1gitdiffcore(7) 2============== 3 4NAME 5---- 6gitdiffcore - Tweaking diff output 7 8SYNOPSIS 9-------- 10[verse] 11'git diff' * 12 13DESCRIPTION 14----------- 15 16The diff commands 'git diff-index', 'git diff-files', and 'git diff-tree' 17can be told to manipulate differences they find in 18unconventional ways before showing 'diff' output. The manipulation 19is collectively called "diffcore transformation". This short note 20describes what they are and how to use them to produce 'diff' output 21that is easier to understand than the conventional kind. 22 23 24The chain of operation 25---------------------- 26 27The 'git diff-{asterisk}' family works by first comparing two sets of 28files: 29 30 - 'git diff-index' compares contents of a "tree" object and the 31 working directory (when `--cached` flag is not used) or a 32 "tree" object and the index file (when `--cached` flag is 33 used); 34 35 - 'git diff-files' compares contents of the index file and the 36 working directory; 37 38 - 'git diff-tree' compares contents of two "tree" objects; 39 40In all of these cases, the commands themselves first optionally limit 41the two sets of files by any pathspecs given on their command-lines, 42and compare corresponding paths in the two resulting sets of files. 43 44The pathspecs are used to limit the world diff operates in. They remove 45the filepairs outside the specified sets of pathnames. E.g. If the 46input set of filepairs included: 47 48------------------------------------------------ 49:100644 100644 bcd1234... 0123456... M junkfile 50------------------------------------------------ 51 52but the command invocation was `git diff-files myfile`, then the 53junkfile entry would be removed from the list because only "myfile" 54is under consideration. 55 56The result of comparison is passed from these commands to what is 57internally called "diffcore", in a format similar to what is output 58when the -p option is not used. E.g. 59 60------------------------------------------------ 61in-place edit :100644 100644 bcd1234... 0123456... M file0 62create :000000 100644 0000000... 1234567... A file4 63delete :100644 000000 1234567... 0000000... D file5 64unmerged :000000 000000 0000000... 0000000... U file6 65------------------------------------------------ 66 67The diffcore mechanism is fed a list of such comparison results 68(each of which is called "filepair", although at this point each 69of them talks about a single file), and transforms such a list 70into another list. There are currently 5 such transformations: 71 72- diffcore-break 73- diffcore-rename 74- diffcore-merge-broken 75- diffcore-pickaxe 76- diffcore-order 77 78These are applied in sequence. The set of filepairs 'git diff-{asterisk}' 79commands find are used as the input to diffcore-break, and 80the output from diffcore-break is used as the input to the 81next transformation. The final result is then passed to the 82output routine and generates either diff-raw format (see Output 83format sections of the manual for 'git diff-{asterisk}' commands) or 84diff-patch format. 85 86 87diffcore-break: For Splitting Up Complete Rewrites 88-------------------------------------------------- 89 90The second transformation in the chain is diffcore-break, and is 91controlled by the -B option to the 'git diff-{asterisk}' commands. This is 92used to detect a filepair that represents "complete rewrite" and 93break such filepair into two filepairs that represent delete and 94create. E.g. If the input contained this filepair: 95 96------------------------------------------------ 97:100644 100644 bcd1234... 0123456... M file0 98------------------------------------------------ 99 100and if it detects that the file "file0" is completely rewritten, 101it changes it to: 102 103------------------------------------------------ 104:100644 000000 bcd1234... 0000000... D file0 105:000000 100644 0000000... 0123456... A file0 106------------------------------------------------ 107 108For the purpose of breaking a filepair, diffcore-break examines 109the extent of changes between the contents of the files before 110and after modification (i.e. the contents that have "bcd1234..." 111and "0123456..." as their SHA-1 content ID, in the above 112example). The amount of deletion of original contents and 113insertion of new material are added together, and if it exceeds 114the "break score", the filepair is broken into two. The break 115score defaults to 50% of the size of the smaller of the original 116and the result (i.e. if the edit shrinks the file, the size of 117the result is used; if the edit lengthens the file, the size of 118the original is used), and can be customized by giving a number 119after "-B" option (e.g. "-B75" to tell it to use 75%). 120 121 122diffcore-rename: For Detecting Renames and Copies 123------------------------------------------------- 124 125This transformation is used to detect renames and copies, and is 126controlled by the -M option (to detect renames) and the -C option 127(to detect copies as well) to the 'git diff-{asterisk}' commands. If the 128input contained these filepairs: 129 130------------------------------------------------ 131:100644 000000 0123456... 0000000... D fileX 132:000000 100644 0000000... 0123456... A file0 133------------------------------------------------ 134 135and the contents of the deleted file fileX is similar enough to 136the contents of the created file file0, then rename detection 137merges these filepairs and creates: 138 139------------------------------------------------ 140:100644 100644 0123456... 0123456... R100 fileX file0 141------------------------------------------------ 142 143When the "-C" option is used, the original contents of modified files, 144and deleted files (and also unmodified files, if the 145"--find-copies-harder" option is used) are considered as candidates 146of the source files in rename/copy operation. If the input were like 147these filepairs, that talk about a modified file fileY and a newly 148created file file0: 149 150------------------------------------------------ 151:100644 100644 0123456... 