When a reflog notation is used for implicit "current branch", we did
not say which branch, and worse said "branch ''".
* rr/die-on-missing-upstream:
sha1_name: fix error message for @{<N>}, @{<date>}
sha1_name: fix error message for @{u}
One test in t1512 that expects a failure incorrectly passed. The
test prepares a commit whose object name begins with ten "0"s, and
also prepares a tag that points at the commit. The object name of
the tag also begins with ten "0"s. There is no other commit-ish
object in the repository whose name begins with such a prefix.
Ideally, in such a repository:
$ git rev-parse --verify 0000000000^{commit}
should yield that commit. If 0000000000 is taken as the commit
0000000000e4f, peeling it to a commmit yields that commit itself,
and if 0000000000 is taken as the tag 0000000000f8f, peeling it to a
commit also yields the same commit, so in that twisted sense, the
extended SHA-1 expression 0000000000^{commit} is unambigous. The
test that expects a failure is to check the above command.
The reason the test expects a failure is that we did not implement
such a "unification" of two candidate objects. What we did (or at
least, meant to) implement was to recognise that a commit-ish is
required to expand 0000000000, and notice that there are two succh
commit-ish, and diagnose the request as ambiguous.
However, there was a bug in the logic to check the candidate
objects. When the code saw 0000000000f8f (a tag) that shared the
shortened prefix (ten "0"s), it tried to make sure that the tag is a
commit-ish by looking at the tag object. Because it incorrectly
used lookup_object() when the tag has not been parsed, however, we
incorrectly declared that the tag is _not_ a commit-ish, leaving the
sole commit in the repository, 0000000000e4f, that has the required
prefix as "unique match", causing the test to pass when it shouldn't.
This fixes the logic to inspect the type of the object a tag refers
to, to make the test that is expected to fail correctly fail.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Currently, when we try to resolve @{<N>} or @{<date>} when the reflog
doesn't go back far enough, we get errors like:
# on branch master
$ git show @{10000}
fatal: Log for '' only has 7 entries.
$ git show @{10000.days.ago}
warning: Log for '' only goes back to Tue, 21 May 2013 14:14:45 +0530.
...
# detached HEAD case
$ git show @{10000}
fatal: Log for '' only has 2005 entries.
$ git show master@{10000}
fatal: Log for 'master' only has 7 entries.
The empty string '' is confusing and does not convey information
about whose logs we are inspecting. Change this so that we get:
# on branch master
$ git show @{10000}
fatal: Log for 'master' only has 7 entries.
$ git show @{10000.days.ago}
warning: Log for 'master' only goes back to Tue, 21 May 2013 14:14:45 +0530.
...
# detached HEAD case
$ git show @{10000}
fatal: Log for 'HEAD' only has 2005 entries.
$ git show master@{10000}
fatal: Log for 'master' only has 7 entries.
Also one of the message strings given to die() now points into
real_ref that was not used in that fashion, so stop freeing the
underlying storage for it.
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Bug-spotted-and-fixed-by: Thomas Rast
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Currently, when no (valid) upstream is configured for a branch, you get
an error like:
$ git show @{u}
error: No upstream configured for branch 'upstream-error'
error: No upstream configured for branch 'upstream-error'
fatal: ambiguous argument '@{u}': unknown revision or path not in the working tree.
Use '--' to separate paths from revisions, like this:
'git <command> [<revision>...] -- [<file>...]'
The "error: " line actually appears twice, and the rest of the error
message is useless. In sha1_name.c:interpret_branch_name(), there is
really no point in processing further if @{u} couldn't be resolved, and
we might as well die() instead of returning an error(). After making
this change, you get:
$ git show @{u}
fatal: No upstream configured for branch 'upstream-error'
Also tweak a few tests in t1507 to expect this output.
This only turns error() that may be called after we know we are
dealing with an @{upstream} marker into die(), without touching
silent error returns "return -1" from the function. Any caller that
wants to handle an error condition itself will not be hurt by this
change, unless they want to see the message from error() and then
exit silently without giving its own message, which needs to be
fixed anyway.
Signed-off-by: Ramkumar Ramachandra <artagnon@gmail.com>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If you were on 'frotz' branch before you checked out your current
branch, "git merge @{-1}~22" means the same as "git merge frotz~22".
