Squashed commit for alternative format processing

Squashed commit of the following: commit ccbc433 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Tue Oct 15 14:13:35 2024 +0200 Fix typo commit 0e4ec8c Merge: 52fe1f7 3f658fd Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Tue Oct 15 14:07:14 2024 +0200 Merge remote-tracking branch 'upstream/main' into pr-refactor-logic commit 52fe1f7 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Mon Oct 14 22:19:16 2024 +0200 Document added regex commit 1418812 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Mon Oct 14 18:15:44 2024 +0200 Fix Clippy warnings commit 6c248db Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Mon Oct 14 18:04:11 2024 +0200 Use `char::width()` for checking wrapping need commit 3f658fd Author: William G Underwood <42812654+WGUNDERWOOD@users.noreply.github.com> Date: Mon Oct 14 15:28:00 2024 +0100 Clippy commit 0dcbc93 Author: William G Underwood <42812654+WGUNDERWOOD@users.noreply.github.com> Date: Mon Oct 14 15:03:15 2024 +0100 Split cli and read files commit 3ca0fd2 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Mon Oct 14 14:09:18 2024 +0200 Remove unused function commit b4e2130 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Mon Oct 14 14:08:45 2024 +0200 Wrap lines based on indent length and only apply indentation after wrapping commit 9bc0a3b Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Sun Oct 13 20:12:44 2024 +0200 Move allocation outside of wrap application commit 76d9ef0 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Sat Oct 12 11:17:43 2024 +0200 Separate indent calculation from application commit eb7971e Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Sat Oct 12 10:46:56 2024 +0200 Extract indent calculation into function commit e41a66d Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Sat Oct 12 10:31:58 2024 +0200 Apply clippy suggestions commit 21bb908 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Fri Oct 11 17:43:14 2024 +0200 Rename ignore check and document commit ac164ad Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Fri Oct 11 17:13:31 2024 +0200 Simplify Option matching commit 8a6e868 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Fri Oct 11 12:58:12 2024 +0200 Remove unnecessary comment check and `Option` in return type when splitting lines commit 1d15b86 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Fri Oct 11 12:57:17 2024 +0200 Change comments around commit e77b629 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Fri Oct 11 12:56:46 2024 +0200 Refactor ignoring check as function commit b10b010 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Fri Oct 11 11:37:18 2024 +0200 Move environment line splitting before indenting This avoids re-adding the start of the split line to the queue. commit 0424d0a Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Fri Oct 11 11:23:01 2024 +0200 Move check for ignore outside of `apply_indent()` commit 70ac614 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Fri Oct 11 09:29:55 2024 +0200 Use new return type for `put_env_new_line` commit fdea16b Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Fri Oct 11 09:28:54 2024 +0200 Fix infinite loop issue when split points are in comments commit 791c7f1 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Fri Oct 11 09:27:23 2024 +0200 Correct to not expect `{` after `\item` but only after `\begin` and `\end` commit 6e5ce6c Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Fri Oct 11 09:26:06 2024 +0200 Match target tex file to source file commit 9562af1 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Thu Oct 10 22:27:53 2024 +0200 Remove debug print statement commit 0f48547 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Thu Oct 10 18:10:00 2024 +0200 Split environment and item lines without allocating This does not compile because the `format` function cannot yet deal with the new return type. commit c49a08c Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Thu Oct 10 13:27:48 2024 +0200 Check indentation return to 0 on `state` instead of `new_text` commit 1b90f21 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Thu Oct 10 13:11:57 2024 +0200 Split condition tree into distinct blocks with `continue` commit e974225 Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Thu Oct 10 11:07:08 2024 +0200 Plan line formatting logic steps commit 4bf802a Author: Cyprien de Saint Guilhem <cyprien.desaintguilhem@3milabs.tech> Date: Thu Oct 10 10:53:42 2024 +0200 Group lines to clean text into one function commit 380380d Author: William G Underwood <42812654+WGUNDERWOOD@users.noreply.github.com> Date: Thu Oct 10 14:33:04 2024 +0100 Release process notes
WGUNDERWOOD · Oct 15, 2024 · db634ec · db634ec
1 parent 945637b
commit db634ec
Show file tree

Hide file tree

Showing 16 changed files with 358 additions and 230 deletions.
diff --git a/Cargo.lock b/Cargo.lock
diff --git a/Cargo.toml b/Cargo.toml
@@ -18,6 +18,7 @@ lazy_static = "1.5.0"
 log = "0.4.22"
 regex = "1.11.0"
 similar = "2.6.0"
+unicode-width = "0.2.0"
 
