1
  2
  3
  4
  5
  6
  7
  8
  9
 10
 11
 12
 13
 14
 15
 16
 17
 18
 19
 20
 21
 22
 23
 24
 25
 26
 27
 28
 29
 30
 31
 32
 33
 34
 35
 36
 37
 38
 39
 40
 41
 42
 43
 44
 45
 46
 47
 48
 49
 50
 51
 52
 53
 54
 55
 56
 57
 58
 59
 60
 61
 62
 63
 64
 65
 66
 67
 68
 69
 70
 71
 72
 73
 74
 75
 76
 77
 78
 79
 80
 81
 82
 83
 84
 85
 86
 87
 88
 89
 90
 91
 92
 93
 94
 95
 96
 97
 98
 99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
//! This crate provides an implementation of
//! [elastic tabstops](http://nickgravgaard.com/elastictabstops/index.html).
//! It is a minimal port of Go's
//! [tabwriter](http://golang.org/pkg/text/tabwriter/) package.
//! Namely, its main mode of operation is to wrap a `Writer` and implement
//! elastic tabstops for the text written to the wrapped `Writer`.
//!
//! This package is also bundled with a program, `tabwriter`,
//! that exposes this functionality at the command line.
//!
//! Here's an example that shows basic alignment:
//!
//! ```rust
//! use std::io::Write;
//! use tabwriter::TabWriter;
//!
//! let mut tw = TabWriter::new(vec![]);
//! write!(&mut tw, "
//! Bruce Springsteen\tBorn to Run
//! Bob Seger\tNight Moves
//! Metallica\tBlack
//! The Boss\tDarkness on the Edge of Town
//! ").unwrap();
//! tw.flush().unwrap();
//!
//! let written = String::from_utf8(tw.into_inner().unwrap()).unwrap();
//! assert_eq!(&*written, "
//! Bruce Springsteen  Born to Run
//! Bob Seger          Night Moves
//! Metallica          Black
//! The Boss           Darkness on the Edge of Town
//! ");
//! ```
//!
//! Note that `flush` **must** be called or else `TabWriter` may never write
//! anything. This is because elastic tabstops requires knowing about future
//! lines in order to align output. More precisely, all text considered in a
//! single alignment must fit into memory.
//!
//! Here's another example that demonstrates how *only* contiguous columns
//! are aligned:
//!
//! ```rust
//! use std::io::Write;
//! use tabwriter::TabWriter;
//!
//! let mut tw = TabWriter::new(vec![]).padding(1);
//! write!(&mut tw, "
//!fn foobar() {{
//!    let mut x = 1+1;\t// addition
//!    x += 1;\t// increment in place
//!    let y = x * x * x * x;\t// multiply!
//!
//!    y += 1;\t// this is another group
//!    y += 2 * 2;\t// that is separately aligned
//!}}
//!").unwrap();
//! tw.flush().unwrap();
//!
//! let written = String::from_utf8(tw.into_inner().unwrap()).unwrap();
//! assert_eq!(&*written, "
//!fn foobar() {
//!    let mut x = 1+1;       // addition
//!    x += 1;                // increment in place
//!    let y = x * x * x * x; // multiply!
//!
//!    y += 1;     // this is another group
//!    y += 2 * 2; // that is separately aligned
//!}
//!");
//! ```

#![deny(missing_docs)]

#[cfg(feature = "ansi_formatting")]
#[macro_use]
extern crate lazy_static;
#[cfg(feature = "ansi_formatting")]
extern crate regex;
extern crate unicode_width;

use std::cmp;
use std::error;
use std::fmt;
use std::io::{self, Write};
use std::iter;
use std::mem;
use std::str;

#[cfg(feature = "ansi_formatting")]
use regex::Regex;
#[cfg(feature = "ansi_formatting")]
use std::borrow::Cow;

#[cfg(test)]
mod test;

/// TabWriter wraps an arbitrary writer and aligns tabbed output.
///
/// Elastic tabstops work by aligning *contiguous* tabbed delimited fields
/// known as *column blocks*. When a line appears that breaks all contiguous
/// blocks, all buffered output will be flushed to the underlying writer.
/// Otherwise, output will stay buffered until `flush` is explicitly called.
#[derive(Debug)]
pub struct TabWriter<W> {
    w: W,
    buf: io::Cursor<Vec<u8>>,
    lines: Vec<Vec<Cell>>,
    curcell: Cell,
    minwidth: usize,
    padding: usize,
}

