* [PATCH 0/2] merge-tree: add new --mergeability-only option @ 2025-05-10 22:02 Elijah Newren via GitGitGadget 2025-05-10 22:02 ` [PATCH 1/2] merge-ort: add a new mergeability_only option Elijah Newren via GitGitGadget ` (3 more replies) 0 siblings, 4 replies; 29+ messages in thread From: Elijah Newren via GitGitGadget @ 2025-05-10 22:02 UTC (permalink / raw) To: git; +Cc: Elijah Newren This adds a new flag, --mergeability-only, to git merge-tree, which suppresses all output and leaves only the exit status (reflecting successful merge or conflict). This is useful for Git Forges in cases where they are only interested in whether two branches can be merged, without needing the actual merge result or conflict details. The advantage of the flag is two fold: * The merge machinery can exit once it detects the first conflict, instead of continuing to compute merge result information * The merge machinery can avoid writing merged blobs and trees to the object store when in the outer layer of the merging process (more details in the first commit message). Elijah Newren (2): merge-ort: add a new mergeability_only option merge-tree: add a new --mergeability-only flag Documentation/git-merge-tree.adoc | 6 +++++ builtin/merge-tree.c | 22 ++++++++++++++++++ merge-ort.c | 38 +++++++++++++++++++++++++------ merge-ort.h | 1 + t/t4301-merge-tree-write-tree.sh | 38 +++++++++++++++++++++++++++++++ 5 files changed, 98 insertions(+), 7 deletions(-) base-commit: 6c0bd1fc70efaf053abe4e57c976afdc72d15377 Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-1920%2Fnewren%2Fmergeability-only-v1 Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-1920/newren/mergeability-only-v1 Pull-Request: https://github.com/gitgitgadget/git/pull/1920 -- gitgitgadget ^ permalink raw reply [flat|nested] 29+ messages in thread
* [PATCH 1/2] merge-ort: add a new mergeability_only option 2025-05-10 22:02 [PATCH 0/2] merge-tree: add new --mergeability-only option Elijah Newren via GitGitGadget @ 2025-05-10 22:02 ` Elijah Newren via GitGitGadget 2025-05-10 22:02 ` [PATCH 2/2] merge-tree: add a new --mergeability-only flag Elijah Newren via GitGitGadget ` (2 subsequent siblings) 3 siblings, 0 replies; 29+ messages in thread From: Elijah Newren via GitGitGadget @ 2025-05-10 22:02 UTC (permalink / raw) To: git; +Cc: Elijah Newren, Elijah Newren From: Elijah Newren <newren@gmail.com> Git Forges may be interested in whether two branches can be merged while not being interested in what the resulting merge tree is nor which files conflicted. For such cases, add a new mergeability_only option. This option allows the merge machinery to, in the outer layer of the merge: * exit upon first conflict * avoid writing merged blobs/trees to the object store Note that since the recursive merge of merge bases (corresponding to call_depth > 0) can conflict without the outer final merge (corresponding to call_depth == 0) conflicting, we can't short-circuit nor avoid writing merges blobs/trees to the object store during those inner merges. There is a further potential micro-optimization here. rename/rename conflicts have the potential for nested conflicts even without recursive merges; because of that, handle_content_merge() can be called multiple times and is done via different paths. Currently, we only exit early in process_entries(), which is where the final handle_content_merge() is invoked. Since rename/rename conflicts have an additional earlier handle_content_merge() call that can be invoked from detect_and_process_renames() (via process_renames()), we could potentially exit earlier at that call point. However, rename/rename conflicts are exceptionally rare, and feeding the extra logic through didn't seem worth it. (And, if we don't exit early at that point, then any resulting blobs need to be written to the store so that subsequent handle_content_merge() calls trying to use the blob don't throw exceptions.) Signed-off-by: Elijah Newren <newren@gmail.com> --- merge-ort.c | 38 +++++++++++++++++++++++++++++++------- merge-ort.h | 1 + 2 files changed, 32 insertions(+), 7 deletions(-) diff --git a/merge-ort.c b/merge-ort.c index 77310a4a52c9..47b3d1730ece 100644 --- a/merge-ort.c +++ b/merge-ort.c @@ -2127,6 +2127,7 @@ static int handle_content_merge(struct merge_options *opt, const struct version_info *b, const char *pathnames[3], const int extra_marker_size, + const int record_object, struct version_info *result) { /* @@ -2214,7 +2215,7 @@ static int handle_content_merge(struct merge_options *opt, ret = -1; } - if (!ret && + if (!ret && record_object && write_object_file(result_buf.ptr, result_buf.size, OBJ_BLOB, &result->oid)) { path_msg(opt, ERROR_OBJECT_WRITE_FAILED, 0, @@ -2897,6 +2898,7 @@ static int process_renames(struct merge_options *opt, struct version_info merged; struct conflict_info *base, *side1, *side2; unsigned was_binary_blob = 0; + const int record_object = true; pathnames[0] = oldpath; pathnames[1] = newpath; @@ -2947,6 +2949,7 @@ static int process_renames(struct merge_options *opt, &side2->stages[2], pathnames, 1 + 2 * opt->priv->call_depth, + record_object, &merged); if (clean_merge < 0) return -1; @@ -3061,6 +3064,7 @@ static int process_renames(struct merge_options *opt, struct conflict_info *base, *side1, *side2; int clean; + const int record_object = true; pathnames[0] = oldpath; pathnames[other_source_index] = oldpath; @@ -3080,6 +3084,7 @@ static int process_renames(struct merge_options *opt, &side2->stages[2], pathnames, 1 + 2 * opt->priv->call_depth, + record_object, &merged); if (clean < 0) return -1; @@ -3931,9 +3936,12 @@ static int write_completed_directory(struct merge_options *opt, * Write out the tree to the git object directory, and also * record the mode and oid in dir_info->result. */ + int record_tree = (!opt->mergeability_only || + opt->priv->call_depth); dir_info->is_null = 0; dir_info->result.mode = S_IFDIR; - if (write_tree(&dir_info->result.oid, &info->versions, offset, + if (record_tree && + write_tree(&dir_info->result.oid, &info->versions, offset, opt->repo->hash_algo->rawsz) < 0) ret = -1; } @@ -4231,10 +4239,13 @@ static int process_entry(struct merge_options *opt, struct version_info *o = &ci->stages[0]; struct version_info *a = &ci->stages[1]; struct version_info *b = &ci->stages[2]; + int record_object = (!opt->mergeability_only || + opt->priv->call_depth); clean_merge = handle_content_merge(opt, path, o, a, b, ci->pathnames, opt->priv->call_depth * 2, + record_object, &merged_file); if (clean_merge < 0) return -1; @@ -4395,6 +4406,8 @@ static int process_entries(struct merge_options *opt, STRING_LIST_INIT_NODUP, NULL, 0 }; int ret = 0; + const int record_tree = (!opt->mergeability_only || + opt->priv->call_depth); trace2_region_enter("merge", "process_entries setup", opt->repo); if (strmap_empty(&opt->priv->paths)) { @@ -4454,6 +4467,12 @@ static int process_entries(struct merge_options *opt, ret = -1; goto cleanup; }; + if (!ci->merged.clean && opt->mergeability_only && + !opt->priv->call_depth) { + ret = 0; + goto cleanup; + } + } } trace2_region_leave("merge", "processing", opt->repo); @@ -4468,7 +4487,8 @@ static int process_entries(struct merge_options *opt, fflush(stdout); BUG("dir_metadata accounting completely off; shouldn't happen"); } - if (write_tree(result_oid, &dir_metadata.versions, 0, + if (record_tree && + write_tree(result_oid, &dir_metadata.versions, 0, opt->repo->hash_algo->rawsz) < 0) ret = -1; cleanup: @@ -4715,6 +4735,8 @@ void merge_display_update_messages(struct merge_options *opt, if (opt->record_conflict_msgs_as_headers) BUG("Either display conflict messages or record them as headers, not both"); + if (opt->mergeability_only) + BUG("Displaying conflict messages incompatible with mergeability-only checks"); trace2_region_enter("merge", "display messages", opt->repo); @@ -5171,10 +5193,12 @@ redo: result->path_messages = &opt->priv->conflicts; if (result->clean >= 0) { - result->tree = parse_tree_indirect(&working_tree_oid); - if (!result->tree) - die(_("unable to read tree (%s)"), - oid_to_hex(&working_tree_oid)); + if (!opt->mergeability_only) { + result->tree = parse_tree_indirect(&working_tree_oid); + if (!result->tree) + die(_("unable to read tree (%s)"), + oid_to_hex(&working_tree_oid)); + } /* existence of conflicted entries implies unclean */ result->clean &= strmap_empty(&opt->priv->conflicted); } diff --git a/merge-ort.h b/merge-ort.h index 30750c03962f..6045579825da 100644 --- a/merge-ort.h +++ b/merge-ort.h @@ -83,6 +83,7 @@ struct merge_options { /* miscellaneous control options */ const char *subtree_shift; unsigned renormalize : 1; + unsigned mergeability_only : 1; /* exit early, write fewer objects */ unsigned record_conflict_msgs_as_headers : 1; const char *msg_header_prefix; -- gitgitgadget ^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH 2/2] merge-tree: add a new --mergeability-only flag 2025-05-10 22:02 [PATCH 0/2] merge-tree: add new --mergeability-only option Elijah Newren via GitGitGadget 2025-05-10 22:02 ` [PATCH 1/2] merge-ort: add a new mergeability_only option Elijah Newren via GitGitGadget @ 2025-05-10 22:02 ` Elijah Newren via GitGitGadget 2025-05-12 17:04 ` [PATCH 0/2] merge-tree: add new --mergeability-only option Junio C Hamano 2025-05-12 23:42 ` [PATCH v2 0/2] merge-tree: add new --dry-run option Elijah Newren via GitGitGadget 3 siblings, 0 replies; 29+ messages in thread From: Elijah Newren via GitGitGadget @ 2025-05-10 22:02 UTC (permalink / raw) To: git; +Cc: Elijah Newren, Elijah Newren From: Elijah Newren <newren@gmail.com> Git Forges may be interested in whether two branches can be merged while not being interested in what the resulting merge tree is nor which files conflicted. For