How does "index f2e4113..d4b9bfc 100644" in git diff correspond to SHA1 ID in gitk? -

what 'index f2e4113..d4b9bfc 100644' mean in output git diff? till have assumed these sha1 id's from_version..to_version, can't find them in gitk.

$ git --version git version 1.8.3.4 $ git diff ae1fdc1 diff --git a/readme b/readme index f2e4113..d4b9bfc 100644 --- a/readme +++ b/readme @@ -1 +1,3 @@ stuff +more +more stuff 

'index f2e4113..d4b9bfc 100644' doesn't part of diff unified format. http://en.wikipedia.org/wiki/diff#unified_format

100644 looks file mode, doesn't correspond mode of readme file (660).

f2e4113..d4b9bfc sha1 id shorties. 'git rev-parse ' gives long sha1. here 3 shorties example above:

ae1fdc1 - ae1fdc1e7b2f018a15c421f2f22b7c77215c5d61 d4b9bfc - d4b9bfcd51a3eaf427d337a30b12d1f3dbdd21b4 f2e4113 - f2e41136eac73c39554dede1fd7e67b12502d577 

but why doesn't ae1fdc1 correspond first or second part of diff? , why can't find d4b9bfc or d4b9bfc using gitk?

edit: @wiswit has pointed this explanation.

f2e4113 , d4b9bfc file-indexes. git show f2e4113 see file before commit , git show d4b9bfc shows file after commit.

the combined diff format in git diff manual pages described.

combined diff format    diff-generating command can take ‘-c` or --cc option produce     combined diff when showing merge. default format when showing     merges git-diff(1) or git-show(1). note can give `-m’    option of these commands force generation of diffs individual     parents of merge.     combined diff format looks this:         diff --combined describe.c        index fabadb8,cc95eb0..4866510        --- a/describe.c        +++ b/describe.c        @@@ -98,20 -98,12 +98,20 @@@                return (a_date > b_date) ? -1 : (a_date == b_date) ? 0 : 1;          }         - static void describe(char *arg)         -static void describe(struct commit *cmit, int last_one)        ++static void describe(char *arg, int last_one)          {         +      unsigned char sha1[20];         +      struct commit *cmit;                struct commit_list *list;                static int initialized = 0;                struct commit_name *n;          +      if (get_sha1(arg, sha1) < 0)         +              usage(describe_usage);         +      cmit = lookup_commit_reference(sha1);         +      if (!cmit)         +              usage(describe_usage);         +                if (!initialized) {                        initialized = 1;                        for_each_ref(get_name);      1. preceded "git diff" header, looks (when -c         option used):             diff --combined file         or (when --cc option used):             diff --cc file      2. followed 1 or more extended header lines (this example shows        merge 2 parents):             index <hash>,<hash>..<hash>            mode <mode>,<mode>..<mode>            new file mode <mode>            deleted file mode <mode>,<mode>         mode <mode>,<mode>..<mode> line appears if @ least 1 of         <mode> different rest. extended headers information          detected contents movement (renames , copying detection)        designed work diff of 2 <tree-ish> , not used combined        diff format.      3. followed two-line from-file/to-file header             --- a/file            +++ b/file         similar two-line header traditional unified diff format,         /dev/null used signal created or deleted files.      4. chunk header format modified prevent people accidentally        feeding patch -p1. combined diff format created review         of merge commit changes, , not meant apply. change         similar change in extended index header:             @@@ <from-file-range> <from-file-range> <to-file-range> @@@         there (number of parents + 1) @ characters in chunk header         combined diff format.     unlike traditional unified diff format, shows 2 files , b    single column has - (minus — appears in removed in b),    + (plus — missing in added b), or " " (space — unchanged) prefix,    format compares 2 or more files file1, file2,... 1 file x,     , shows how x differs each of filen. 1 column each of filen    prepended output line note how x’s line different it.     - character in column n means line appears in filen    not appear in result. + character in column n means    line appears in result, , filen not have line (in     other words, line added, point of view of parent).     in above example output, function signature changed both    files (hence 2 - removals both file1 , file2, plus ++ mean 1     line added not appear in either file1 nor file2). 8     other lines same file1 not appear in file2 (hence     prefixed +).     when shown git diff-tree -c, compares parents of merge commit    merge result (i.e. file1..filen parents). when shown     git diff-files -c, compares 2 unresolved merge parents     working tree file (i.e. file1 stage 2 aka "our version", file2 stage     3 aka "their version").