| 1 | diff-highlight |
| 2 | ============== |
| 3 | |
| 4 | Line oriented diffs are great for reviewing code, because for most |
| 5 | hunks, you want to see the old and the new segments of code next to each |
| 6 | other. Sometimes, though, when an old line and a new line are very |
| 7 | similar, it's hard to immediately see the difference. |
| 8 | |
| 9 | You can use "--color-words" to highlight only the changed portions of |
| 10 | lines. However, this can often be hard to read for code, as it loses |
| 11 | the line structure, and you end up with oddly formatted bits. |
| 12 | |
| 13 | Instead, this script post-processes the line-oriented diff, finds pairs |
| 14 | of lines, and highlights the differing segments. It's currently very |
| 15 | simple and stupid about doing these tasks. In particular: |
| 16 | |
| 17 | 1. It will only highlight hunks in which the number of removed and |
| 18 | added lines is the same, and it will pair lines within the hunk by |
| 19 | position (so the first removed line is compared to the first added |
| 20 | line, and so forth). This is simple and tends to work well in |
| 21 | practice. More complex changes don't highlight well, so we tend to |
| 22 | exclude them due to the "same number of removed and added lines" |
| 23 | restriction. Or even if we do try to highlight them, they end up |
| 24 | not highlighting because of our "don't highlight if the whole line |
| 25 | would be highlighted" rule. |
| 26 | |
| 27 | 2. It will find the common prefix and suffix of two lines, and |
| 28 | consider everything in the middle to be "different". It could |
| 29 | instead do a real diff of the characters between the two lines and |
| 30 | find common subsequences. However, the point of the highlight is to |
| 31 | call attention to a certain area. Even if some small subset of the |
| 32 | highlighted area actually didn't change, that's OK. In practice it |
| 33 | ends up being more readable to just have a single blob on the line |
| 34 | showing the interesting bit. |
| 35 | |
| 36 | The goal of the script is therefore not to be exact about highlighting |
| 37 | changes, but to call attention to areas of interest without being |
| 38 | visually distracting. Non-diff lines and existing diff coloration is |
| 39 | preserved; the intent is that the output should look exactly the same as |
| 40 | the input, except for the occasional highlight. |
| 41 | |
| 42 | Build/Install |
| 43 | ------------- |
| 44 | |
| 45 | You can build the `diff-highlight` script by running `make` from within |
| 46 | the diff-highlight directory. There is no `make install` target; you can |
| 47 | copy the built script to your $PATH. |
| 48 | |
| 49 | You can run diff-highlight's internal tests by running `make test`. Note |
| 50 | that you must also build Git itself first (by running `make` from the |
| 51 | top-level of the project). |
| 52 | |
| 53 | Use |
| 54 | --- |
| 55 | |
| 56 | You can try out the built diff-highlight program with: |
| 57 | |
| 58 | --------------------------------------------- |
| 59 | git log -p --color | /path/to/diff-highlight |
| 60 | --------------------------------------------- |
| 61 | |
| 62 | If you want to use it all the time, drop it in your $PATH and put the |
| 63 | following in your git configuration: |
| 64 | |
| 65 | --------------------------------------------- |
| 66 | [pager] |
| 67 | log = diff-highlight | less |
| 68 | show = diff-highlight | less |
| 69 | diff = diff-highlight | less |
| 70 | --------------------------------------------- |
| 71 | |
| 72 | If you use the interactive patch mode of `git add -p`, `git checkout |
| 73 | -p`, etc, you may also want to configure it to be used there: |
| 74 | |
| 75 | --------------------------------------------- |
| 76 | [interactive] |
| 77 | diffFilter = diff-highlight |
| 78 | --------------------------------------------- |
| 79 | |
| 80 | |
| 81 | Color Config |
| 82 | ------------ |
| 83 | |
| 84 | You can configure the highlight colors and attributes using git's |
| 85 | config. The colors for "old" and "new" lines can be specified |
| 86 | independently. There are two "modes" of configuration: |
| 87 | |
| 88 | 1. You can specify a "highlight" color and a matching "reset" color. |
| 89 | This will retain any existing colors in the diff, and apply the |
| 90 | "highlight" and "reset" colors before and after the highlighted |
| 91 | portion. |
| 92 | |
| 93 | 2. You can specify a "normal" color and a "highlight" color. In this |
| 94 | case, existing colors are dropped from that line. The non-highlighted |
| 95 | bits of the line get the "normal" color, and the highlights get the |
| 96 | "highlight" color. |
| 97 | |
| 98 | If no "new" colors are specified, they default to the "old" colors. If |
| 99 | no "old" colors are specified, the default is to reverse the foreground |
| 100 | and background for highlighted portions. |
| 101 | |
| 102 | Examples: |
| 103 | |
| 104 | --------------------------------------------- |
| 105 | # Underline highlighted portions |
| 106 | [color "diff-highlight"] |
| 107 | oldHighlight = ul |
| 108 | oldReset = noul |
| 109 | --------------------------------------------- |
| 110 | |
| 111 | --------------------------------------------- |
| 112 | # Varying background intensities |
| 113 | [color "diff-highlight"] |
| 114 | oldNormal = "black #f8cbcb" |
| 115 | oldHighlight = "black #ffaaaa" |
| 116 | newNormal = "black #cbeecb" |
| 117 | newHighlight = "black #aaffaa" |
| 118 | --------------------------------------------- |
| 119 | |