#[derive(Debug)]
struct Cell {
    start: usize, // offset into TabWriter.buf
    width: usize, // in characters
    size: usize,  // in bytes
}

impl<W: io::Write> TabWriter<W> {
    /// Create a new `TabWriter` from an existing `Writer`.
    ///
    /// All output written to `Writer` is passed through `TabWriter`.
    /// Contiguous column blocks indicated by tabs are aligned.
    ///
    /// Note that `flush` must be called to guarantee that `TabWriter` will
    /// write to the given writer.
    pub fn new(w: W) -> TabWriter<W> {
        TabWriter {
            w: w,
            buf: io::Cursor::new(Vec::with_capacity(1024)),
            lines: vec!(vec!()),
            curcell: Cell::new(0),
            minwidth: 2,
            padding: 2,
        }
    }

    /// Set the minimum width of each column. That is, all columns will have
    /// *at least* the size given here. If a column is smaller than `minwidth`,
    /// then it is passed with spaces.
    ///
    /// The default minimum width is `2`.
    pub fn minwidth(mut self, minwidth: usize) -> TabWriter<W> {
        self.minwidth = minwidth;
        self
    }

    /// Set the padding between columns. All columns will be separated by
    /// *at least* the number of spaces indicated by `padding`. If `padding`
    /// is zero, then columns may run up against each other without any
    /// separation.
    ///
    /// The default padding is `2`.
    pub fn padding(mut self, padding: usize) -> TabWriter<W> {
        self.padding = padding;
        self
    }

    /// Unwraps this `TabWriter`, returning the underlying writer.
    ///
    /// This internal buffer is flushed before returning the writer. If the
    /// flush fails, then an error is returned.
    pub fn into_inner(mut self) -> Result<W, IntoInnerError<TabWriter<W>>> {
        match self.flush() {
            Ok(()) => Ok(self.w),
            Err(err) => Err(IntoInnerError(self, err)),
        }
    }

    /// Resets the state of the aligner. Once the aligner is reset, all future
    /// writes will start producing a new alignment.
    fn reset(&mut self) {
        self.buf = io::Cursor::new(Vec::with_capacity(1024));
        self.lines = vec!(vec!());
        self.curcell = Cell::new(0);
    }

    /// Adds the bytes received into the buffer and updates the size of
    /// the current cell.
    fn add_bytes(&mut self, bytes: &[u8]) {
        self.curcell.size += bytes.len();
        let _ = self.buf.write_all(bytes); // cannot fail
    }

    /// Ends the current cell, updates the UTF8 width of the cell and starts
    /// a fresh cell.
    fn term_curcell(&mut self) {
        let mut curcell = Cell::new(self.buf.position() as usize);
        mem::swap(&mut self.curcell, &mut curcell);

        curcell.update_width(&self.buf.get_ref());
        self.curline_mut().push(curcell);
    }

    /// Return a view of the current line of cells.
    fn curline(&mut self) -> &[Cell] {
        let i = self.lines.len() - 1;
        &*self.lines[i]
    }

    /// Return a mutable view of the current line of cells.
    fn curline_mut(&mut self) -> &mut Vec<Cell> {
        let i = self.lines.len() - 1;
        &mut self.lines[i]
    }
}

impl Cell {
    fn new(start: usize) -> Cell {
        Cell { start: start, width: 0, size: 0 }
    }

    fn update_width(&mut self, buf: &[u8]) {
        let end = self.start + self.size;
        self.width = display_columns(&buf[self.start..end]);
    }
}

impl<W: io::Write> io::Write for TabWriter<W> {
    fn write(&mut self, buf: &[u8]) -> io::Result<usize> {
        let mut lastterm = 0usize;
        for (i, &c) in buf.iter().enumerate() {
            match c {
                b'\t' | b'\n' => {
                    self.add_bytes(&buf[lastterm..i]);
                    self.term_curcell();
                    lastterm = i + 1;
                    if c == b'\n' {
                        let ncells = self.curline().len();
                        self.lines.push(vec!());
                        // Having a single cell means that *all* previous
                        // columns have been broken, so we should just flush.
                        if ncells == 1 {
                            try!(self.flush());
                        }
                    }
                }
                _ => {}
            }
        }
        self.add_bytes(&buf[lastterm..]);
        Ok(buf.len())
    }

    fn flush(&mut self) -> io::Result<()> {
        if self.curcell.size > 0 {
            self.term_curcell();
        }
        let widths = cell_widths(&self.lines, self.minwidth);