 [profile.release]
 codegen-units = 1

diff --git a/notes.org b/notes.org
@@ -1 +1,11 @@
 #+title: tex-fmt
+* Release process
+** Update version number in Cargo.toml
+** Push to GitHub and check tests pass
+** Create a git tag
+*** git tag vX.X.X
+*** git push --tags
+** Publish to crates.io with cargo publish
+** Publish GitHub release with notes
+*** GitHub binaries published automatically with actions
+** Publish in nixpkgs when bot makes pull request
diff --git a/src/parse.rs → src/cli.rs b/src/parse.rs → src/cli.rs
@@ -1,12 +1,9 @@
 //! Utilities for reading the command line arguments
 
 use crate::logging::*;
-use crate::regexes::*;
 use clap::Parser;
-use log::Level::{Error, Trace};
+use log::Level::Error;
 use log::LevelFilter;
-use std::fs;
-use std::io::Read;
 
 /// Command line arguments
 #[allow(missing_docs)]
@@ -112,48 +109,3 @@ impl Cli {
         }
     }
 }
-
-/// Add a missing extension and read the file
-pub fn read(file: &str, logs: &mut Vec<Log>) -> Option<(String, String)> {
-    // check if file has an accepted extension
-    let has_ext = EXTENSIONS.iter().any(|e| file.ends_with(e));
-    // if no valid extension, try adding .tex
-    let mut new_file = file.to_owned();
-    if !has_ext {
-        new_file.push_str(".tex");
-    };
-    if let Ok(text) = fs::read_to_string(&new_file) {
-        return Some((new_file, text));
-    }
-    if has_ext {
-        record_file_log(logs, Error, file, "Could not open file.");
-    } else {
-        record_file_log(logs, Error, file, "File type invalid.");
-    }
-    None
-}
-
-/// Attempt to read from STDIN, return filename `<STDIN>` and text
-pub fn read_stdin(logs: &mut Vec<Log>) -> Option<(String, String)> {
-    let mut text = String::new();
-    match std::io::stdin().read_to_string(&mut text) {
-        Ok(bytes) => {
-            record_file_log(
-                logs,
-                Trace,
-                "<STDIN>",
-                &format!("Read {bytes} bytes."),
-            );
-            Some((String::from("<STDIN>"), text))
-        }
-        Err(e) => {
-            record_file_log(
-                logs,
-                Error,
-                "<STDIN>",
-                &format!("Could not read from STDIN: {e}"),
-            );
-            None
-        }
-    }
-}
diff --git a/src/comments.rs b/src/comments.rs
@@ -16,8 +16,3 @@ pub fn find_comment_index(line: &str) -> Option<usize> {
 pub fn remove_comment(line: &str, comment: Option<usize>) -> &str {
     comment.map_or_else(|| line, |c| &line[0..c])
 }
-
-/// Extract a comment from the end of a line
-pub fn get_comment(line: &str, comment: Option<usize>) -> &str {
-    comment.map_or_else(|| "", |c| &line[c..])
-}
diff --git a/src/format.rs b/src/format.rs
@@ -1,9 +1,9 @@
 //! Core methodology for formatting a file
 
+use crate::cli::*;
 use crate::ignore::*;
 use crate::indent::*;
 use crate::logging::*;
-use crate::parse::*;
 use crate::regexes::{ENV_BEGIN, ENV_END, ITEM};
 use crate::subs::*;
 use crate::verbatim::*;
@@ -14,77 +14,106 @@ use std::iter::zip;
 
 /// Central function to format a file
 pub fn format_file(
-    text: &str,
+    old_text: &str,
     file: &str,
     args: &Cli,
     logs: &mut Vec<Log>,
 ) -> String {
     record_file_log(logs, Info, file, "Formatting started.");
-    let mut old_text = remove_extra_newlines(text);
-    if !args.usetabs {
-        old_text = remove_tabs(&old_text, args);
-    }
-    old_text = remove_trailing_spaces(&old_text);
 
+    // Clean the source file and zip its lines with line numbers
+    let old_text = clean_text(old_text, args);
+    let mut old_lines = zip(1.., old_text.lines());
+
+    // Initialise
     let mut state = State::new();
-    let old_lines = old_text.lines();
-    let mut old_lines = zip(1.., old_lines);
     let mut queue: Vec<(usize, String)> = vec![];
-    let mut new_text = String::with_capacity(2 * text.len());
+    let mut new_text = String::with_capacity(2 * old_text.len());
+
+    // Select the character used for indentation.
     let indent_char = if args.usetabs { "\t" } else { " " };
 