1234567... M fileY 152:000000 100644 0000000... bcd3456... A file0 153------------------------------------------------ 154 155the original contents of fileY and the resulting contents of 156file0 are compared, and if they are similar enough, they are 157changed to: 158 159------------------------------------------------ 160:100644 100644 0123456... 1234567... M fileY 161:100644 100644 0123456... bcd3456... C100 fileY file0 162------------------------------------------------ 163 164In both rename and copy detection, the same "extent of changes" 165algorithm used in diffcore-break is used to determine if two 166files are "similar enough", and can be customized to use 167a similarity score different from the default of 50% by giving a 168number after the "-M" or "-C" option (e.g. "-M8" to tell it to use 1698/10 = 80%). 170 171Note. When the "-C" option is used with `--find-copies-harder` 172option, 'git diff-{asterisk}' commands feed unmodified filepairs to 173diffcore mechanism as well as modified ones. This lets the copy 174detector consider unmodified files as copy source candidates at 175the expense of making it slower. Without `--find-copies-harder`, 176'git diff-{asterisk}' commands can detect copies only if the file that was 177copied happened to have been modified in the same changeset. 178 179 180diffcore-merge-broken: For Putting Complete Rewrites Back Together 181------------------------------------------------------------------ 182 183This transformation is used to merge filepairs broken by 184diffcore-break, and not transformed into rename/copy by 185diffcore-rename, back into a single modification. This always 186runs when diffcore-break is used. 187 188For the purpose of merging broken filepairs back, it uses a 189different "extent of changes" computation from the ones used by 190diffcore-break and diffcore-rename. It counts only the deletion 191from the original, and does not count insertion. If you removed 192only 10 lines from a 100-line document, even if you added 910 193new lines to make a new 1000-line document, you did not do a 194complete rewrite. diffcore-break breaks such a case in order to 195help diffcore-rename to consider such filepairs as candidate of 196rename/copy detection, but if filepairs broken that way were not 197matched with other filepairs to create rename/copy, then this 198transformation merges them back into the original 199"modification". 200 201The "extent of changes" parameter can be tweaked from the 202default 80% (that is, unless more than 80% of the original 203material is deleted, the broken pairs are merged back into a 204single modification) by giving a second number to -B option, 205like these: 206 207* -B50/60 (give 50% "break score" to diffcore-break, use 60% 208 for diffcore-merge-broken). 209 210* -B/60 (the same as above, since diffcore-break defaults to 50%). 211 212Note that earlier implementation left a broken pair as a separate 213creation and deletion patches. This was an unnecessary hack and 214the latest implementation always merges all the broken pairs 215back into modifications, but the resulting patch output is 216formatted differently for easier review in case of such 217a complete rewrite by showing the entire contents of old version 218prefixed with '-', followed by the entire contents of new 219version prefixed with '+'. 220 221 222diffcore-pickaxe: For Detecting Addition/Deletion of Specified String 223--------------------------------------------------------------------- 224 225This transformation limits the set of filepairs to those that change 226specified strings between the preimage and the postimage in a certain 227way. -S<block of text> and -G<regular expression> options are used to 228specify different ways these strings are sought. 229 230"-S<block of text>" detects filepairs whose preimage and postimage 231have different number of occurrences of the specified block of text. 232By definition, it will not detect in-file moves. Also, when a 233changeset moves a file wholesale without affecting the interesting 234string, diffcore-rename kicks in as usual, and `-S` omits the filepair 235(since the number of occurrences of that string didn't change in that 236rename-detected filepair). When used with `--pickaxe-regex`, treat 237the <block of text> as an extended POSIX regular expression to match, 238instead of a literal string. 239 240"-G<regular expression>" (mnemonic: grep) detects filepairs whose 241textual diff has an added or a deleted line that matches the given 242regular expression. This means that it will detect in-file (or what 243rename-detection considers the same file) moves, which is noise. The 244implementation runs diff twice and greps, and this can be quite 245expensive. 246 247When `-S` or `-G` are used without `--pickaxe-all`, only filepairs 248that match their respective criterion are kept in the output. When 249`--pickaxe-all` is used, if even one filepair matches their respective 250criterion in a changeset, the entire changeset is kept. This behavior 251is designed to make reviewing changes in the context of the whole 252changeset easier. 253 254diffcore-order: For Sorting the Output Based on Filenames 255--------------------------------------------------------- 256 257This is used to reorder the filepairs according to the user's 258(or project's) taste, and is controlled by the -O option to the 259'git diff-{asterisk}' commands. 260 261This takes a text file each of whose lines is a shell glob 262pattern. Filepairs that match a glob pattern on an earlier line 263in the file are output before ones that match a later line, and 264filepairs that do not match any glob pattern are output last. 265 266As an example, a typical orderfile for the core Git probably 267would look like this: 268 269------------------------------------------------ 270README 271Makefile 272Documentation 273*.h 274*.c 275t 276------------------------------------------------ 277 278SEE ALSO 279-------- 280linkgit:git-diff[1], 281linkgit:git-diff-files[1], 282linkgit:git-diff-index[1], 283linkgit:git-diff-tree[1], 284linkgit:git-format-patch[1], 285linkgit:git-log[1], 286linkgit:gitglossary[7], 287link:user-manual.html[The Git User's Manual] 288 289GIT 290--- 291Part of the linkgit:git[1] suite