From: Christian Couder <christian.couder@gmail.com>
To: git@vger.kernel.org
Cc: Junio C Hamano <gitster@pobox.com>, Jeff King <peff@peff.net>,
Duy Nguyen <pclouds@gmail.com>,
Johannes Schindelin <Johannes.Schindelin@gmx.de>,
Stefan Beller <sbeller@google.com>,
Christian Couder <chriscool@tuxfamily.org>
Subject: [PATCH v3 6/8] t: add t5319-delta-islands.sh
Date: Thu, 9 Aug 2018 17:55:30 +0200 [thread overview]
Message-ID: <20180809155532.26151-7-chriscool@tuxfamily.org> (raw)
In-Reply-To: <20180809155532.26151-1-chriscool@tuxfamily.org>
From: Jeff King <peff@peff.net>
Signed-off-by: Jeff King <peff@peff.net>
Signed-off-by: Christian Couder <chriscool@tuxfamily.org>
---
t/t5319-delta-islands.sh | 143 +++++++++++++++++++++++++++++++++++++++
1 file changed, 143 insertions(+)
create mode 100755 t/t5319-delta-islands.sh
diff --git a/t/t5319-delta-islands.sh b/t/t5319-delta-islands.sh
new file mode 100755
index 0000000000..fea92a5777
--- /dev/null
+++ b/t/t5319-delta-islands.sh
@@ -0,0 +1,143 @@
+#!/bin/sh
+
+test_description='exercise delta islands'
+. ./test-lib.sh
+
+# returns true iff $1 is a delta based on $2
+is_delta_base () {
+ delta_base=$(echo "$1" | git cat-file --batch-check='%(deltabase)') &&
+ echo >&2 "$1 has base $delta_base" &&
+ test "$delta_base" = "$2"
+}
+
+# generate a commit on branch $1 with a single file, "file", whose
+# content is mostly based on the seed $2, but with a unique bit
+# of content $3 appended. This should allow us to see whether
+# blobs of different refs delta against each other.
+commit() {
+ blob=$({ test-tool genrandom "$2" 10240 && echo "$3"; } |
+ git hash-object -w --stdin) &&
+ tree=$(printf '100644 blob %s\tfile\n' "$blob" | git mktree) &&
+ commit=$(echo "$2-$3" | git commit-tree "$tree" ${4:+-p "$4"}) &&
+ git update-ref "refs/heads/$1" "$commit" &&
+ eval "$1"'=$(git rev-parse $1:file)' &&
+ eval "echo >&2 $1=\$$1"
+}
+
+test_expect_success 'setup commits' '
+ commit one seed 1 &&
+ commit two seed 12
+'
+
+# Note: This is heavily dependent on the "prefer larger objects as base"
+# heuristic.
+test_expect_success 'vanilla repack deltas one against two' '
+ git repack -adf &&
+ is_delta_base $one $two
+'
+
+test_expect_success 'island repack with no island definition is vanilla' '
+ git repack -adfi &&
+ is_delta_base $one $two
+'
+
+test_expect_success 'island repack with no matches is vanilla' '
+ git -c "pack.island=refs/foo" repack -adfi &&
+ is_delta_base $one $two
+'
+
+test_expect_success 'separate islands disallows delta' '
+ git -c "pack.island=refs/heads/(.*)" repack -adfi &&
+ ! is_delta_base $one $two &&
+ ! is_delta_base $two $one
+'
+
+test_expect_success 'same island allows delta' '
+ git -c "pack.island=refs/heads" repack -adfi &&
+ is_delta_base $one $two
+'
+
+test_expect_success 'coalesce same-named islands' '
+ git \
+ -c "pack.island=refs/(.*)/one" \
+ -c "pack.island=refs/(.*)/two" \
+ repack -adfi &&
+ is_delta_base $one $two
+'
+
+test_expect_success 'island restrictions drop reused deltas' '
+ git repack -adfi &&
+ is_delta_base $one $two &&
+ git -c "pack.island=refs/heads/(.*)" repack -adi &&
+ ! is_delta_base $one $two &&
+ ! is_delta_base $two $one
+'
+
+test_expect_success 'island regexes are left-anchored' '
+ git -c "pack.island=heads/(.*)" repack -adfi &&
+ is_delta_base $one $two
+'
+
+test_expect_success 'island regexes follow last-one-wins scheme' '
+ git \
+ -c "pack.island=refs/heads/(.*)" \
+ -c "pack.island=refs/heads/" \
+ repack -adfi &&
+ is_delta_base $one $two
+'
+
+test_expect_success 'setup shared history' '
+ commit root shared root &&
+ commit one shared 1 root &&
+ commit two shared 12-long root
+'
+
+# We know that $two will be preferred as a base from $one,
+# because we can transform it with a pure deletion.