     loop {
         if let Some((linum_old, mut line)) = queue.pop() {
+            // Read the patterns present on this line.
             let pattern = Pattern::new(&line);
-            let temp_state: State;
-            (line, temp_state) = apply_indent(
+
+            // Temporary state for working on this line.
+            let mut temp_state = state.clone();
+
+            // Update the state with the line number from the queue.
+            temp_state.linum_old = linum_old;
+
+            // If the line should not be ignored ...
+            if !set_ignore_and_report(
                 &line,
-                linum_old,
-                &state,
+                &mut temp_state,
                 logs,
                 file,
-                args,
                 &pattern,
-                indent_char,
-            );
-            if needs_env_new_line(&line, &temp_state, &pattern) {
-                let env_lines =
-                    put_env_new_line(&line, &temp_state, file, args, logs);
-                if env_lines.is_some() {
-                    queue.push((linum_old, env_lines.clone().unwrap().1));
-                    queue.push((linum_old, env_lines.clone().unwrap().0));
-                } else {
-                    state = temp_state;
-                    new_text.push_str(&line);
-                    new_text.push_str(LINE_END);
-                    state.linum_new += 1;
-                };
-            } else if needs_wrap(&line, &temp_state, args) {
-                let wrapped_lines =
-                    apply_wrap(&line, &temp_state, file, args, logs);
-                if wrapped_lines.is_some() {
-                    queue.push((linum_old, wrapped_lines.clone().unwrap().1));
-                    queue.push((linum_old, wrapped_lines.clone().unwrap().0));
-                } else {
-                    state = temp_state;
-                    new_text.push_str(&line);
-                    new_text.push_str(LINE_END);
-                    state.linum_new += 1;
-                };
-            } else {
-                state = temp_state;
-                new_text.push_str(&line);
-                new_text.push_str(LINE_END);
-                state.linum_new += 1;
+            ) {
+                // Check if the line should be split because of a pattern that should begin on a new line.
+                if needs_env_new_line(&line, &temp_state, &pattern) {
+                    // Split the line into two ...
+                    let (this_line, next_line) =
+                        put_env_new_line(&line, &temp_state, file, args, logs);
+                    // ... and queue the second part for formatting.
+                    queue.push((linum_old, next_line.to_string()));
+                    line = this_line.to_string();
+                }
+
+                // Calculate the indent based on the current state and the patterns in the line.
+                let indent = calculate_indent(
+                    &line,
+                    &mut temp_state,
+                    logs,
+                    file,
+                    args,
+                    &pattern,
+                );
+
+                let indent_length = usize::try_from(indent.visual * args.tab)
+                    .expect("Visual indent is non-negative.");
+
+                // Wrap the line before applying the indent, and loop back if the line needed wrapping.
+                if needs_wrap(line.trim_start(), indent_length, args) {
+                    let wrapped_lines = apply_wrap(
+                        line.trim_start(),
+                        indent_length,
+                        &temp_state,
+                        file,
+                        args,
+                        logs,
+                    );
+                    if let Some([this_line, next_line_start, next_line]) =
+                        wrapped_lines
+                    {
+                        queue.push((
+                            linum_old,
+                            [next_line_start, next_line].concat(),
+                        ));
+                        queue.push((linum_old, this_line.to_string()));
+                        continue;
+                    }
+                }
+
+                // Lastly, apply the indent if the line didn't need wrapping.
+                line = apply_indent(&line, &indent, args, indent_char);
             }
+
+            // Add line to new text
+            state = temp_state;
+            new_text.push_str(&line);
+            new_text.push_str(LINE_END);
+            state.linum_new += 1;
         } else if let Some((linum_old, line)) = old_lines.next() {
             queue.push((linum_old, line.to_string()));
         } else {
             break;
         }
     }
 
-    if !indents_return_to_zero(&new_text) {
+    if !indents_return_to_zero(&state) {
         record_file_log(logs, Warn, file, "Indent does not return to zero.");
     }
 
@@ -93,6 +122,35 @@ pub fn format_file(
     new_text
 }
 
+/// Sets the `ignore` and `verbatim` flags in the given [State] based on `line` and returns whether `line` should be
+/// ignored by formatting.
+fn set_ignore_and_report(
+    line: &str,
+    temp_state: &mut State,
+    logs: &mut Vec<Log>,
+    file: &str,
+    pattern: &Pattern,
+) -> bool {
+    temp_state.ignore = get_ignore(line, temp_state, logs, file, true);
+    temp_state.verbatim =
+        get_verbatim(line, temp_state, logs, file, true, pattern);
+
+    temp_state.verbatim.visual || temp_state.ignore.visual
+}
+
+/// Cleans the given text by removing extra line breaks and trailing spaces, and tabs if they shouldn't be used.
+fn clean_text(text: &str, args: &Cli) -> String {
+    let mut text = remove_extra_newlines(text);
+
+    if !args.usetabs {
+        text = remove_tabs(&text, args);
+    }
+
+    text = remove_trailing_spaces(&text);
+
+    text
+}
+
 /// Information on the current state during formatting
 #[derive(Clone, Debug)]
 pub struct State {
@@ -143,6 +201,7 @@ impl Pattern {
 }
 
 /// Ensure that the indentation returns to zero at the end of the file
-fn indents_return_to_zero(text: &str) -> bool {
-    !text.lines().last().unwrap_or_default().starts_with(' ')
+fn indents_return_to_zero(state: &State) -> bool {
+    #![allow(clippy::missing_const_for_fn)]
+    state.indent.actual == 0
 }