such cases, add a new --mergeability-only flag which will make use of the new mergeability_only flag added to merge-ort in the previous commit. This option allows the merge machinery to, in the outer layer of the merge: * exit upon first conflict * avoid writing merged blobs/trees to the object store Signed-off-by: Elijah Newren <newren@gmail.com> --- Documentation/git-merge-tree.adoc | 6 +++++ builtin/merge-tree.c | 22 ++++++++++++++++++ t/t4301-merge-tree-write-tree.sh | 38 +++++++++++++++++++++++++++++++ 3 files changed, 66 insertions(+) diff --git a/Documentation/git-merge-tree.adoc b/Documentation/git-merge-tree.adoc index cf0578f9b5e8..7dcc17806191 100644 --- a/Documentation/git-merge-tree.adoc +++ b/Documentation/git-merge-tree.adoc @@ -65,6 +65,12 @@ OPTIONS default is to include these messages if there are merge conflicts, and to omit them otherwise. +--mergeability-only:: + Disable all output from the program. Useful when you are only + interested in the exit status. Allows merge-tree to exit + early on the first conflict it finds, and allows it to avoid + writing most objects created by merges. + --allow-unrelated-histories:: merge-tree will by default error out if the two branches specified share no common history. This flag can be given to override that diff --git a/builtin/merge-tree.c b/builtin/merge-tree.c index 4aafa73c6155..48179d83596c 100644 --- a/builtin/merge-tree.c +++ b/builtin/merge-tree.c @@ -490,6 +490,9 @@ static int real_merge(struct merge_tree_options *o, if (result.clean < 0) die(_("failure to merge")); + if (o->merge_options.mergeability_only) + goto cleanup; + if (show_messages == -1) show_messages = !result.clean; @@ -522,6 +525,8 @@ static int real_merge(struct merge_tree_options *o, } if (o->use_stdin) putchar(line_termination); + +cleanup: merge_finalize(&opt, &result); clear_merge_options(&opt); return !result.clean; /* result.clean < 0 handled above */ @@ -538,6 +543,7 @@ int cmd_merge_tree(int argc, int original_argc; const char *merge_base = NULL; int ret; + int mergeability_only = 0; const char * const merge_tree_usage[] = { N_("git merge-tree [--write-tree] [<options>] <branch1> <branch2>"), @@ -552,6 +558,10 @@ int cmd_merge_tree(int argc, N_("do a trivial merge only"), MODE_TRIVIAL), OPT_BOOL(0, "messages", &o.show_messages, N_("also show informational/conflict messages")), + OPT_BOOL_F(0, "mergeability-only", + &mergeability_only, + N_("suppress all output; only exit status wanted"), + PARSE_OPT_NONEG), OPT_SET_INT('z', NULL, &line_termination, N_("separate paths with the NUL character"), '\0'), OPT_BOOL_F(0, "name-only", @@ -583,6 +593,18 @@ int cmd_merge_tree(int argc, argc = parse_options(argc, argv, prefix, mt_options, merge_tree_usage, PARSE_OPT_STOP_AT_NON_OPTION); + if (mergeability_only && o.show_messages == -1) + o.show_messages = 0; + o.merge_options.mergeability_only = mergeability_only; + die_for_incompatible_opt2(mergeability_only, "--mergeability-only", + o.show_messages, "--messages"); + die_for_incompatible_opt2(mergeability_only, "--mergeability-only", + o.name_only, "--name-only"); + die_for_incompatible_opt2(mergeability_only, "--mergeability-only", + o.use_stdin, "--stdin"); + die_for_incompatible_opt2(mergeability_only, "--mergeability-only", + !line_termination, "-z"); + if (xopts.nr && o.mode == MODE_TRIVIAL) die(_("--trivial-merge is incompatible with all other options")); for (size_t x = 0; x < xopts.nr; x++) diff --git a/t/t4301-merge-tree-write-tree.sh b/t/t4301-merge-tree-write-tree.sh index f9c5883a7f7c..3170830dd79d 100755 --- a/t/t4301-merge-tree-write-tree.sh +++ b/t/t4301-merge-tree-write-tree.sh @@ -54,6 +54,25 @@ test_expect_success setup ' git commit -m first-commit ' +test_expect_success '--mergeability-only on clean merge' ' + # Get rid of loose objects to start with + git gc && + echo "0 objects, 0 kilobytes" >expect && + git count-objects >actual && + test_cmp expect actual && + + # Ensure merge is successful (exit code of 0) + git merge-tree --write-tree --mergeability-only side1 side3 >output && + + # Ensure there is no output + test_must_be_empty output && + + # Ensure no loose objects written (all new objects written would have + # been in "outer layer" of the merge) + git count-objects >actual && + test_cmp expect actual +' + test_expect_success 'Clean merge' ' TREE_OID=$(git merge-tree --write-tree side1 side3) && q_to_tab <<-EOF >expect && @@ -72,6 +91,25 @@ test_expect_success 'Failed merge without rename detection' ' grep "CONFLICT (modify/delete): numbers deleted" out ' +test_expect_success '--mergeability-only on conflicted merge' ' + # Get rid of loose objects to start with + git gc && + echo "0 objects, 0 kilobytes" >expect && + git count-objects >actual && + test_cmp expect actual && + + # Ensure merge has conflict + test_expect_code 1 git merge-tree --write-tree --mergeability-only side1 side2 >output && + + # Ensure there is no output + test_must_be_empty output && + + # Ensure no loose objects written (all new objects written would have + # been in "outer layer" of the merge) + git count-objects >actual && + test_cmp expect actual +' + test_expect_success 'Content merge and a few conflicts' ' git checkout side1^0 && test_must_fail git merge side2 && -- gitgitgadget ^ permalink raw reply related [flat|nested] 29+ messages in thread
