[FILE] Docs update

From: David Greaves <david@dgreaves.com>
Date: Apr 24, 2005, 19:46 UTC
Message-ID: <426BF790.9070406@dgreaves.com>

Now contains strawman reference documentation for all the core git commands

It would probably be a good idea for people who know what's going on to validate these docs lest the partially sighted author unwittingly leads too many potential gitters over conceptual cliffs...

And I've attached this as a file rather than a patch to make it easier for people to read.

Signed-off-by: David Greaves <david@dgreaves.com>
---

This file contains reference information for the core git commands. It is actually based on the source from Petr Baudis' tree and may therefore contain a few 'extras' that may or may not make it upstream.

The README contains much useful definition and clarification info - read that first. And of the commands, I suggest reading 'update-cache' and 'read-tree' first - I wish I had!

David Greaves - 24/4/05

Identifier terminology used:

<object>
	Indicates any object sha1 identifier

<blob>
	Indicates a blob object sha1 identifier

<tree>
	Indicates a tree object sha1 identifier

<commit>
	Indicates a commit object sha1 identifier

<tree/commit>
	Indicates a tree or commit object sha1 identifier (usually
	because the command can read the <tree> a <commit> contains).
	[Eventually may be replaced with <tree> if <tree> means
	<tree/commit> in all commands]

<type>
	Indicates that an object type is required.
	Currently one of: blob/tree/commit

<file>
	Indicates a filename - often includes leading path

<path>
	Indicates the path of a file (is this ever useful?)

################################################################
cat-file
	cat-file (-t | <type>) <object>

Provide contents or type of objects in the repository. The type is required if -t is not being used to find the object type.

<object>
	The sha1 identifier of the object.

-t
	show the object type identified by <object>

<type>
	One of: blob/tree/commit

Output

If -t is specified, one of:
        blob/tree/commit

Otherwise the raw (though uncompressed) contents of the <object> will be returned.

################################################################
check-files
	check-files <file>...

Check that a list of files are up-to-date between the filesystem and the cache. Used to verify a patch target before doing a patch.

Files that do not exist on the filesystem are considered up-to-date (whether or not they are in the cache).

Emits an error message on failure.
preparing to update existing file <file> not in cache
	  <file> exists but is not in the cache

preparing to update file <file> not uptodate in cache
	  <file> on disk is not up-to-date with the cache

exits with a status code indicating success if all files are up-to-date.

see also: update-cache

################################################################
checkout-cache
	checkout-cache [-q] [-a] [-f] [--prefix=<string>] [--] <file>...

Will copy all files listed from the cache to the working directory (not overwriting existing files). Note that the file contents are restored - NOT the file permissions.

-q
	be quiet if files exist or are not in the cache

-f
	forces overwrite of existing files

-a
	checks out all files in the cache (will then continue to
	process listed files).

--prefix=<string>
	When creating files, prepend <string> (usually a directory
	including a trailing /)

--
	Do not interpret any more arguments as options.

Note that the order of the flags matters:

	checkout-cache -a -f file.c

will first check out all files listed in the cache (but not overwrite any old ones), and then force-checkout file.c a second time (ie that one _will_ overwrite any old contents with the same filename).

Also, just doing "checkout-cache" does nothing. You probably meant "checkout-cache -a". And if you want to force it, you want "checkout-cache -f -a".

Intuitiveness is not the goal here. Repeatability is. The reason for the "no arguments means no work" thing is that from scripts you are supposed to be able to do things like

	find . -name '*.h' -print0 | xargs -0 checkout-cache -f --

which will force all existing *.h files to be replaced with their cached copies. If an empty command line implied "all", then this would force-refresh everything in the cache, which was not the point.

Oh, and the "--" is just a good idea when you know the rest will be filenames. Just so that you wouldn't have a filename of "-a" causing problems (not possible in the above example, but get used to it in scripting!).

The prefix ability basically makes it trivial to use checkout-cache as
a "export as tree" function. Just read the desired tree into the
index, and do a
  
        checkout-cache --prefix=export-dir/ -a
  
and checkout-cache will "export" the cache into the specified
directory.
  
NOTE! The final "/" is important. The exported name is literally just
prefixed with the specified string, so you can also do something like
  
        checkout-cache --prefix=.merged- Makefile
  
to check out the currently cached copy of "Makefile" into the file
".merged-Makefile".

################################################################
commit-tree
	commit-tree <tree> [-p <parent tree>]*   < changelog

Creates a new commit object based on the provided tree object and emits the new commit object id on stdout. If no parent is given then it is considered to be an initial tree.