The strbuf_branchname() function, when interpret_branch_name() gives
up resolving "@{-1}~22" fully, returns "frotz" and tells the caller
that it only resolved "@{-1}" part of the input, mistakes this as a
total failure, and appends the whole thing to the result, yielding
"frotz@{-1}~22", which does not make any sense.
Inspect the return value from interpret_branch_name() a bit more
carefully. When it errored out without consuming anything, we will
get -1 and we should return the whole thing. Otherwise, we should
append the remainder (i.e. "~22" in the earlier example) to the
partially resolved name (i.e. "frotz").
The test suite adds enough number of checkout to make @{-12} in the
last test in t0100 that tried to check "we haven't flipped branches
that many times" error case succeed; raise the number to a hundred.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There was no good way to ask "I have a random string that came from
outside world. I want to turn it into a 40-hex object name while
making sure such an object exists". A new peeling suffix ^{object}
can be used for that purpose, together with "rev-parse --verify".
* jc/sha1-name-object-peeler:
peel_onion(): teach $foo^{object} peeler
peel_onion: disambiguate to favor tree-ish when we know we want a tree-ish
A string that names an object can be suffixed with ^{type} peeler to
say "I have this object name; peel it until you get this type. If
you cannot do so, it is an error". v1.8.2^{commit} asks for a commit
that is pointed at an annotated tag v1.8.2; v1.8.2^{tree} unwraps it
further to the top-level tree object. A special suffix ^{} (i.e. no
type specified) means "I do not care what it unwraps to; just peel
annotated tag until you get something that is not a tag".
When you have a random user-supplied string, you can turn it to a
bare 40-hex object name, and cause it to error out if such an object
does not exist, with:
git rev-parse --verify "$userstring^{}"
for most objects, but this does not yield the tag object name when
$userstring refers to an annotated tag.
Introduce a new suffix, ^{object}, that only makes sure the given
name refers to an existing object. Then
git rev-parse --verify "$userstring^{object}"
becomes a way to make sure $userstring refers to an existing object.
This is necessary because the plumbing "rev-parse --verify" is only
about "make sure the argument is something we can feed to get_sha1()
and turn it into a raw 20-byte object name SHA-1" and is not about
"make sure that 20-byte object name SHA-1 refers to an object that
exists in our object store". When the given $userstring is already
a 40-hex, by definition "rev-parse --verify $userstring" can turn it
into a raw 20-byte object name. With "$userstring^{object}", we can
make sure that the 40-hex string names an object that exists in our
object store before "--verify" kicks in.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The function already knows when interpreting $foo^{commit} to tell
the underlying get_sha1_1() to expect a commit-ish while evaluating
$foo. Teach it to do the same when asked for $foo^{tree}; we are
expecting a tree-ish and $foo should be disambiguated in favor of a
tree-ish, discarding a possible ambiguous match with a blob object.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
An internal function used to implement "git checkout @{-1}" was
hard to use correctly.
* jc/reflog-reverse-walk:
refs.c: fix fread error handling
reflog: add for_each_reflog_ent_reverse() API
for_each_recent_reflog_ent(): simplify opening of a reflog file
for_each_reflog_ent(): extract a helper to process a single entry
The only caller of diagnose_invalid_sha1_path() extracts a substring from
an object name by creating a NUL-terminated copy of the interesting part.
Add a length parameter to the function and thus avoid the need for an
allocation, thereby simplifying the code.
Signed-off-by: Rene Scharfe <rene.scharfe@lsrfire.ath.cx>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
"git checkout -" is a short-hand for "git checkout @{-1}" and the
"@{nth}" notation for a negative number is to find nth previous
checkout in the reflog of the HEAD to determine the name of the
branch the user was on. We would want to find the nth most recent
reflog entry that matches "checkout: moving from X to Y" for this.
Unfortunately, reflog is implemented as an append-only file, and the
API to iterate over its entries, for_each_reflog_ent(), reads the
file in order, giving the entries from the oldest to newer. For the
purpose of finding nth most recent one, this API forces us to record
the last n entries in a rotating buffer and give the result out only
after we read everything. To optimize for a common case of finding
the nth most recent one for a small value of n, we also have a side
API for_each_recent_reflog_ent() that starts reading near the end of
the file, but it still has to read the entries in the "wrong" order.
The implementation of understanding @{-1} uses this interface.