* Re: [PATCH 0/2] merge-tree: add new --mergeability-only option 2025-05-10 22:02 [PATCH 0/2] merge-tree: add new --mergeability-only option Elijah Newren via GitGitGadget 2025-05-10 22:02 ` [PATCH 1/2] merge-ort: add a new mergeability_only option Elijah Newren via GitGitGadget 2025-05-10 22:02 ` [PATCH 2/2] merge-tree: add a new --mergeability-only flag Elijah Newren via GitGitGadget @ 2025-05-12 17:04 ` Junio C Hamano 2025-05-12 17:41 ` Elijah Newren 2025-05-12 23:42 ` [PATCH v2 0/2] merge-tree: add new --dry-run option Elijah Newren via GitGitGadget 3 siblings, 1 reply; 29+ messages in thread From: Junio C Hamano @ 2025-05-12 17:04 UTC (permalink / raw) To: Elijah Newren via GitGitGadget; +Cc: git, Elijah Newren "Elijah Newren via GitGitGadget" <gitgitgadget@gmail.com> writes: > This adds a new flag, --mergeability-only, to git merge-tree, which > suppresses all output and leaves only the exit status (reflecting successful > merge or conflict). This is useful for Git Forges in cases where they are > only interested in whether two branches can be merged, without needing the > actual merge result or conflict details. Sounds useful, but wouldn't that usually called --dry-run? ^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH 0/2] merge-tree: add new --mergeability-only option 2025-05-12 17:04 ` [PATCH 0/2] merge-tree: add new --mergeability-only option Junio C Hamano @ 2025-05-12 17:41 ` Elijah Newren 2025-05-12 18:27 ` Junio C Hamano 0 siblings, 1 reply; 29+ messages in thread From: Elijah Newren @ 2025-05-12 17:41 UTC (permalink / raw) To: Junio C Hamano; +Cc: Elijah Newren via GitGitGadget, git On Mon, May 12, 2025 at 10:04 AM Junio C Hamano <gitster@pobox.com> wrote: > > "Elijah Newren via GitGitGadget" <gitgitgadget@gmail.com> writes: > > > This adds a new flag, --mergeability-only, to git merge-tree, which > > suppresses all output and leaves only the exit status (reflecting successful > > merge or conflict). This is useful for Git Forges in cases where they are > > only interested in whether two branches can be merged, without needing the > > actual merge result or conflict details. > > Sounds useful, but wouldn't that usually called --dry-run? I thought about that, but I was worried that folks would expect "--dry-run" to not make any changes. This mode does not prevent writing objects to the object store, it merely avoids it in the "outer layer" of the merge. More precisely, objects will still be written to the object store for the merging of merge bases, and also be written to the object store in the case of rename/rename conflicts if the contents of the files involved in the conflicting renames were also modified by both sides. ^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH 0/2] merge-tree: add new --mergeability-only option 2025-05-12 17:41 ` Elijah Newren @ 2025-05-12 18:27 ` Junio C Hamano 2025-05-12 18:37 ` Elijah Newren 0 siblings, 1 reply; 29+ messages in thread From: Junio C Hamano @ 2025-05-12 18:27 UTC (permalink / raw) To: Elijah Newren; +Cc: Elijah Newren via GitGitGadget, git Elijah Newren <newren@gmail.com> writes: > I thought about that, but I was worried that folks would expect > "--dry-run" to not make any changes. This mode does not prevent > writing objects to the object store, it merely avoids it in the "outer > layer" of the merge. I think we have already precedence to call something that creates new objects in the object database, as long as the resulting objects are not made reachable ("git fetch --dry-run" probably falls into that category). The idea is that it does not make a change that is "observable" by end-users (and what "gc" sees is not part of what the users would be observaing). We have "--check" (in "git apply"), which is an exact counterpart in the patch based workflow to this thing. It reads Instead of applying the patch, see if the patch is applicable to the current working tree and/or the index file and detects errors. Turns off "apply". I feel that `apply --check` should have been `apply --dry-run`, so I would not recommend calling it `--check` for `merge-tree`, though. Thanks. ^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH 0/2] merge-tree: add new --mergeability-only option 2025-05-12 18:27 ` Junio C Hamano @ 2025-05-12 18:37 ` Elijah Newren 0 siblings, 0 replies; 29+ messages in thread From: Elijah Newren @ 2025-05-12 18:37 UTC (permalink / raw) To: Junio C Hamano; +Cc: Elijah Newren via GitGitGadget, git On Mon, May 12, 2025 at 11:27 AM Junio C Hamano <gitster@pobox.com> wrote: > > Elijah Newren <newren@gmail.com> writes: > > > I thought about that, but I was worried that folks would expect > > "--dry-run" to not make any changes. This mode does not prevent > > writing objects to the object store, it merely avoids it in the "outer > > layer" of the merge. > > I think we have already precedence to call something that creates > new objects in the object database, as long as the resulting objects > are not made reachable ("git fetch --dry-run" probably falls into > that category). The idea is that it does not make a change that is > "observable" by end-users (and what "gc" sees is not part of what > the users would be observaing). Oh, I was unaware of `git fetch --dry-run` for some reason. And its documentation even states "without making any changes" despite the fact that it downloads more objects to the object store, so it indeed sounds like a good precedent. I'll switch the flag name to --dry-run. (I have a suspicion, however, that the primary users of this new merge-tree flag will care about whether objects are created, so I still want the documentation to call it out, unlike git fetch's --dry-run option.) > We have "--check" (in "git apply"), which is an exact counterpart in > the patch based workflow to this thing. It reads > > Instead of applying the patch, see if the patch is > applicable to the current working tree and/or the index > file and detects errors. Turns off "apply". > > I feel that `apply --check` should have been `apply --dry-run`, so I > would not recommend calling it `--check` for `merge-tree`, though. Makes sense; thanks for the pointers. ^ permalink raw reply [flat|nested] 29+ messages in thread
* [PATCH v2 0/2] merge-tree: add new --dry-run option 2025-05-10 22:02 [PATCH 0/2] merge-tree: add new --mergeability-only option Elijah Newren via GitGitGadget ` (2 preceding siblings ...) 2025-05-12 17:04 ` [PATCH 0/2] merge-tree: add new --mergeability-only option Junio C Hamano @ 2025-05-12 23:42 ` Elijah Newren via GitGitGadget 2025-05-12 23:42 ` [PATCH v2 1/2] merge-ort: add a new mergeability_only option Elijah Newren via GitGitGadget ` (2 more replies) 3 siblings, 3 replies; 29+ messages in thread From: Elijah Newren via GitGitGadget @ 2025-05-12 23:42 UTC (permalink / raw) To: git; +Cc: Elijah Newren, Elijah Newren Changes since v1: * Renamed --mergeability-only flag to --dry-run, as per suggestion from Junio * added some commit message clarifications This adds a new flag, --dry-run, to git merge-tree, which suppresses all output and leaves only the exit status (reflecting successful merge or conflict). This is useful for Git Forges in cases where they are only interested in whether two branches can be merged, without needing the actual merge result or conflict details. The advantage of the flag is two fold: * The merge machinery can exit once it detects a conflict, instead of continuing to compute merge result information * The merge machinery can avoid writing merged blobs and trees to the object store when in the outer layer of the merging process (more details in the first commit message). Elijah Newren (2): merge-ort: add a new mergeability_only option merge-tree: add a new --dry-run flag Documentation/git-merge-tree.adoc | 6 +++++ builtin/merge-tree.c | 22 ++++++++++++++++++ merge-ort.c | 38 +++++++++++++++++++++++++------ merge-ort.h | 1 + t/t4301-merge-tree-write-tree.sh | 38 +++++++++++++++++++++++++++++++ 5 files changed, 98 insertions(+), 7 deletions(-) base-commit: 6c0bd1fc70efaf053abe4e57c976afdc72d15377 Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-1920%2Fnewren%2Fmergeability-only-v2 Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-1920/newren/mergeability-only-v2 Pull-Request: https://github.com/gitgitgadget/git/pull/1920 Range-diff vs v1: 1: 09292804cff ! 1: 4757c4810d3 merge-ort: add a new mergeability_only option @@ Commit message Git Forges may be interested in whether two branches can be merged while not being interested in what the resulting merge tree is nor which files conflicted. For such cases, add a new mergeability_only option. This - option allows the merge machinery to, in the outer layer of the merge: - * exit upon first conflict - * avoid writing merged blobs/trees to the object store + option allows the merge machinery to, in the "outer layer" of the merge: + * exit upon first[-ish] conflict + * avoid (not prevent) writing merged blobs/trees to the object store + + I have a number of qualifiers there, so let me explain each: + + "outer layer": Note that since the recursive merge of merge bases (corresponding to - call_depth > 0) can conflict without the outer final merge (corresponding - to call_depth == 0) conflicting, we can't short-circuit nor