A commit object usually has 1 parent (a commit after a change) or up to 16 parents. More than one parent represents merge of branches that led to them.

While a tree represents a particular directory state of a working directory, a commit represents that state in "time", and explains how to get there.

Normally a commit would identify a new "HEAD" state, and while git doesn't care where you save the note about that state, in practice we tend to just write the result to the file ".git/HEAD", so that we can always see what the last committed state was.

Options

<tree>
	An existing tree object

-p <parent tree>
	Each -p indicates a the id of a parent commit object.

Commit Information

A commit encapsulates:
	all parent object ids
	author name, email and date
	committer name and email and the commit time.

If not provided, commit-tree uses your name, hostname and domain to
provide author and committer info. This can be overridden using the
following environment variables.
	AUTHOR_NAME
	AUTHOR_EMAIL
	AUTHOR_DATE
	COMMIT_AUTHOR_NAME
	COMMIT_AUTHOR_EMAIL
(nb <,> and '\n's are stripped)

A commit comment is read from stdin (max 999 chars). If a changelog entry is not provided via '<' redirection, commit-tree will just wait for one to be entered and terminated with ^D

see also: write-tree

################################################################
diff-cache
	diff-cache [-r] [-z] [--cached] <tree/commit>

Compares the content and mode of the blobs found via a tree object with the content of the current cache and, optionally ignoring the stat state of the file on disk.

(This is basically a special case of diff-tree that works with the current cache as the first tree.)

<tree/commit>
	The id of a tree or commit object to diff against.

-r
	recurse

-z
	\0 line termination on output

--cached
	do not consider the on-disk file at all

Output format:

For files in the tree but not in the cache -<mode>\t <type>\t <object>\t <path><file>

For files in the cache but not in the tree +<mode>\t <type>\t <object>\t <path><file>

For files that differ: *<tree-mode>-><cache-mode>\t <type>\t <tree-sha1>-><cache-sha1>\t <path><file>

In the special case of the file being changed on disk and out of sync with the cache, the sha1 is all 0's. Example:

	*100644->100660 blob    5be4a414b32cf4204f889469942986d3d783da84->0000000000000000000000000000000000000000      file.c

Operating Modes

You can choose whether you want to trust the index file entirely (using the "--cached" flag) or ask the diff logic to show any files that don't match the stat state as being "tentatively changed". Both of these operations are very useful indeed.

Cached Mode

If --cached is specified, it allows you to ask:
	show me the differences between HEAD and the current index
	contents (the ones I'd write with a "write-tree")

For example, let's say that you have worked on your index file, and are ready to commit. You want to see eactly _what_ you are going to commit is without having to write a new tree object and compare it that way, and to do that, you just do

	diff-cache --cached $(cat .git/HEAD)

Example: let's say I had renamed "commit.c" to "git-commit.c", and I had 
done an "upate-cache" to make that effective in the index file. 
"show-diff" wouldn't show anything at all, since the index file matches 
my working directory. But doing a diff-cache does:
	torvalds@ppc970:~/git> diff-cache --cached $(cat .git/HEAD)
	-100644 blob    4161aecc6700a2eb579e842af0b7f22b98443f74        commit.c
	+100644 blob    4161aecc6700a2eb579e842af0b7f22b98443f74        git-commit.c

And as you can see, the output matches "diff-tree -r" output (we always do "-r", since the index is always fully populated ??CHECK??). You can trivially see that the above is a rename.

In fact, "diff-cache --cached" _should_ always be entirely equivalent to actually doing a "write-tree" and comparing that. Except this one is much nicer for the case where you just want to check where you are.

So doing a "diff-cache --cached" is basically very useful when you are asking yourself "what have I already marked for being committed, and what's the difference to a previous tree".

Non-cached Mode

The "non-cached" mode takes a different approach, and is potentially the even more useful of the two in that what it does can't be emulated with a "write-tree + diff-tree". Thus that's the default mode. The non-cached version asks the question

   "show me the differences between HEAD and the currently checked out 
    tree - index contents _and_ files that aren't up-to-date"

which is obviously a very useful question too, since that tells you what you _could_ commit. Again, the output matches the "diff-tree -r" output to a tee, but with a twist.