+#
+# We also expect $root as a delta against $two by the "longest is base" rule.
+test_expect_success 'vanilla delta goes between branches' '
+ git repack -adf &&
+ is_delta_base $one $two &&
+ is_delta_base $root $two
+'
+
+# Here we should allow $one to base itself on $root; even though
+# they are in different islands, the objects in $root are in a superset
+# of islands compared to those in $one.
+#
+# Similarly, $two can delta against $root by our rules. And unlike $one,
+# in which we are just allowing it, the island rules actually put $root
+# as a possible base for $two, which it would not otherwise be (due to the size
+# sorting).
+test_expect_success 'deltas allowed against superset islands' '
+ git -c "pack.island=refs/heads/(.*)" repack -adfi &&
+ is_delta_base $one $root &&
+ is_delta_base $two $root
+'
+
+# We are going to test the packfile order here, so we again have to make some
+# assumptions. We assume that "$root", as part of our core "one", must come
+# before "$two". This should be guaranteed by the island code. However, for
+# this test to fail without islands, we are also assuming that it would not
+# otherwise do so. This is true by the current write order, which will put
+# commits (and their contents) before their parents.
+test_expect_success 'island core places core objects first' '
+ cat >expect <<-EOF &&
+ $root
+ $two
+ EOF
+ git -c "pack.island=refs/heads/(.*)" \
+ -c "pack.islandcore=one" \
+ repack -adfi &&
+ git verify-pack -v .git/objects/pack/*.pack |
+ cut -d" " -f1 |
+ egrep "$root|$two" >actual &&
+ test_cmp expect actual
+'
+
+test_expect_success 'unmatched island core is not fatal' '
+ git -c "pack.islandcore=one" repack -adfi
+'
+
+test_done
--
2.18.0.555.g17f9c4abba
next prev parent reply other threads:[~2018-08-09 15:56 UTC|newest]
Thread overview: 14+ messages / expand[flat|nested] mbox.gz Atom feed top
2018-08-09 15:55 [PATCH v3 0/8] Add delta islands support Christian Couder
2018-08-09 15:55 ` [PATCH v3 1/8] packfile: make get_delta_base() non static Christian Couder
2018-08-09 15:55 ` [PATCH v3 2/8] Add delta-islands.{c,h} Christian Couder
2018-08-11 9:04 ` SZEDER Gábor
2018-08-11 10:32 ` Christian Couder
2018-08-11 14:12 ` Jeff King
2018-08-12 4:32 ` Christian Couder
2018-08-09 15:55 ` [PATCH v3 3/8] pack-objects: refactor code into compute_layer_order() Christian Couder
2018-08-09 15:55 ` [PATCH v3 4/8] pack-objects: add delta-islands support Christian Couder
2018-08-09 15:55 ` [PATCH v3 5/8] repack: " Christian Couder
2018-08-09 15:55 ` Christian Couder [this message]
2018-08-09 15:55 ` [PATCH v3 7/8] pack-objects: move tree_depth into 'struct packing_data' Christian Couder
2018-08-09 15:55 ` [PATCH v3 8/8] pack-objects: move 'layer' " Christian Couder
2018-08-09 19:34 ` [PATCH v3 0/8] Add delta islands support Christian Couder
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=20180809155532.26151-7-chriscool@tuxfamily.org \
--to=christian.couder@gmail.com \
--cc=Johannes.Schindelin@gmx.de \
--cc=chriscool@tuxfamily.org \
--cc=git@vger.kernel.org \
--cc=gitster@pobox.com \
--cc=pclouds@gmail.com \
--cc=peff@peff.net \
--cc=sbeller@google.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).