This all becomes unnecessary if we add an API to let us iterate over
reflog entries in the reverse order, from the newest to older.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Teaches the object name parser things like a "git describe" output
is always a commit object, "A" in "git log A" must be a committish,
and "A" and "B" in "git log A...B" both must be committish, etc., to
prolong the lifetime of abbreviated object names.
* jc/sha1-name-more: (27 commits)
t1512: match the "other" object names
t1512: ignore whitespaces in wc -l output
rev-parse --disambiguate=<prefix>
rev-parse: A and B in "rev-parse A..B" refer to committish
reset: the command takes committish
commit-tree: the command wants a tree and commits
apply: --build-fake-ancestor expects blobs
sha1_name.c: add support for disambiguating other types
revision.c: the "log" family, except for "show", takes committish
revision.c: allow handle_revision_arg() to take other flags
sha1_name.c: introduce get_sha1_committish()
sha1_name.c: teach lookup context to get_sha1_with_context()
sha1_name.c: many short names can only be committish
sha1_name.c: get_sha1_1() takes lookup flags
sha1_name.c: get_describe_name() by definition groks only commits
sha1_name.c: teach get_short_sha1() a commit-only option
sha1_name.c: allow get_short_sha1() to take other flags
get_sha1(): fix error status regression
sha1_name.c: restructure disambiguation of short names
sha1_name.c: correct misnamed "canonical" and "res"
...
The new option allows you to feed an ambiguous prefix and enumerate
all the objects that share it as a prefix of their object names.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This teaches the revision parser that in "$name:$path" (used for a
blob object name), "$name" must be a tree-ish.
There are many more places where we know what types of objects are
called for. This patch adds support for "commit", "treeish", "tree",
and "blob", which could be used in the following contexts:
- "git apply --build-fake-ancestor" reads the "index" lines from
the patch; they must name blob objects (not even "blob-ish");
- "git commit-tree" reads a tree object name (not "tree-ish"), and
zero or more commit object names (not "committish");
- "git reset $rev" wants a committish; "git reset $rev -- $path"
wants a treeish.
They will come in later patches in the series.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Many callers know that the user meant to name a committish by
syntactical positions where the object name appears. Calling this
function allows the machinery to disambiguate shorter-than-unique
abbreviated object names between committish and others.
Note that this does NOT error out when the named object is not a
committish. It is merely to give a hint to the disambiguation
machinery.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The function takes user input string and returns the object name
(binary SHA-1) with mode bits and path when the object was looked
up in a tree.
Additionally give hints to help disambiguation of abbreviated object
names when the caller knows what it is looking for.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We know that the token "$name" that appear in "$name^{commit}",
"$name^4", "$name~4" etc. can only name a committish (either a
commit or a tag that peels to a commit). Teach get_short_sha1() to
take advantage of that knowledge when disambiguating an abbreviated
SHA-1 given as an object name.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This is to pass the disambiguation hints from the caller down the
callchain. Nothing is changed in this step, as everybody just
passes 0 in the flag.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Teach get_describe_name() to pass the disambiguation hint down the
callchain to get_short_sha1().
Also add tests to show various syntactic elements that we could take
advantage of the object type information to help disambiguration of
abbreviated object names. Many of them are marked as broken, and
some of them will be fixed in later patches in this series.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When the caller knows that the parameter is meant to name a commit,
e.g. "56789a" in describe name "v1.2.3-4-g56789a", pass that as a
hint so that lower level can use it to disambiguate objects when
there is only one commit whose name begins with 56789a even if there
are objects of other types whose names share the same prefix.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Instead of a separate "int quietly" argument, make it take "unsigned
flags" so that we can pass other options to it.
The bit assignment of this flag word is exposed in cache.h because
the mechanism will be exposed to callers of the higher layer in
later commits in this series.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
In finish_object_disambiguation(), if the candidate hasn't been
checked, there are two cases:
- It is the first and only object that match the prefix; or
- It replaced another object that matched the prefix but that
object did not satisfy ds->fn() callback.
And the former case we set ds->candidate_ok to true without doing
anything else, while for the latter we check the candidate, which
may set ds->candidate_ok to false.
At this point in the code, ds->candidate_ok can be false only if
this last-round check found that the candidate does not pass the
check, because the state after update_candidates() returns cannot
satisfy
!ds->ambiguous && ds->candidate_exists && ds->candidate_checked
and !ds->canidate_ok at the same time.
Hence, when we execute this "return", we know we have seen more than
one object that match the prefix (and none of them satisfied ds->fn),
meaning that we should say "the short name is ambiguous", not "there
is no object that matches the prefix".
