From: Derrick Stolee <stolee@gmail.com>
To: Junio C Hamano <gitster@pobox.com>,
Derrick Stolee via GitGitGadget <gitgitgadget@gmail.com>
Cc: git@vger.kernel.org, newren@gmail.com,
"René Scharfe" <l.s.r@web.de>,
"Derrick Stolee" <derrickstolee@github.com>,
"Derrick Stolee" <dstolee@microsoft.com>
Subject: Re: [PATCH v2 6/9] index-format: update preamble to cached tree extension
Date: Thu, 7 Jan 2021 06:51:03 -0500 [thread overview]
Message-ID: <5a1a51c1-7ada-6f66-a6d0-2fc01303d334@gmail.com> (raw)
In-Reply-To: <xmqqble1o5sc.fsf@gitster.c.googlers.com>
On 1/6/2021 9:10 PM, Junio C Hamano wrote:
> "Derrick Stolee via GitGitGadget" <gitgitgadget@gmail.com> writes:
>
>> Subject: Re: [PATCH v2 6/9] index-format: update preamble to cached tree extension
>
> By the way, the name of the extension is "cache tree".
>
> git grep -i 'cached[- ]tree' ':!po/'
>
> reveals there are a handful of mistakes already present but their
> number is dwarfed when we check:
>
> git grep -i 'cache tree' ':!po/'
I will fix my own additions and add a patch that fixes these mistakes.
>> + Since the index does not record entries for directories, the cache
>> + entries cannot describe tree objects that already exist in the object
>> + database for regions of the index that are unchanged from an existing
>> + commit. The cached tree extension stores a recursive tree structure that
>> + describes the trees that already exist and completely match sections of
>> + the cache entries. This speeds up tree object generation from the index
>> + for a new commit by only computing the trees that are "new" to that
>> + commit.
>
> The original motivation was the above one. A cache of tree objects
> that correspond to unmodified part of the directory structure helps
> writing out a new tree out of modified index.
>
> We later found out that we rather often compare the index against
> the tree of HEAD (think: "git status"), and diff-lib.c::diff_cache()
> does take advantage of the fact that an entire directory can be
> skipped if the tree object taken from the HEAD side exactly matches
> the tree recorded for the subdirectory in the cache tree extension.
I need to read more about this. traverse_by_cache_tree() seems to
be a good place to start. Thanks.
>> + The recursive tree structure uses nodes that store a number of cache
>> + entries, a list of subnodes, and an object ID (OID). The OID references
>> + the exising tree for that node, if it is known to exist. The subnodes
>> + correspond to subdirectories that themselves have cached tree nodes. The
>> + number of cache entries corresponds to the number of cache entries in
>> + the index that describe paths within that tree's directory.
s/exising/existing/
>
> OK.
>
>> + Note that the path for a given tree is part of the parent node in-memory
>
> Sorry, I am not sure if I follow. The top-level in-core cache_tree
> object records the number of entries, tree object name for the
> entire tree (if valid), and cache_tree_sub structures, one for each
> subdirectory. Each of the cache_tree_sub structure describes the
> "child" directory, including the path to it.
>
>> + but is part of the child in the file format. The root tree has an empty
>> + string for its name and its name does not exist in-memory.
>
> It's more like we could have consistently used cache_tree_sub
> instances to represent each and every level (i.e. I consider that
> cache_tree_sub is what represents a directory, with cache_tree being
> a record of just one aspect of it) including the root of the
> hierarchy, but because there wasn't much point in giving a name to
> the root level, I cheated and avoided wasting a cache_tree_sub for
> it. So from that point of view, the path belongs to the node in
> each level in both in-core and on-disk representations.
That's a good point. I'll retract my statement here.
>> + When a path is updated in index, Git invalidates all nodes of the
>> + recurisive cached tree corresponding to the parent directories of that
>> + path. We store these tree nodes as being "invalid" by using "-1" as the
>> + number of cache entries.
>
> Correct.
Making note of my s/recurisive/recursive/ typo here.
>> + To create trees corresponding to the current
>> + index, Git only walks the invalid tree nodes and uses the cached OIDs
>> + for the valid trees to construct new trees.
>
> I wonder if the above is sufficiently clear, or "Git only has to
> walk the spans of index entries that corresponds to the invalid
> trees, while reusing the ..." is too long and detailed.
I will try to simplify.