avoid - writing merges blobs/trees to the object store during those inner - merges. + call_depth > 0) can conflict without the outer final merge + (corresponding to call_depth == 0) conflicting, we can't short-circuit + nor avoid writing merged blobs/trees to the object store during those + inner merges. + + "first-ish conflict": + + The current patch only exits early from process_entries() on the first + conflict it detects, but conflicts could have been detected in a + previous function call, namely detect_and_process_renames(). However: + * conflicts detected by detect_and_process_renames() are quite rare + conflict types + * the detection would still come after regular rename detection + (which is the expensive part of detect_and_process_renames()), so + it is not saving us much in computation time given that + process_entries() directly follows detect_and_process_renames() + * [this overlaps with the next bullet point] process_entries() is the + place where virtually all object writing occurs (object writing is + sometimes more of a concern for Forges than computation time), so + exiting early here isn't saving us much in object writes either + * the code changes needed to handle an earlier exit are slightly + more invasive in detect_and_process_renames() than for + process_entries(). + Given the rareness of the even earlier conflicts, the limited savings + we'd get from exiting even earlier, and in an attempt to keep this + patch simpler, we don't guarantee that we actually exit on the first + conflict detected. We can always revisit this decision later if we + decide that a further micro-optimization to exit slightly earlier in + rare cases is worthwhile. + + "avoid (not prevent) writing objects": - There is a further potential micro-optimization here. rename/rename - conflicts have the potential for nested conflicts even without recursive - merges; because of that, handle_content_merge() can be called multiple - times and is done via different paths. Currently, we only exit early in - process_entries(), which is where the final handle_content_merge() is - invoked. Since rename/rename conflicts have an additional earlier - handle_content_merge() call that can be invoked from - detect_and_process_renames() (via process_renames()), we could - potentially exit earlier at that call point. However, rename/rename - conflicts are exceptionally rare, and feeding the extra logic through - didn't seem worth it. (And, if we don't exit early at that point, then - any resulting blobs need to be written to the store so that subsequent - handle_content_merge() calls trying to use the blob don't throw - exceptions.) + The detect_and_process_renames() call can also write objects to the + object store, when rename/rename conflicts involve one (or more) files + that have also been modified on both sides. Because of this alternate + call path leading to handle_content_merges(), our "early exit" does not + prevent writing objects entirely, even within the "outer layer" + (i.e. even within call_depth == 0). I figure that's fine though, since + we're already writing objects for the inner merges (i.e. for call_depth + > 0), which are likely going to represent vastly more objects than files + involved in rename/rename+modify/modify cases in the outer merge, on + average. Signed-off-by: Elijah Newren <newren@gmail.com> 2: 0b32c80d286 ! 2: 1d18ab7feb8 merge-tree: add a new --mergeability-only flag @@ Metadata Author: Elijah Newren <newren@gmail.com> ## Commit message ## - merge-tree: add a new --mergeability-only flag + merge-tree: add a new --dry-run flag Git Forges may be interested in whether two branches can be merged while not being interested in what the resulting merge tree is nor which files - conflicted. For such cases, add a new --mergeability-only flag which + conflicted. For such cases, add a new --dry-run flag which will make use of the new mergeability_only flag added to merge-ort in the previous commit. This option allows the merge machinery to, in the outer layer of the merge: - * exit upon first conflict - * avoid writing merged blobs/trees to the object store + * exit early when a conflict is detected + * avoid writing (most) merged blobs/trees to the object store Signed-off-by: Elijah Newren <newren@gmail.com> @@ builtin/merge-tree.c: int cmd_merge_tree(int argc, int original_argc; const char *merge_base = NULL; int ret; -+ int mergeability_only = 0; ++ int dry_run = 0; const char * const merge_tree_usage[] = { N_("git merge-tree [--write-tree] [<options>] <branch1> <branch2>"), @@ builtin/merge-tree.c: int cmd_merge_tree(int argc, N_("do a trivial merge only"), MODE_TRIVIAL), OPT_BOOL(0, "messages", &o.show_messages, N_("also show informational/conflict messages")), -+ OPT_BOOL_F(0, "mergeability-only", -+ &mergeability_only, ++ OPT_BOOL_F(0, "dry-run", ++ &dry_run, + N_("suppress