        // This is a trick to avoid allocating padding for every cell.
        // Just allocate the most we'll ever need and borrow from it.
        let biggest_width = widths.iter()
                                  .map(|ws| ws.iter().map(|&w|w).max()
                                              .unwrap_or(0))
                                  .max().unwrap_or(0);
        let padding: String =
            iter::repeat(' ').take(biggest_width + self.padding).collect();

        let mut first = true;
        for (line, widths) in self.lines.iter().zip(widths.iter()) {
            if !first { try!(self.w.write_all(b"\n")); } else { first = false }
            for (i, cell) in line.iter().enumerate() {
                let bytes = &self.buf.get_ref()[cell.start..cell.start + cell.size];
                try!(self.w.write_all(bytes));
                if i >= widths.len() {
                    assert_eq!(i, line.len()-1);
                } else {
                    assert!(widths[i] >= cell.width);
                    let padsize = self.padding + widths[i] - cell.width;
                    try!(write!(&mut self.w, "{}", &padding[0..padsize]));
                }
            }
        }

        self.reset();
        Ok(())
    }
}

/// An error returned by `into_inner`.
///
/// This combines the error that happened while flushing the buffer with the
/// `TabWriter` itself.
pub struct IntoInnerError<W>(W, io::Error);

impl<W> IntoInnerError<W> {
    /// Returns the error which caused the `into_error()` call to fail.
    pub fn error(&self) -> &io::Error {
        &self.1
    }

    /// Returns the `TabWriter` instance which generated the error.
    pub fn into_inner(self) -> W {
        self.0
    }
}

impl<W> fmt::Debug for IntoInnerError<W> {
    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
        self.error().fmt(f)
    }
}

impl<W> fmt::Display for IntoInnerError<W> {
    fn fmt(&self, f: &mut fmt::Formatter) -> fmt::Result {
        self.error().fmt(f)
    }
}

impl<W: ::std::any::Any> error::Error for IntoInnerError<W> {
    fn description(&self) -> &str {
        self.error().description()
    }

    fn cause(&self) -> Option<&error::Error> {
        Some(self.error())
    }
}

fn cell_widths(lines: &Vec<Vec<Cell>>, minwidth: usize) -> Vec<Vec<usize>> {
    // Naively, this algorithm looks like it could be O(n^2m) where `n` is
    // the number of lines and `m` is the number of contiguous columns.
    //
    // However, I claim that it is actually O(nm). That is, the width for
    // every contiguous column is computed exactly once.
    let mut ws: Vec<_> = (0..lines.len()).map(|_| vec![]).collect();
    for (i, iline) in lines.iter().enumerate() {
        if iline.is_empty() {
            continue
        }
        for col in ws[i].len()..(iline.len()-1) {
            let mut width = minwidth;
            let mut contig_count = 0;
            for line in lines[i..].iter() {
                if col + 1 >= line.len() { // ignores last column
                    break
                }
                contig_count += 1;
                width = cmp::max(width, line[col].width);
            }
            assert!(contig_count >= 1);
            for j in i..(i+contig_count) {
                ws[j].push(width);
            }
        }
    }
    ws
}

#[cfg(not(feature = "ansi_formatting"))]
fn display_columns(bytes: &[u8]) -> usize {
    use unicode_width::UnicodeWidthChar;

    // If we have a Unicode string, then attempt to guess the number of
    // *display* columns used.
    match str::from_utf8(bytes) {
        Err(_) => bytes.len(),
        Ok(s) => s.chars()
                  .map(|c| UnicodeWidthChar::width(c).unwrap_or(0))
                  .fold(0, |sum, width| sum + width),
    }
}

#[cfg(feature = "ansi_formatting")]
fn display_columns(bytes: &[u8]) -> usize {
    use unicode_width::UnicodeWidthChar;

    // If we have a Unicode string, then attempt to guess the number of
    // *display* columns used.
    match str::from_utf8(bytes) {
        Err(_) => bytes.len(),
        Ok(s) => strip_formatting(s).chars()
                    .map(|c| UnicodeWidthChar::width(c).unwrap_or(0))
                    .fold(0, |sum, width| sum + width),
    }
}

#[cfg(feature = "ansi_formatting")]
fn strip_formatting<'t>(input: &'t str) -> Cow<'t, str> {
    // use lazy_static to avoid compiling the regex every time
    // this function is called
    lazy_static! {
        static ref RE: regex::Regex = Regex::new(r#"\x1B\[.+?m"#).unwrap();
    }
    RE.replace_all(input, "")
}