2007-11-25 08:48:04 +01:00
|
|
|
tree walking API
|
|
|
|
================
|
|
|
|
|
2009-08-04 06:13:21 +02:00
|
|
|
The tree walking API is used to traverse and inspect trees.
|
2007-11-25 08:48:04 +01:00
|
|
|
|
2009-08-04 06:13:21 +02:00
|
|
|
Data Structures
|
|
|
|
---------------
|
2007-11-25 08:48:04 +01:00
|
|
|
|
2009-08-04 06:13:21 +02:00
|
|
|
`struct name_entry`::
|
|
|
|
|
|
|
|
An entry in a tree. Each entry has a sha1 identifier, pathname, and
|
|
|
|
mode.
|
|
|
|
|
|
|
|
`struct tree_desc`::
|
|
|
|
|
|
|
|
A semi-opaque data structure used to maintain the current state of the
|
|
|
|
walk.
|
|
|
|
+
|
|
|
|
* `buffer` is a pointer into the memory representation of the tree. It always
|
|
|
|
points at the current entry being visited.
|
|
|
|
|
|
|
|
* `size` counts the number of bytes left in the `buffer`.
|
|
|
|
|
|
|
|
* `entry` points to the current entry being visited.
|
|
|
|
|
|
|
|
`struct traverse_info`::
|
|
|
|
|
|
|
|
A structure used to maintain the state of a traversal.
|
|
|
|
+
|
|
|
|
* `prev` points to the traverse_info which was used to descend into the
|
|
|
|
current tree. If this is the top-level tree `prev` will point to
|
|
|
|
a dummy traverse_info.
|
|
|
|
|
|
|
|
* `name` is the entry for the current tree (if the tree is a subtree).
|
|
|
|
|
|
|
|
* `pathlen` is the length of the full path for the current tree.
|
|
|
|
|
|
|
|
* `conflicts` can be used by callbacks to maintain directory-file conflicts.
|
|
|
|
|
|
|
|
* `fn` is a callback called for each entry in the tree. See Traversing for more
|
|
|
|
information.
|
|
|
|
|
|
|
|
* `data` can be anything the `fn` callback would want to use.
|
|
|
|
|
2010-08-11 10:38:07 +02:00
|
|
|
* `show_all_errors` tells whether to stop at the first error or not.
|
|
|
|
|
2009-08-04 06:13:21 +02:00
|
|
|
Initializing
|
|
|
|
------------
|
|
|
|
|
|
|
|
`init_tree_desc`::
|
|
|
|
|
|
|
|
Initialize a `tree_desc` and decode its first entry. The buffer and
|
|
|
|
size parameters are assumed to be the same as the buffer and size
|
|
|
|
members of `struct tree`.
|
|
|
|
|
|
|
|
`fill_tree_descriptor`::
|
|
|
|
|
2017-08-12 10:32:59 +02:00
|
|
|
Initialize a `tree_desc` and decode its first entry given the
|
|
|
|
object ID of a tree. Returns the `buffer` member if the latter
|
|
|
|
is a valid tree identifier and NULL otherwise.
|
2009-08-04 06:13:21 +02:00
|
|
|
|
|
|
|
`setup_traverse_info`::
|
|
|
|
|
|
|
|
Initialize a `traverse_info` given the pathname of the tree to start
|
setup_traverse_info(): stop copying oid
We assume that if setup_traverse_info() is passed a non-empty "base"
string, that string is pointing into a tree object and we can read the
object oid by skipping past the trailing NUL.
As it turns out, this is not true for either of the two calls, and we
may end up reading garbage bytes:
1. In git-merge-tree, our base string is either empty (in which case
we'd never run this code), or it comes from our traverse_path()
helper. The latter overallocates a buffer by the_hash_algo->rawsz
bytes, but then fills it with only make_traverse_path(), leaving
those extra bytes uninitialized (but part of a legitimate heap
buffer).
2. In unpack_trees(), we pass o->prefix, which is some arbitrary
string from the caller. In "git read-tree --prefix=foo", for
instance, it will point to the command-line parameter, and we'll
read 20 bytes past the end of the string.
Interestingly, tools like ASan do not detect (2) because the process
argv is part of a big pre-allocated buffer. So we're reading trash, but
it's trash that's probably part of the next argument, or the
environment.
You can convince it to fail by putting something like this at the
beginning of common-main.c's main() function:
{
int i;
for (i = 0; i < argc; i++)
argv[i] = xstrdup_or_null(argv[i]);
}
That puts the arguments into their own heap buffers, so running:
make SANITIZE=address test
will find problems when "read-tree --prefix" is used (e.g., in t3030).
Doubly interesting, even with the hackery above, this does not fail
prior to ea82b2a085 (tree-walk: store object_id in a separate member,
2019-01-15). That commit switched setup_traverse_info() to actually
copying the hash, rather than simply pointing to it. That pointer was
always pointing to garbage memory, but that commit started actually
dereferencing the bytes, which is what triggers ASan.