The twist is that if some file doesn't match the cache, we don't have a backing store thing for it, and we use the magic "all-zero" sha1 to show that. So let's say that you have edited "kernel/sched.c", but have not actually done an update-cache on it yet - there is no "object" associated with the new state, and you get:

	torvalds@ppc970:~/v2.6/linux> diff-cache $(cat .git/HEAD )
	*100644->100664 blob    7476bbcfe5ef5a1dd87d745f298b831143e4d77e->0000000000000000000000000000000000000000      kernel/sched.c

ie it shows that the tree has changed, and that "kernel/sched.c" has is not up-to-date and may contain new stuff. The all-zero sha1 means that to get the real diff, you need to look at the object in the working directory directly rather than do an object-to-object diff.

NOTE! As with other commands of this type, "diff-cache" does not actually look at the contents of the file at all. So maybe "kernel/sched.c" hasn't actually changed, and it's just that you touched it. In either case, it's a note that you need to upate-cache it to make the cache be in sync.

NOTE 2! You can have a mixture of files show up as "has been updated" and "is still dirty in the working directory" together. You can always tell which file is in which state, since the "has been updated" ones show a valid sha1, and the "not in sync with the index" ones will always have the special all-zero sha1.

################################################################
diff-tree
	diff-tree [-r] [-z] <tree/commit> <tree/commit>

Compares the content and mode of the blobs found via two tree objects.

Note that diff-tree can use the tree encapsulated in a commit object.

<tree sha1>
	The id of a tree or commit object.

-r
	recurse

-z
	\0 line termination on output

Output format:

For files in tree1 but not in tree2 -<mode>\t <type>\t <object>\t <path><file>

For files not in tree1 but in tree2 +<mode>\t <type>\t <object>\t <path><file>

For files that differ: *<tree1-mode>-><tree2-mode>\t <type>\t <tree1 sha1>-><tree2 sha1>\t <path><file>

An example of normal usage is:

	torvalds@ppc970:~/git> diff-tree 5319e4d609cdd282069cc4dce33c1db559539b03 b4e628ea30d5ab3606119d2ea5caeab141d38df7
	*100664->100664 blob    ac348b7d5278e9d04e3a1cd417389379c32b014f->a01513ed4d4d565911a60981bfb4173311ba3688      fsck-cache.c

which tells you that the last commit changed just one file (it's from this one:

	commit 3c6f7ca19ad4043e9e72fa94106f352897e651a8
	tree 5319e4d609cdd282069cc4dce33c1db559539b03
	parent b4e628ea30d5ab3606119d2ea5caeab141d38df7
	author Linus Torvalds <torvalds@ppc970.osdl.org> Sat Apr 9 12:02:30 2005
	committer Linus Torvalds <torvalds@ppc970.osdl.org> Sat Apr 9 12:02:30 2005

	Make "fsck-cache" print out all the root commits it finds.

	Once I do the reference tracking, I'll also make it print out all the
	HEAD commits it finds, which is even more interesting.

in case you care).

################################################################
fsck-cache
	fsck-cache [[--unreachable] <commit>*]

Verifies the connectivity and validity of the objects in the database.

<commit>
	A commit object to treat as the head of an unreachability
	trace

--unreachable
	print out objects that exist but that aren't readable from any
	of the specified root nodes

It tests SHA1 and general object sanity, but it does full tracking of the resulting reachability and everything else. It prints out any corruption it finds (missing or bad objects), and if you use the "--unreachable" flag it will also print out objects that exist but that aren't readable from any of the specified root nodes.

So for example

	fsck-cache --unreachable $(cat .git/HEAD)

or, for Cogito users:

	fsck-cache --unreachable $(cat .git/heads/*)

will do quite a _lot_ of verification on the tree. There are a few extra validity tests to be added (make sure that tree objects are sorted properly etc), but on the whole if "fsck-cache" is happy, you do have a valid tree.

Any corrupt objects you will have to find in backups or other archives (ie you can just remove them and do an "rsync" with some other site in the hopes that somebody else has the object you have corrupted).