Noticed by Jeff King.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
We try to find zero, one or more matches from loose objects and
packed objects independently and then decide if the given short
object name is unique across them.
Instead, introduce a "struct disambiguate_state" that keeps track of
what we have found so far, that can be one of:
- We have seen one object that _could_ be what we are looking for;
- We have also checked that object for additional constraints (if any),
and found that the object satisfies it;
- We have also checked that object for additional constraints (if any),
and found that the object does not satisfy it; or
- We have seen more than one objects that satisfy the constraints.
and pass it to the enumeration functions for loose and packed
objects. The disambiguation state can optionally take a callback
function that takes a candidate object name and reports if the
object satisifies additional criteria (e.g. when the caller knows
that the short name must refer to a commit, this mechanism can be
used to check the type of the given object).
Compared to the earlier attempt, this round avoids the optional
check if there is only one candidate that matches the short name in
the first place.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
These are hexadecimal and binary representation of the short object
name given to the callchain as its input. Rename them with _pfx
suffix to make it clear they are prefixes, and call them hex and bin
respectively.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Extract the logic to find object(s) that match a given prefix inside
a single pack into a separate helper function, and give it a bit more
comment.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This variable points at the element we are currently looking at, and
does not have anything to do with the current time which the name
"now" implies.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
There are only two callers, and they will benefit from being able to
pass disambiguation hints to underlying get_sha1_with_context() API
once it happens.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
The only external caller is setup.c that tries to give a nicer error
message when an object name is misspelt (e.g. "HEAD:cashe.h").
Retire it and give the caller a dedicated and more intuitive API
function maybe_die_on_misspelt_object_name().
Signed-off-by: Junio C Hamano <gitster@pobox.com>
diagnose_invalid_sha1_path() is meant to be called to diagnose a
misspelt <treeish>:<pathname> when <pathname> does not exist in
<treeish>. However, the code may call it if <treeish>:<pathname> is
invalid (which triggers another call with only_to_die == 1), but for
another reason. This happens when calling e.g.
git log existing-file HEAD:existing-file
because existing-file is a path and not a revision, the code
verifies that the arguments that follow to be paths. This leads to
an incorrect message like "existing-file does not exist in HEAD",
even though the path exists in HEAD.
Check that the search for <pathname> in <treeish> fails before
triggering the diagnosis.
Signed-off-by: Matthieu Moy <Matthieu.Moy@imag.fr>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
If the branch configured as upstream didn't have a local tracking
branch, git said "Upstream branch not found". We can be more helpful,
and separate the cases when upstream is not configured, and when it is
configured, but the upstream branch is not tracked in a local branch.
The following configuration leads to the second scenario:
[remote "origin"]
url = ...
fetch = refs/heads/master
[branch "master"]
remote = origin
merge = refs/heads/master
'git pull' will work on master, but master@{upstream} is not defined.
Signed-off-by: Zbigniew Jędrzejewski-Szmek <zbyszek@in.waw.pl>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Instead of just saying that no upstream exists for such branch,
which is true but not very helpful, check that there's no
refs/heads/barnhc_wiht_tpyo and tell it to the user.
Signed-off-by: Zbigniew Jędrzejewski-Szmek <zbyszek@in.waw.pl>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
When using @{u} or @{upstream} it is common to omit the branch name,
implying current branch. If the upstream is not configured, the error
message was "No upstream branch found for ''".
When resolving '@{u}', branch_get() is called, which almost always
returns a description of a branch. This allows us to use a branch name
in the error message, even if the user said something like '@{u}'.
The only case when branch_get() returns NULL is when HEAD points to so
something which is not a branch. Of course this also means that no
upstream is configured, but it is better to directly say that HEAD
does not point to a branch.