That also implies that nobody actually cares about reading these oid
bytes anyway (or at least no path covered by our tests). And manual
inspection of the code backs that up (I'll follow this patch with some
cleanups that show definitively this is the case, but they're quite
invasive, so it's worth doing this fix on its own).
So let's drop the bogus hashcpy(), along with the confusing oversizing
in merge-tree.
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Junio C Hamano <gitster@pobox.com>
2019-07-31 06:38:11 +02:00
|
|
|
traversing from.
|
2009-08-04 06:13:21 +02:00
|
|
|
|
|
|
|
Walking
|
|
|
|
-------
|
|
|
|
|
|
|
|
`tree_entry`::
|
|
|
|
|
|
|
|
Visit the next entry in a tree. Returns 1 when there are more entries
|
|
|
|
left to visit and 0 when all entries have been visited. This is
|
|
|
|
commonly used in the test of a while loop.
|
|
|
|
|
|
|
|
`tree_entry_len`::
|
|
|
|
|
|
|
|
Calculate the length of a tree entry's pathname. This utilizes the
|
|
|
|
memory structure of a tree entry to avoid the overhead of using a
|
|
|
|
generic strlen().
|
|
|
|
|
|
|
|
`update_tree_entry`::
|
|
|
|
|
|
|
|
Walk to the next entry in a tree. This is commonly used in conjunction
|
|
|
|
with `tree_entry_extract` to inspect the current entry.
|
|
|
|
|
|
|
|
`tree_entry_extract`::
|
|
|
|
|
|
|
|
Decode the entry currently being visited (the one pointed to by
|
|
|
|
`tree_desc's` `entry` member) and return the sha1 of the entry. The
|
|
|
|
`pathp` and `modep` arguments are set to the entry's pathname and mode
|
|
|
|
respectively.
|
|
|
|
|
|
|
|
`get_tree_entry`::
|
|
|
|
|
|
|
|
Find an entry in a tree given a pathname and the sha1 of a tree to
|
|
|
|
search. Returns 0 if the entry is found and -1 otherwise. The third
|
|
|
|
and fourth parameters are set to the entry's sha1 and mode
|
|
|
|
respectively.
|
|
|
|
|
|
|
|
Traversing
|
|
|
|
----------
|
|
|
|
|
|
|
|
`traverse_trees`::
|
|
|
|
|
|
|
|
Traverse `n` number of trees in parallel. The `fn` callback member of
|
|
|
|
`traverse_info` is called once for each tree entry.
|
|
|
|
|
|
|
|
`traverse_callback_t`::
|
|
|
|
The arguments passed to the traverse callback are as follows:
|
|
|
|
+
|
|
|
|
* `n` counts the number of trees being traversed.
|
|
|
|
|
|
|
|
* `mask` has its nth bit set if something exists in the nth entry.
|
|
|
|
|
|
|
|
* `dirmask` has its nth bit set if the nth tree's entry is a directory.
|
|
|
|
|
|
|
|
* `entry` is an array of size `n` where the nth entry is from the nth tree.
|
|
|
|
|
|
|
|
* `info` maintains the state of the traversal.
|
|
|
|
|
|
|
|
+
|
|
|
|
Returning a negative value will terminate the traversal. Otherwise the
|
|
|
|
return value is treated as an update mask. If the nth bit is set the nth tree
|
|
|
|
will be updated and if the bit is not set the nth tree entry will be the
|
|
|
|
same in the next callback invocation.
|
|
|
|
|
|
|
|
`make_traverse_path`::
|
|
|
|
|
|
|
|
Generate the full pathname of a tree entry based from the root of the
|
|
|
|
traversal. For example, if the traversal has recursed into another
|
|
|
|
tree named "bar" the pathname of an entry "baz" in the "bar"
|
|
|
|
tree would be "bar/baz".
|
|
|
|
|
|
|
|
`traverse_path_len`::
|
|
|
|
|
|
|
|
Calculate the length of a pathname returned by `make_traverse_path`.
|
|
|
|
This utilizes the memory structure of a tree entry to avoid the
|
|
|
|
overhead of using a generic strlen().
|
|
|
|
|
2019-07-31 06:38:23 +02:00
|
|
|
`strbuf_make_traverse_path`::
|
|
|
|
|
|
|
|
Convenience wrapper to `make_traverse_path` into a strbuf.
|
|
|
|
|
2009-08-04 06:13:21 +02:00
|
|
|
Authors
|
|
|
|
-------
|
|
|
|
|
|
|
|
Written by Junio C Hamano <gitster@pobox.com> and Linus Torvalds
|
|
|
|
<torvalds@linux-foundation.org>
|