Of course, "valid tree" doesn't mean that it wasn't generated by some evil person, and the end result might be crap. Git is a revision tracking system, not a quality assurance system ;)

Extracted Diagnostics

expect dangling commits - potential heads - due to lack of head information
	You haven't specified any nodes as heads so it won't be
	possible to differentiate between un-parented commits and
	root nodes.

missing sha1 directory '<dir>'
	The directory holding the sha1 objects is missing.

unreachable <type> <object>
	The <type> object <object>, isn't actually referred to directly
	or indirectly in any of the trees or commits seen. This can
	mean that there's another root na SHA1_ode that you're not specifying
	or that the tree is corrupt. If you haven't missed a root node
	then you might as well delete unreachable nodes since they
	can't be used.

missing <type> <object>
	The <type> object <object>, is referred to but isn't present in
	the database.

dangling <type> <object>
	The <type> object <object>, is present in the database but never
	_directly_ used. A dangling commit could be a root node.

warning: fsck-cache: tree <tree> has full pathnames in it
	And it shouldn't...

sha1 mismatch <object>
	The database has an object who's sha1 doesn't match the
	database value.
	This indicates a ??serious?? data integrity problem.
	(note: this error occured during early git development when
	the database format changed.)

Environment Variables

SHA1_FILE_DIRECTORY
	used to specify the object database root (usually .git/objects)

################################################################
git-export
	git-export top [base]

probably deprecated: On Wed, 20 Apr 2005, Petr Baudis wrote:

Show 5 quoted lines

>> I will probably not buy git-export, though. (That is, it is merged, but
>> I won't make git frontend for it.) My "git export" already does
>> something different, but more importantly, "git patch" of mine already
>> does effectively the same thing as you do, just for a single patch; so I
>> will probably just extend it to do it for an (a,b] range of patches.

That's fine. It was a quick hack, just to show that if somebody wants to, the data is trivially exportable.

		Linus

Although in Linus' distribution show-diff is not part of 'core' git.

################################################################
init-db
	init-db

This simply creates an empty git object database - basically a .git directory.

If the object storage directory is specified via the SHA1_FILE_DIRECTORY environment variable then the sha1 directories are created underneath - otherwise the default .git/objects directory is used.

init-db won't hurt an existing repository.

################################################################
ls-tree
	ls-tree [-r] [-z] <tree/commit>

convert the tree object to a human readable (and script processable) form.

<tree/commit>
	Id of a tree or commit object.
-r
	recurse into sub-trees

-z
	\0 line termination on output

Output Format <mode>\t <type>\t <object>\t <path><file>

################################################################
merge-base
	merge-base <commit> <commit>

merge-base finds one of the best common ancestors of a pair of commits. In particular, it finds one of the ones which is fewest commits away from the further of the heads.

When it has a tree with equally likely common ancestors like this:

  A -- C
    \/   \
    /\   /
  B -- D

It essentially returns either A or B randomly.

The merge-base algorithm is still in flux.

################################################################
merge-cache
	merge-cache <merge-program> (-a | -- | <file>*)

This looks up the <file>(s) in the cache and, if there are any merge entries, unpacks all of them (which may be just one file, of course) into up to three separate temporary files, and then executes the supplied <merge-program> with those three files as arguments 1,2,3 (empty argument if no file), and <file> as argument 4.

--
	Interpret all future arguments as filenames

-a
	Run merge against all files in the cache that need merging.

If merge-cache is called with multiple <file>s (or -a) then it processes them in turn only stopping if merge returns a non-zero exit code.

Typically this is run with the a script calling the merge command from the RCS package.

Example script:

	#!/bin/sh
	#
	# This is the git merge script, called with
	#
	#   $1 - original file (or empty string)
	#   $2 - file in branch1 (or empty string)
	#   $3 - file in branch2 (or empty string)
	#   $4 - pathname in repository
	#
	#
	# Case 1: file removed in both
	#
	if [ -z "$2" && -z "$3" ]; then
		rm -- "$4"
		update-cache --remove -- "$4"
		exit 0
	fi
	#
	# Case 2: file exists in just one
	#
	if [ -z "$2" || -z "$3" ] then
		cat "$2""$3" > "$4"
		update-cache --add -- "$4"
		exit 0
	fi
	#
	# Case 3: file exists in both
	#
	src="$1"
	if [ -z "$1" ]; then
		src=/den/null
	fi	
	merge "$3" "$src" "$2" && cp "$3" "$4" && update-cache --add -- "$4"

ALERT ALERT ALERT! The git "merge object order" is different from the RCS "merge" program merge object order. In the above ordering, the original is first. But the argument order to the 3-way merge program "merge" is to have the original in the middle. Don't ask me why.

Examples:

	torvalds@ppc970:~/merge-test> merge-cache cat MM
	This is MM from the original tree.			# original
	This is modified MM in the branch A.			# merge1
	This is modified MM in the branch B.			# merge2
	This is modified MM in the branch B.			# current contents

or

	torvalds@ppc970:~/merge-test> merge-cache cat AA MM
	cat: : No such file or directory
	This is added AA in the branch A.
	This is added AA in the branch B.
	This is added AA in the branch B.
	fatal: merge program failed

where the latter example shows how "merge-cache" will stop trying to merge once anything has returned an error (ie "cat" returned an error for the AA file, because it didn't exist in the original, and thus "merge-cache" didn't even try to merge the MM thing).

