The add_extra_ref() interface is used to add an extra-ref that is _not_
our ref for the purpose of helping auto-following of tags and reducing
object transfer from remote repository, and they are typically formatted
as a tagname followed by ^{} to make sure no valid refs match that
pattern. In other words, these entries are deliberately formatted not to
pass check-refname-format test.

A recent series however added a test unconditionally to the add_ref()
function that is called from add_extra_ref(). The check may be sensible
for other two callsites of the add_ref() interface, but definitely is
a wrong thing to do in add_extra_ref(). Disable it.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>
Acked-by: NMichael Haggerty <mhagger@alum.mit.edu>

"git log frotz" can DWIM to "refs/remotes/frotz/HEAD", but in the remote
access context, "git fetch frotz" to fetch what the other side happened to
have fetched from what it calls 'frotz' (which may not have any relation
to what we consider is 'frotz') the last time would not make much sense,
so the fetch rules table did not include "refs/remotes/%.*s/HEAD".

When the user really wants to, "git fetch $there remotes/frotz/HEAD" would
let her do so anyway, so this is not about safety or security; it merely
is about confusion avoidance and discouraging meaningless usage.

Specifically, it is _not_ about ambiguity avoidance. A name that would
become ambiguous if we use the same rules table for both fetch and local
rev-parse would be ambiguous locally at the remote side.

So for the same reason as we added rule to allow "git fetch $there v1.0"
instead of "git fetch $there tags/v1.0" in the previous commit, here is a
bit longer rope for the users, which incidentally simplifies our code.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

You can already do so with "git fetch $there tags/v1.0" but if it is not
ambiguous there is no reason to force users to type more.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

629cd3ac (resolve_ref(): emit warnings for improperly-formatted references,
2011-09-15) made resolve_ref() warn against files that are found in the
directories the ref dwimmery looks at. The intent may be good, but these
messages come from a wrong level of the API hierarchy.

Instead record the breakage in "flags" whose purpose is to explain the
result of the function to the caller, who is in a much better position to
make intelligent decision based on the information.

This updates sha1_name.c::dwim_ref() to warn against such a broken
candidate only when it does not appear directly below $GIT_DIR to restore
the traditional behaviour, as we know many files directly underneath
$GIT_DIR/ are not refs.

Warning against "git show config --" with "$GIT_DIR/config does not look
like a well-formed ref" does not make sense, and we may later tweak the
dwimmery not to even consider them as candidates, but that is a longer
term topic.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Instead of keeping this as an internal API, let the callers find
out the reason why resolve_ref() returned NULL is not because there
was no such file in $GIT_DIR but because a file was corrupt.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

2c5c66be (Merge branch 'jp/get-ref-dir-unsorted', 2011-10-10) merged a
topic that forked from the mainline before a new helper function
get_packed_refs() refactored code to read packed-refs file. The merge made
the call to the helper function with an incorrect argument. The parameter
to the function has to be a path to the submodule.

Fix the mismerge.
Helped-by: NMark Levedahl <mlevedahl@gmail.com>
Helped-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

clear_ref_cache() was only called from one place, so inline it
there.
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Since write_ref_sha1() can only write loose refs and cannot write
symbolic refs, there is no need for it to invalidate the packed ref
cache.

Suggested by: Martin Fick <mfick@codeaurora.org>
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Extract two new functions from clear_cached_refs():
clear_loose_ref_cache() and clear_packed_ref_cache().
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

...for consistency with the rest of this module.
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Make invalidate_ref_cache() an official part of the refs API.  It is
currently a fact of life that code outside of refs.c mucks about with
references.  This change gives such code a way of informing the refs
module that it should no longer trust its cache.
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Instead of invalidating the ref cache on an all-or-nothing basis,
invalidate the cache for a specific submodule.
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

It is the cache that is being invalidated, not the references, and the
new name makes this unambiguous.  Rename other items analogously:

* struct cached_refs -> struct ref_cache
* cached_refs (the variable) -> ref_cache
* clear_cached_refs() -> clear_ref_cache()
* create_cached_refs() -> create_ref_cache()
* get_cached_refs() -> get_ref_cache()
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Both dwim_ref()/dwim_log() functions are intimately related to the ref
parsing rules defined in refs.c and better fits there. Move them together
with substitute_branch_name(), a file scope static helper function.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The commit message for c976d415 (git-branch: add options and tests for
branch renaming, 2006-11-28) mentions RENAME_REF but otherwise this is not
documented anywhere, and it does not appear in any of the tests.

Worse yet, the name of the actual file is "RENAMED-REF".

This was supposed to hold the commit object name at the tip of the branch
the most recent "branch -m/-M" renamed, but that is not necessary in order
to be able to recover from a mistake. Even when "branch -M A B" overwrites
an existing branch B, what is kept in RENAMED-REF is the commit at the tip
of the original branch A, not the commit B from the now-lost branch.

Just remove this unused "feature".
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NBrandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The bsearch() implementation on IRIX 6.5 segfaults if it is passed NULL
for the base array argument even if number-of-elements is zero.  So, let's
work around it by detecting an empty array and aborting early.

This is a useful optimization in its own right anyway, since we avoid a
useless allocation and initialization of the ref_entry when the ref array
is empty.
Signed-off-by: NBrandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The variable "refs" is allocated on the stack but is not initialized.  It
is passed to read_packed_refs(), and its struct members may eventually be
passed to add_ref() and ALLOC_GROW().  Since the structure has not been
initialized, its members may contain random non-zero values.  So let's
initialize it.

