.
This works out, because the table module depended on everything in the columns module. It opens the door for other only-table-specific things to be included.
The casualty was that by making it non-Clone and non-PartialEq, a bunch of other #[derive]-d types had to have their derivions removed too.
The Environment struct only used the Default trait so it could have the same call for both Environment<UsersCache> and Environment<MockUsers>. There’s no reason to keep it around anymore.
There was a bug where if you tried to recurse into a directory you didn’t have permission to read the contents of, the error would be ignored.
It now displays the errors.
The goal of this part of the refactoring, if you wondered, is to make it so only the tree module is aware that it needs ‘depth’ and ‘last’ values to draw the tree.
As far as the details module is concerned, it should just be doing something to produce TreeParams values which it later consumes; that’s it.
This change should make it easier to have tables that may or may not have a tree in them.
