Merge branch 'develop'

.gitignore

@@ -5,3 +5,5 @@ target/
 result
 *.html
 *.log
+*.svg
+perf.data*

README.md

@@ -9,9 +9,17 @@ https://mit-license.org/)
 
 A LaTeX formatter written in Rust.
 
-## Formatting style example
-
-Before formatting `example.tex`:
+<table width="100%">
+<tr>
+<td>
+<b>Input</b>
+</td>
+<td>
+<b>Output</b>
+</td>
+</tr>
+<tr>
+<td>
 
 ``` tex
 \documentclass{article}
@@ -29,8 +37,8 @@ E = m c^2
 
 \end{document}
 ```
-
-After running `tex-fmt example.tex`
+</td>
+<td>
 
 ``` tex
 \documentclass{article}
@@ -48,6 +56,9 @@ After running `tex-fmt example.tex`
 
 \end{document}
 ```
+</td>
+</tr>
+</table>
 
 ## Installation
 

notes.org

@@ -1,32 +1,31 @@
 #+title: tex-fmt
 * Tests
-** One feature per file
-** Benchmark running all tests
-*** Give number of files and total line numbers
-** Look for problem cases in other documents
+** Look for problem cases in other latex documents
 * Features
-** Commands on own lines
-*** Implement this by regex on whole document at start
-*** Care about multiple or long arguments
-*** Look into bracket matching
-*** \begin and \end
-*** \section, \section*, \chapter, \subsection etc
-*** \author and \title etc
-*** \usepackage and \documentclass
-*** \label
-*** \centering
-*** \includegraphics
-*** \caption
-*** \newpage
-*** \appendix
-** Fold long lines
+** TODO Environments should start and end on new lines
+*** New line before begin/end with regex replace
+*** Care with comments
+** Fold long lines to 80 characters
+*** Care with trailing comments
+*** No folding in verbatim environments, just warn
+** Merge short alphanumeric lines
 ** Flags
-*** Dry run (do not modify files)
-*** Print to STDOUT
-*** Debug (display log file)
+*** -d Dry run
+*** -p Print to STDOUT
+*** -v Info verbose
+*** -vv Debug verbose
 ** Log file
+** Ignore source lines
+*** Line-by-line ignore
+*** Block ignore
 * Bugs
-** Better errors including line numbers in source file
 ** Check multiple begins or ends on the same line
+*** This should never happen once new lines are implemented
+** Better errors including line numbers in source file
 ** Check begins or ends with more brackets on the same line
 ** Handle list items properly
+* Structure
+** Perform indenting
+** While long lines are present
+*** Trim long lines
+*** Perform indenting

shell.nix

@@ -4,5 +4,6 @@ pkgs.mkShell {
   buildInputs = with pkgs; [
     rustfmt
     clippy
+    cargo-flamegraph
   ];
 }

src/format.rs

@@ -0,0 +1,53 @@
+use crate::indent::*;
+use crate::subs::*;
+use crate::TAB;
+
+pub fn format_file(file: String, debug: bool) -> String {
+    // preformat
+    let mut new_file = remove_extra_newlines(&file);
+    //new_file = begin_end_environments_new_line(&new_file);
+    new_file = remove_tabs(&new_file);
+    new_file = remove_trailing_spaces(&new_file);
+    let lines: Vec<&str> = new_file.lines().collect();
+
+    // set up variables
+    let n_lines = lines.len();
+    let mut indent = Indent {
+        actual: 0,
+        visual: 0,
+    };
+    let mut new_lines = vec!["".to_owned(); n_lines];
+
+    // main loop through file
+    for i in 0..n_lines {
+        // calculate indent
+        let line = lines[i];
+        let line_strip = &remove_comment(line);
+        indent = get_indent(line_strip, indent);
+        if !debug {
+            dbg!(&line);
+            assert!(indent.actual >= 0, "line {}", i);
+            assert!(indent.visual >= 0, "line {}", i);
+        };
+
+        // apply indent
+        let mut new_line = line.trim_start().to_string();
+        if !new_line.is_empty() {
+            let n_spaces = indent.visual * TAB;
+            let spaces: String = (0..n_spaces).map(|_| " ").collect();
+            new_line.insert_str(0, &spaces);
+        }
+        new_lines[i] = new_line
+    }
+
+    // check indents return to zero
+    if !debug {
+        assert!(indent.actual == 0);
+        assert!(indent.visual == 0);
+    }
+
+    // prepare indented file
+    let mut new_file = new_lines.join("\n");
+    new_file.push('\n');
+    new_file
+}

src/indent.rs

@@ -0,0 +1,94 @@
+use crate::regexes::*;
+use core::cmp::max;
+
+const OPENS: [char; 3] = ['(', '[', '{'];
+const CLOSES: [char; 3] = [')', ']', '}'];
+
+#[derive(Debug)]
+pub struct Indent {
+    /// actual running indentation count at end of current line
+    pub actual: i8,
+    /// visual indentation of current line
+    pub visual: i8,
+}
+
+/// calculate total indentation change due to current line
+pub fn get_diff(line: &str) -> i8 {
+    // documents get no global indentation
+    if RE_DOCUMENT_BEGIN.is_match(line) || RE_DOCUMENT_END.is_match(line) {
+        return 0;
+    };
+
+    // list environments get double indents
+    let mut diff: i8 = 0;
+    for re_list_begin in RE_LISTS_BEGIN.iter() {
+        if re_list_begin.is_match(line) {
+            diff += 1
+        };
+    }
+    for re_list_end in RE_LISTS_END.iter() {
+        if re_list_end.is_match(line) {
+            diff -= 1
+        };
+    }
+
+    // other environments get single indents
+    if RE_ENV_BEGIN.is_match(line) {
+        diff += 1
+    };
+    if RE_ENV_END.is_match(line) {
+        diff -= 1
+    };
+
+    // indent for delimiters
+    for c in OPENS {
+        diff += line.chars().filter(|&x| x == c).count() as i8;
+    }
+    for c in CLOSES {
+        diff -= line.chars().filter(|&x| x == c).count() as i8;
+    }
+
+    diff
+}
+
+/// calculate dedentation for current line compared to previous
+pub fn get_back(line: &str) -> i8 {
+    // documents get no global indentation
+    if RE_DOCUMENT_END.is_match(line) {
+        return 0;
+    };
+
+    // list environments get double indents for indenting items
+    for re_list_end in RE_LISTS_END.iter() {
+        if re_list_end.is_match(line) {
+            return 2;
+        };
+    }
+
+    // other environments get single indents
+    if RE_ENV_END.is_match(line) {
+        return 1;
+    };
+
+    // deindent items to make the rest of item environment appear indented
+    if RE_ITEM.is_match(line) {
+        return 1;
+    };
+
+    let mut back: i8 = 0;
+    let mut cumul: i8 = 0;
+    for c in line.chars() {
+        cumul -= OPENS.contains(&c) as i8;
+        cumul += CLOSES.contains(&c) as i8;
+        back = max(cumul, back);
+    }
+    back
+}
+
+pub fn get_indent(line: &str, prev_indent: Indent) -> Indent {
+    let diff = get_diff(line);
+    let back = get_back(line);
+    let actual = prev_indent.actual + diff;
+    let visual: i8 = prev_indent.actual - back;
+    Indent { actual, visual }
+}