| 120 | |
| 121 | Using diff-highlight as a module |
| 122 | -------------------------------- |
| 123 | |
| 124 | If you want to pre- or post- process the highlighted lines as part of |
| 125 | another perl script, you can use the DiffHighlight module. You can |
| 126 | either "require" it or just cat the module together with your script (to |
| 127 | avoid run-time dependencies). |
| 128 | |
| 129 | Your script may set up one or more of the following variables: |
| 130 | |
| 131 | - $DiffHighlight::line_cb - this should point to a function which is |
| 132 | called whenever DiffHighlight has lines (which may contain |
| 133 | highlights) to output. The default function prints each line to |
| 134 | stdout. Note that the function may be called with multiple lines. |
| 135 | |
| 136 | - $DiffHighlight::flush_cb - this should point to a function which |
| 137 | flushes the output (because DiffHighlight believes it has completed |
| 138 | processing a logical chunk of input). The default function flushes |
| 139 | stdout. |
| 140 | |
| 141 | - @DiffHighlight::OLD_HIGHLIGHT and @DiffHighlight::NEW_HIGHLIGHT - these |
| 142 | arrays specify the normal, highlighted, and reset colors (in that order) |
| 143 | for old/new lines. If unset, values will be retrieved by calling `git |
| 144 | config` (see "Color Config" above). Note that these should be the literal |
| 145 | color bytes (starting with an ANSI escape code), not color names. |
| 146 | |
| 147 | The script may then feed lines, one at a time, to DiffHighlight::handle_line(). |
| 148 | When lines are done processing, they will be fed to $line_cb. Note that |
| 149 | DiffHighlight may queue up many input lines (to analyze a whole hunk) |
| 150 | before calling $line_cb. After providing all lines, call |
| 151 | DiffHighlight::flush() to flush any unprocessed lines. |
| 152 | |
| 153 | If you just want to process stdin, DiffHighlight::highlight_stdin() |
| 154 | is a convenience helper which will loop and flush for you. |
| 155 | |
| 156 | |
| 157 | Bugs |
| 158 | ---- |
| 159 | |
| 160 | Because diff-highlight relies on heuristics to guess which parts of |
| 161 | changes are important, there are some cases where the highlighting is |
| 162 | more distracting than useful. Fortunately, these cases are rare in |
| 163 | practice, and when they do occur, the worst case is simply a little |
| 164 | extra highlighting. This section documents some cases known to be |
| 165 | sub-optimal, in case somebody feels like working on improving the |
| 166 | heuristics. |
| 167 | |
| 168 | 1. Two changes on the same line get highlighted in a blob. For example, |
| 169 | highlighting: |
| 170 | |
| 171 | ---------------------------------------------- |
| 172 | -foo(buf, size); |
| 173 | +foo(obj->buf, obj->size); |
| 174 | ---------------------------------------------- |
| 175 | |
| 176 | yields (where the inside of "+{}" would be highlighted): |
| 177 | |
| 178 | ---------------------------------------------- |
| 179 | -foo(buf, size); |
| 180 | +foo(+{obj->buf, obj->}size); |
| 181 | ---------------------------------------------- |
| 182 | |
| 183 | whereas a more semantically meaningful output would be: |
| 184 | |
| 185 | ---------------------------------------------- |
| 186 | -foo(buf, size); |
| 187 | +foo(+{obj->}buf, +{obj->}size); |
| 188 | ---------------------------------------------- |
| 189 | |
| 190 | Note that doing this right would probably involve a set of |
| 191 | content-specific boundary patterns, similar to word-diff. Otherwise |
| 192 | you get junk like: |
| 193 | |
| 194 | ----------------------------------------------------- |
| 195 | -this line has some -{i}nt-{ere}sti-{ng} text on it |
| 196 | +this line has some +{fa}nt+{a}sti+{c} text on it |
| 197 | ----------------------------------------------------- |
| 198 | |
| 199 | which is less readable than the current output. |
| 200 | |
| 201 | 2. The multi-line matching assumes that lines in the pre- and post-image |
| 202 | match by position. This is often the case, but can be fooled when a |
| 203 | line is removed from the top and a new one added at the bottom (or |
| 204 | vice versa). Unless the lines in the middle are also changed, diffs |
| 205 | will show this as two hunks, and it will not get highlighted at all |
| 206 | (which is good). But if the lines in the middle are changed, the |
| 207 | highlighting can be misleading. Here's a pathological case: |
| 208 | |
| 209 | ----------------------------------------------------- |
| 210 | -one |
| 211 | -two |
| 212 | -three |
| 213 | -four |
| 214 | +two 2 |
| 215 | +three 3 |
| 216 | +four 4 |
| 217 | +five 5 |
| 218 | ----------------------------------------------------- |
| 219 | |
| 220 | which gets highlighted as: |
| 221 | |
| 222 | ----------------------------------------------------- |
| 223 | -one |
| 224 | -t-{wo} |
| 225 | -three |
| 226 | -f-{our} |
| 227 | +two 2 |
| 228 | +t+{hree 3} |
| 229 | +four 4 |
| 230 | +f+{ive 5} |
| 231 | ----------------------------------------------------- |
| 232 | |
| 233 | because it matches "two" to "three 3", and so forth. It would be |
| 234 | nicer as: |
| 235 | |
| 236 | ----------------------------------------------------- |
| 237 | -one |
| 238 | -two |
| 239 | -three |
| 240 | -four |
| 241 | +two +{2} |
| 242 | +three +{3} |
| 243 | +four +{4} |
| 244 | +five 5 |
| 245 | ----------------------------------------------------- |
| 246 | |
| 247 | which would probably involve pre-matching the lines into pairs |
| 248 | according to some heuristic. |