Thanks,
-Stolee
next prev parent reply other threads:[~2021-01-07 11:51 UTC|newest]
Thread overview: 53+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-12-30 19:26 [PATCH 0/8] Cleanups around index operations Derrick Stolee via GitGitGadget
2020-12-30 19:26 ` [PATCH 1/8] tree-walk: report recursion counts Derrick Stolee via GitGitGadget
2020-12-30 19:42 ` Elijah Newren
2020-12-30 19:51 ` Derrick Stolee
2020-12-30 19:26 ` [PATCH 2/8] unpack-trees: add trace2 regions Derrick Stolee via GitGitGadget
2020-12-30 19:45 ` Elijah Newren
2020-12-30 19:26 ` [PATCH 3/8] cache-tree: use trace2 in cache_tree_update() Derrick Stolee via GitGitGadget
2020-12-30 19:26 ` [PATCH 4/8] cache-tree: trace regions for I/O Derrick Stolee via GitGitGadget
2020-12-30 19:26 ` [PATCH 5/8] cache-tree: trace regions for prime_cache_tree Derrick Stolee via GitGitGadget
2020-12-30 19:48 ` Elijah Newren
2020-12-30 19:53 ` Derrick Stolee
2020-12-30 19:26 ` [PATCH 6/8] index-format: update preamble to cached tree extension Derrick Stolee via GitGitGadget
2020-12-30 20:00 ` Elijah Newren
2020-12-30 19:26 ` [PATCH 7/8] index-format: discuss recursion of cached-tree better Derrick Stolee via GitGitGadget
2020-12-30 19:26 ` [PATCH 8/8] cache-tree: avoid path comparison loop when silent Derrick Stolee via GitGitGadget
2020-12-30 20:14 ` Elijah Newren
2021-01-06 8:55 ` Junio C Hamano
2021-01-06 12:08 ` Derrick Stolee
2020-12-31 12:34 ` René Scharfe
2020-12-31 16:46 ` Derrick Stolee
2021-01-01 13:30 ` René Scharfe
2021-01-02 15:19 ` [PATCH] cache-tree: use ce_namelen() instead of strlen() René Scharfe
2021-01-04 1:26 ` Derrick Stolee
2021-01-05 12:05 ` Junio C Hamano
2021-01-02 15:31 ` [PATCH 8/8] cache-tree: avoid path comparison loop when silent René Scharfe
2020-12-30 20:19 ` [PATCH 0/8] Cleanups around index operations Elijah Newren
2020-12-30 20:24 ` Derrick Stolee
2021-01-04 3:09 ` [PATCH v2 0/9] " Derrick Stolee via GitGitGadget
2021-01-04 3:09 ` [PATCH v2 1/9] tree-walk: report recursion counts Derrick Stolee via GitGitGadget
2021-01-04 3:09 ` [PATCH v2 2/9] unpack-trees: add trace2 regions Derrick Stolee via GitGitGadget
2021-01-04 3:09 ` [PATCH v2 3/9] cache-tree: use trace2 in cache_tree_update() Derrick Stolee via GitGitGadget
2021-01-04 3:09 ` [PATCH v2 4/9] cache-tree: trace regions for I/O Derrick Stolee via GitGitGadget
2021-01-04 3:09 ` [PATCH v2 5/9] cache-tree: trace regions for prime_cache_tree Derrick Stolee via GitGitGadget
2021-01-04 3:09 ` [PATCH v2 6/9] index-format: update preamble to cached tree extension Derrick Stolee via GitGitGadget
2021-01-07 2:10 ` Junio C Hamano
2021-01-07 11:51 ` Derrick Stolee [this message]
2021-01-07 20:12 ` Junio C Hamano
2021-01-07 21:26 ` Junio C Hamano
2021-01-04 3:09 ` [PATCH v2 7/9] index-format: discuss recursion of cached-tree better Derrick Stolee via GitGitGadget
2021-01-04 3:09 ` [PATCH v2 8/9] cache-tree: use ce_namelen() instead of strlen() René Scharfe via GitGitGadget
2021-01-04 3:09 ` [PATCH v2 9/9] cache-tree: speed up consecutive path comparisons Derrick Stolee via GitGitGadget
2021-01-07 16:32 ` [PATCH v3 00/10] Cleanups around index operations Derrick Stolee via GitGitGadget
2021-01-07 16:32 ` [PATCH v3 01/10] tree-walk: report recursion counts Derrick Stolee via GitGitGadget
2021-01-07 16:32 ` [PATCH v3 02/10] unpack-trees: add trace2 regions Derrick Stolee via GitGitGadget
2021-01-07 16:32 ` [PATCH v3 03/10] cache-tree: use trace2 in cache_tree_update() Derrick Stolee via GitGitGadget
2021-01-07 16:32 ` [PATCH v3 04/10] cache-tree: trace regions for I/O Derrick Stolee via GitGitGadget
2021-01-07 16:32 ` [PATCH v3 05/10] cache-tree: trace regions for prime_cache_tree Derrick Stolee via GitGitGadget
2021-01-07 16:32 ` [PATCH v3 06/10] index-format: use 'cache tree' over 'cached tree' Derrick Stolee via GitGitGadget
2021-01-07 16:32 ` [PATCH v3 07/10] index-format: update preamble to cache tree extension Derrick Stolee via GitGitGadget
2021-01-07 16:32 ` [PATCH v3 08/10] index-format: discuss recursion of cached-tree better Derrick Stolee via GitGitGadget
2021-01-07 16:32 ` [PATCH v3 09/10] cache-tree: use ce_namelen() instead of strlen() René Scharfe via GitGitGadget
2021-01-07 16:32 ` [PATCH v3 10/10] cache-tree: speed up consecutive path comparisons Derrick Stolee via GitGitGadget
2021-01-16 6:58 ` [PATCH v3 00/10] Cleanups around index operations Junio C Hamano
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=5a1a51c1-7ada-6f66-a6d0-2fc01303d334@gmail.com \
--to=stolee@gmail.com \
--cc=derrickstolee@github.com \
--cc=dstolee@microsoft.com \
--cc=git@vger.kernel.org \
--cc=gitgitgadget@gmail.com \
--cc=gitster@pobox.com \
--cc=l.s.r@web.de \
--cc=newren@gmail.com \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).