The call sequence looks something like this:

   resolve_gitlink_packed_ref(...) {

       struct cached_refs refs;
       ...
       read_packed_refs(f, &refs);
       ...
   }

   read_packed_refs(FILE*, struct cached_refs *cached_refs) {
       ...
       add_ref(name, sha1, flag, &cached_refs->packed, &last);
       ...
   }

   add_ref(..., struct ref_array *refs, struct ref_entry **) {
       ...
       ALLOC_GROW(refs->refs, refs->nr + 1, refs->alloc);
   }
Signed-off-by: NBrandon Casey <casey@nrlssc.navy.mil>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

In add_ref(), verify that the refname is formatted correctly before
adding it to the ref_list.  Here we have to allow refname components
that start with ".", since (for example) the remote protocol uses
synthetic reference name ".have".  So add a new REFNAME_DOT_COMPONENT
flag that can be passed to check_refname_format() to allow leading
dots.
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Record information about resolve_ref(), hard-won via reverse
engineering, in a comment for future spelunkers.
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

If the SHA1 in a reference file is not terminated by a space or
end-of-file, consider it malformed and emit a warning.
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

While resolving references, if a reference is found that is in an
unrecognized format, emit a warning (and then fail, as before).
Wouldn't *you* want to know?
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Emit a warning and fail if a symbolic reference refers to an
incorrectly-formatted refname.
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Making it a function and giving it a name makes the code clearer.  I
also have a strong suspicion that the function will find other uses in
the future.
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Immediately strip off trailing spaces and null-terminate the string
holding the contents of the reference file; this allows the use of
string functions and avoids the need to keep separate track of the
string's length.  (get_sha1_hex() fails automatically if the string is
too short.)
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Terminate the link content string one step earlier, allowing
prefixcmp() to be used instead of the less clear memcmp().
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Previously the failure came later, after a few steps in which the
length was treated like the actual length of a string.  Even though
the old code gave the same answers, it was somewhat misleading.
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Since much of the infrastructure does not work correctly with
unnormalized refnames, change check_refname_format() to reject them.

Similarly, change "git check-ref-format" to reject unnormalized
refnames by default.  But add an option --normalize, which causes "git
check-ref-format" to normalize the refname before checking its format,
and print the normalized refname.  This is exactly the behavior of the
old --print option, which is retained but deprecated.
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Allowing any refname component to end with ".lock" is looking for
trouble; for example,

    $ git br foo.lock/bar
    $ git br foo
    fatal: Unable to create '[...]/.git/refs/heads/foo.lock': File exists.

Therefore, do not allow any refname component to end with ".lock".
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Among other things, extract a function check_refname_component().
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Change check_ref_format() to take a flags argument that indicates what
is acceptable in the reference name (analogous to "git
check-ref-format"'s "--allow-onelevel" and "--refspec-pattern").  This
is more convenient for callers and also fixes a failure in the test
suite (and likely elsewhere in the code) by enabling "onelevel" and
"refspec-pattern" to be allowed independently of each other.

Also rename check_ref_format() to check_refname_format() to make it
obvious that it deals with refnames rather than references themselves.
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Previously most bad characters were indicated by returning 1, but "*"
was special-cased to return 2 instead of 1.  One caller examined the
return value to see whether the special case occurred.

But it is easier (to document and understand) for bad_ref_char()
simply to return a boolean value, treating "*" like any other bad
character.  Special-case the handling of "*" (which only occurs in
very specific circumstances) at the caller.  The resulting calling
code thereby also becomes more transparent.
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

Currently we linearly search through lists of refs when we need to
find a specific ref.  This can be very slow if we need to lookup a
large number of refs.  By changing to a binary search we can make this
faster.

In order to be able to use a binary search we need to change from
using linked lists to arrays, which we can manage using ALLOC_GROW.

We can now also use the standard library qsort function to sort the
refs arrays.
Signed-off-by: NJulian Phillips <julian@quantumfyre.co.uk>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

When a push specifies deletion of non-existent refs, the post post-receive and
post-update hooks receive them as input/arguments.

For instance, for the following push, where refs/heads/nonexistent is a ref
which does not exist on the remote side:

	git push origin :refs/heads/nonexistent

the post-receive hook receives from standard input:

	<null-sha1> SP <null-sha1> SP refs/heads/nonexistent

and the post-update hook receives as arguments:

	refs/heads/nonexistent

which does not make sense since it is a no-op.

Teach receive-pack not to pass non-existent refs to the post-receive and
post-update hooks. If the push only attempts to delete non-existent refs,
these hooks are not even called.

The update and pre-receive hooks are still notified about attempted
deletion of non-existent refs to give them a chance to inspect the
situation and act on it.

[jc: mild fix-ups to avoid introducing an extra list; also added fixes to
some tests]
Signed-off-by: NPang Yan Han <pangyanhan@gmail.com>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

get_ref_dir is called recursively for subdirectories, which means that
we were calling sort_ref_list for each directory of refs instead of
once for all the refs.  This is a massive wast of processing, so now
just call sort_ref_list on the result of the top-level get_ref_dir, so
that the sort is only done once.

In the common case of only a few different directories of refs the
difference isn't very noticable, but it becomes very noticeable when
you have a large number of direcotries containing refs (e.g. as
created by Gerrit).

Reported by Martin Fick.
Signed-off-by: NJulian Phillips <julian@quantumfyre.co.uk>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

There is nobody outside that calls into this helper function.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

DEL is an ASCII control character and therefore should not be
permitted in reference names.  Add tests for this and other unusual
characters.
Signed-off-by: NMichael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: NJunio C Hamano <gitster@pobox.com>

The function was not gentle at all to the callers and died without giving
them a chance to deal with possible errors. Rename it to read_gitfile(),
and update all the callers.

As no existing caller needs a true "gently" variant, we do not bother
adding one at this point.
Signed-off-by: NJunio C Hamano <gitster@pobox.com>