all output; only exit status wanted"), + PARSE_OPT_NONEG), OPT_SET_INT('z', NULL, &line_termination, @@ builtin/merge-tree.c: int cmd_merge_tree(int argc, argc = parse_options(argc, argv, prefix, mt_options, merge_tree_usage, PARSE_OPT_STOP_AT_NON_OPTION); -+ if (mergeability_only && o.show_messages == -1) ++ if (dry_run && o.show_messages == -1) + o.show_messages = 0; -+ o.merge_options.mergeability_only = mergeability_only; -+ die_for_incompatible_opt2(mergeability_only, "--mergeability-only", ++ o.merge_options.mergeability_only = dry_run; ++ die_for_incompatible_opt2(dry_run, "--mergeability-only", + o.show_messages, "--messages"); -+ die_for_incompatible_opt2(mergeability_only, "--mergeability-only", ++ die_for_incompatible_opt2(dry_run, "--mergeability-only", + o.name_only, "--name-only"); -+ die_for_incompatible_opt2(mergeability_only, "--mergeability-only", ++ die_for_incompatible_opt2(dry_run, "--mergeability-only", + o.use_stdin, "--stdin"); -+ die_for_incompatible_opt2(mergeability_only, "--mergeability-only", ++ die_for_incompatible_opt2(dry_run, "--mergeability-only", + !line_termination, "-z"); + if (xopts.nr && o.mode == MODE_TRIVIAL) @@ t/t4301-merge-tree-write-tree.sh: test_expect_success setup ' git commit -m first-commit ' -+test_expect_success '--mergeability-only on clean merge' ' ++test_expect_success '--dry-run on clean merge' ' + # Get rid of loose objects to start with + git gc && + echo "0 objects, 0 kilobytes" >expect && @@ t/t4301-merge-tree-write-tree.sh: test_expect_success setup ' + test_cmp expect actual && + + # Ensure merge is successful (exit code of 0) -+ git merge-tree --write-tree --mergeability-only side1 side3 >output && ++ git merge-tree --write-tree --dry-run side1 side3 >output && + + # Ensure there is no output + test_must_be_empty output && @@ t/t4301-merge-tree-write-tree.sh: test_expect_success 'Failed merge without rena grep "CONFLICT (modify/delete): numbers deleted" out ' -+test_expect_success '--mergeability-only on conflicted merge' ' ++test_expect_success '--dry-run on conflicted merge' ' + # Get rid of loose objects to start with + git gc && + echo "0 objects, 0 kilobytes" >expect && @@ t/t4301-merge-tree-write-tree.sh: test_expect_success 'Failed merge without rena + test_cmp expect actual && + + # Ensure merge has conflict -+ test_expect_code 1 git merge-tree --write-tree --mergeability-only side1 side2 >output && ++ test_expect_code 1 git merge-tree --write-tree --dry-run side1 side2 >output && + + # Ensure there is no output + test_must_be_empty output && -- gitgitgadget ^ permalink raw reply [flat|nested] 29+ messages in thread
* [PATCH v2 1/2] merge-ort: add a new mergeability_only option 2025-05-12 23:42 ` [PATCH v2 0/2] merge-tree: add new --dry-run option Elijah Newren via GitGitGadget @ 2025-05-12 23:42 ` Elijah Newren via GitGitGadget 2025-05-12 23:42 ` [PATCH v2 2/2] merge-tree: add a new --dry-run flag Elijah Newren via GitGitGadget 2025-05-14 0:24 ` [PATCH v3 0/2] merge-tree: add new --dry-run option Elijah Newren via GitGitGadget 2 siblings, 0 replies; 29+ messages in thread From: Elijah Newren via GitGitGadget @ 2025-05-12 23:42 UTC (permalink / raw) To: git; +Cc: Elijah Newren, Elijah Newren, Elijah Newren From: Elijah Newren <newren@gmail.com> Git Forges may be interested in whether two branches can be merged while not being interested in what the resulting merge tree is nor which files conflicted. For such cases, add a new mergeability_only option. This option allows the merge machinery to, in the "outer layer" of the merge: * exit upon first[-ish] conflict * avoid (not prevent) writing merged blobs/trees to the object store I have a number of qualifiers there, so let me explain each: "outer layer": Note that since the recursive merge of merge bases (corresponding to call_depth > 0) can conflict without the outer final merge (corresponding to call_depth == 0) conflicting, we can't short-circuit nor avoid writing merged blobs/trees to the object store during those inner merges. "first-ish conflict": The current patch only exits early from process_entries() on the first conflict it detects, but conflicts could have been detected in a previous function call, namely detect_and_process_renames(). However: * conflicts detected by detect_and_process_renames() are quite rare conflict types * the detection would still come after regular rename detection (which is the expensive part of detect_and_process_renames()), so it is not saving us much in computation time given that process_entries() directly follows detect_and_process_renames() * [this overlaps with the next bullet point] process_entries() is the place where virtually all object writing occurs (object writing is sometimes more of a concern for Forges than computation time), so exiting early here isn't saving us much in object writes either * the