################################################################
read-tree
	read-tree (<tree> | -m <tree1> [<tree2> <tree3>])"

Reads the tree information given by <tree> into the directory cache, but does not actually _update_ any of the files it "caches". (see: checkout-cache)

Optionally, it can merge a tree into the cache or perform a 3-way merge.

Trivial merges are done by read-tree itself. Only conflicting paths will be in unmerged state when read-tree returns.

-m
	Perform a merge, not just a read

<tree#>
	The id of the tree object(s) to be read/merged.

Merging If -m is specified, read-tree performs 2 kinds of merge, a single tree merge if only 1 tree is given or a 3-way merge if 3 trees are provided.

Single Tree Merge If only 1 tree is specified, read-tree operates as if the user did not specify "-m", except that if the original cache has an entry for a given pathname; and the contents of the path matches with the tree being read, the stat info from the cache is used. (In other words, the cache's stat()s take precedence over the merged tree's)

That means that if you do a "read-tree -m <newtree>" followed by a "checkout-cache -f -a", the checkout-cache only checks out the stuff that really changed.

This is used to avoid unnecessary false hits when show-diff is run after read-tree.

3-Way Merge Each "index" entry has two bits worth of "stage" state. stage 0 is the normal one, and is the only one you'd see in any kind of normal use.

However, when you do "read-tree" with multiple trees, the "stage" starts out at 0, but increments for each tree you read. And in particular, the "-m" flag means "start at stage 1" instead.

This means that you can do

	read-tree -m <tree1> <tree2> <tree3>

and you will end up with an index with all of the <tree1> entries in "stage1", all of the <tree2> entries in "stage2" and all of the <tree3> entries in "stage3".

Furthermore, "read-tree" has special-case logic that says: if you see a file that matches in all respects in the following states, it "collapses" back to "stage0":

   - stage 2 and 3 are the same; take one or the other (it makes no
     difference - the same work has been done on stage 2 and 3)

   - stage 1 and stage 2 are the same and stage 3 is different; take
     stage 3 (some work has been done on stage 3)

   - stage 1 and stage 3 are the same and stage 2 is different take
     stage 2 (some work has been done on stage 2)

Write-tree refuses to write a nonsensical tree, so write-tree will complain about unmerged entries if it sees a single entry that is not stage 0".

Ok, this all sounds like a collection of totally nonsensical rules, but it's actually exactly what you want in order to do a fast merge. The different stages represent the "result tree" (stage 0, aka "merged"), the original tree (stage 1, aka "orig"), and the two trees you are trying to merge (stage 2 and 3 respectively).

In fact, the way "read-tree" works, it's entirely agnostic about how you assign the stages, and you could really assign them any which way, and the above is just a suggested way to do it (except since "write-tree" refuses to write anything but stage0 entries, it makes sense to always consider stage 0 to be the "full merge" state).

So what happens? Try it out. Select the original tree, and two trees to merge, and look how it works:

 - if a file exists in identical format in all three trees, it will 
   automatically collapse to "merged" state by the new read-tree.

 - a file that has _any_ difference what-so-ever in the three trees
   will stay as separate entries in the index. It's up to "script
   policy" to determine how to remove the non-0 stages, and insert a
   merged version.  But since the index is always sorted, they're easy
   to find: they'll be clustered together.

 - the index file saves and restores with all this information, so you
   can merge things incrementally, but as long as it has entries in
   stages 1/2/3 (ie "unmerged entries") you can't write the result.

So now the merge algorithm ends up being really simple:

 - you walk the index in order, and ignore all entries of stage 0,
   since they've already been done.

 - if you find a "stage1", but no matching "stage2" or "stage3", you
   know it's been removed from both trees (it only existed in the
   original tree), and you remove that entry.  - if you find a
   matching "stage2" and "stage3" tree, you remove one of them, and
   turn the other into a "stage0" entry. Remove any matching "stage1"
   entry if it exists too.  .. all the normal trivial rules ..

Incidentally - it also means that you don't even have to have a separate subdirectory for this. All the information literally is in the index file, which is a temporary thing anyway. There is no need to worry about what is in the working directory, since it is never shown and never used.