Signed-off-by: Zbigniew Jędrzejewski-Szmek <zbyszek@in.waw.pl>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* jc/broken-ref-dwim-fix:
resolve_ref(): report breakage to the caller without warning
resolve_ref(): expose REF_ISBROKEN flag
refs.c: move dwim_ref()/dwim_log() from sha1_name.c
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: Junio C Hamano <gitster@pobox.com>
* mh/check-ref-format-3: (23 commits)
add_ref(): verify that the refname is formatted correctly
resolve_ref(): expand documentation
resolve_ref(): also treat a too-long SHA1 as invalid
resolve_ref(): emit warnings for improperly-formatted references
resolve_ref(): verify that the input refname has the right format
remote: avoid passing NULL to read_ref()
remote: use xstrdup() instead of strdup()
resolve_ref(): do not follow incorrectly-formatted symbolic refs
resolve_ref(): extract a function get_packed_ref()
resolve_ref(): turn buffer into a proper string as soon as possible
resolve_ref(): only follow a symlink that contains a valid, normalized refname
resolve_ref(): use prefixcmp()
resolve_ref(): explicitly fail if a symlink is not readable
Change check_refname_format() to reject unnormalized refnames
Inline function refname_format_print()
Make collapse_slashes() allocate memory for its result
Do not allow ".lock" at the end of any refname component
Refactor check_refname_format()
Change check_ref_format() to take a flags argument
Change bad_ref_char() to return a boolean value
...
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: Michael Haggerty <mhagger@alum.mit.edu>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
It probably is not such a good idea to use ":/<pattern>" to specify which
commit to merge, as ":/<pattern>" can often hit unexpected commits, but
somebody tried it and got a nonsense error message:
fatal: ':/Foo bar' does not point to a commit
So here is a for-the-sake-of-consistency update that is fairly useless
that allows users to carefully try not shooting in the foot.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* jc/magic-pathspec:
setup.c: Fix some "symbol not declared" sparse warnings
t3703: Skip tests using directory name ":" on Windows
revision.c: leave a note for "a lone :" enhancement
t3703, t4208: add test cases for magic pathspec
rev/path disambiguation: further restrict "misspelled index entry" diag
fix overslow :/no-such-string-ever-existed diagnostics
fix overstrict :<path> diagnosis
grep: use get_pathspec() correctly
pathspec: drop "lone : means no pathspec" from get_pathspec()
Revert "magic pathspec: add ":(icase)path" to match case insensitively"
magic pathspec: add ":(icase)path" to match case insensitively
magic pathspec: futureproof shorthand form
magic pathspec: add tentative ":/path/from/top/level" pathspec support
"git cmd :/no-such-string-ever-existed" runs an extra round of get_sha1()
since 009fee4 (Detailed diagnosis when parsing an object name fails.,
2009-12-07). Once without error diagnosis to see there is no commit with
such a string in the log message (hence "it cannot be a ref"), and after
seeing that :/no-such-string-ever-existed is not a filename (hence "it
cannot be a path, either"), another time to give "better diagnosis".
The thing is, the second time it runs, we already know that traversing the
history all the way down to the root will _not_ find any matching commit.
Rename misguided "gently" parameter, which is turned off _only_ when the
"detailed diagnosis" codepath knows that it cannot be a ref and making the
call only for the caller to die with a message. Flip its meaning (and
adjust the callers) and call it "only_to_die", which is not a great name,
but it describes far more clearly what the codepaths that switches their
behaviour based on this variable do.
On my box, the command spends ~1.8 seconds without the patch to make the
report; with the patch it spends ~1.12 seconds.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Given "git log :", we get a disambiguation message that tries to be
helpful and yet totally misses the point, i.e.
$ git log :
fatal: Path '' does not exist (neither on disk nor in the index).
$ git log :/
fatal: Path '/' exists on disk, but not in the index.
An empty path nor anything that begins with '/' cannot possibly in the
index, and it is wrong to guess that the user might have meant to access
such an index entry.
It should yield the same error message as "git log '*.c'", i.e.
$ git log '*.c'
fatal: ambiguous argument '*.c': unknown revision or path not in the working tree.
Use '--' to separate paths from revisions
Signed-off-by: Junio C Hamano <gitster@pobox.com>
Currently, the "Did you mean..." message suggests "commit:fullpath"
only. Extend this to show the more convenient "commit:./file" form also.
Signed-off-by: Michael J Gruber <git@drmicha.warpmail.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
This reverts commit 72a5b561fc, as adding
fixed number of hexdigits more than necessary to make one object name
locally unique does not help in futureproofing the uniqueness of names
we generate today.
Signed-off-by: Junio C Hamano <gitster@pobox.com>
* nd/oneline-sha1-name-from-specific-ref:
get_sha1: handle special case $commit^{/}
get_sha1: support $commit^{/regex} syntax
get_sha1_oneline: make callers prepare the commit list to traverse
get_sha1_oneline: fix lifespan rule of temp_commit_buffer variable