code changes needed to handle an earlier exit are slightly more invasive in detect_and_process_renames() than for process_entries(). Given the rareness of the even earlier conflicts, the limited savings we'd get from exiting even earlier, and in an attempt to keep this patch simpler, we don't guarantee that we actually exit on the first conflict detected. We can always revisit this decision later if we decide that a further micro-optimization to exit slightly earlier in rare cases is worthwhile. "avoid (not prevent) writing objects": The detect_and_process_renames() call can also write objects to the object store, when rename/rename conflicts involve one (or more) files that have also been modified on both sides. Because of this alternate call path leading to handle_content_merges(), our "early exit" does not prevent writing objects entirely, even within the "outer layer" (i.e. even within call_depth == 0). I figure that's fine though, since we're already writing objects for the inner merges (i.e. for call_depth > 0), which are likely going to represent vastly more objects than files involved in rename/rename+modify/modify cases in the outer merge, on average. Signed-off-by: Elijah Newren <newren@gmail.com> --- merge-ort.c | 38 +++++++++++++++++++++++++++++++------- merge-ort.h | 1 + 2 files changed, 32 insertions(+), 7 deletions(-) diff --git a/merge-ort.c b/merge-ort.c index 77310a4a52c9..47b3d1730ece 100644 --- a/merge-ort.c +++ b/merge-ort.c @@ -2127,6 +2127,7 @@ static int handle_content_merge(struct merge_options *opt, const struct version_info *b, const char *pathnames[3], const int extra_marker_size, + const int record_object, struct version_info *result) { /* @@ -2214,7 +2215,7 @@ static int handle_content_merge(struct merge_options *opt, ret = -1; } - if (!ret && + if (!ret && record_object && write_object_file(result_buf.ptr, result_buf.size, OBJ_BLOB, &result->oid)) { path_msg(opt, ERROR_OBJECT_WRITE_FAILED, 0, @@ -2897,6 +2898,7 @@ static int process_renames(struct merge_options *opt, struct version_info merged; struct conflict_info *base, *side1, *side2; unsigned was_binary_blob = 0; + const int record_object = true; pathnames[0] = oldpath; pathnames[1] = newpath; @@ -2947,6 +2949,7 @@ static int process_renames(struct merge_options *opt, &side2->stages[2], pathnames, 1 + 2 * opt->priv->call_depth, + record_object, &merged); if (clean_merge < 0) return -1; @@ -3061,6 +3064,7 @@ static int process_renames(struct merge_options *opt, struct conflict_info *base, *side1, *side2; int clean; + const int record_object = true; pathnames[0] = oldpath; pathnames[other_source_index] = oldpath; @@ -3080,6 +3084,7 @@ static int process_renames(struct merge_options *opt, &side2->stages[2], pathnames, 1 + 2 * opt->priv->call_depth, + record_object, &merged); if (clean < 0) return -1; @@ -3931,9 +3936,12 @@ static int write_completed_directory(struct merge_options *opt, * Write out the tree to the git object directory, and also * record the mode and oid in dir_info->result. */ + int record_tree = (!opt->mergeability_only || + opt->priv->call_depth); dir_info->is_null = 0; dir_info->result.mode = S_IFDIR; - if (write_tree(&dir_info->result.oid, &info->versions, offset, + if (record_tree && + write_tree(&dir_info->result.oid, &info->versions, offset, opt->repo->hash_algo->rawsz) < 0) ret = -1; } @@ -4231,10 +4239,13 @@ static int process_entry(struct merge_options *opt, struct version_info *o = &ci->stages[0]; struct version_info *a = &ci->stages[1]; struct version_info *b = &ci->stages[2]; + int record_object = (!opt->mergeability_only || + opt->priv->call_depth); clean_merge = handle_content_merge(opt, path, o, a, b, ci->pathnames, opt->priv->call_depth * 2, + record_object, &merged_file); if (clean_merge < 0) return -1; @@ -4395,6 +4406,8 @@ static int process_entries(struct merge_options *opt, STRING_LIST_INIT_NODUP, NULL, 0 }; int ret = 0; + const int record_tree = (!opt->mergeability_only || + opt->priv->call_depth); trace2_region_enter("merge", "process_entries setup", opt->repo); if (strmap_empty(&opt->priv->paths)) { @@ -4454,6 +4467,12 @@ static int process_entries(struct merge_options *opt, ret = -1; goto cleanup; }; + if (!ci->merged.clean && opt->mergeability_only && + !opt->priv->call_depth) { + ret = 0; + goto cleanup; + } + } } trace2_region_leave("merge", "processing", opt->repo); @@ -4468,7 +4487,8 @@ static int process_entries(struct merge_options *opt, fflush(stdout); BUG("dir_metadata accounting completely off; shouldn't happen"); } - if (write_tree(result_oid, &dir_metadata.versions, 0, + if (record_tree && + write_tree(result_oid, &dir_metadata.versions, 0, opt->repo->hash_algo->rawsz) < 0) ret = -1; cleanup: @@ -4715,6 +4735,8 @@ void merge_display_update_messages(struct merge_options *opt, if (opt->record_conflict_msgs_as_headers) BUG("Either display conflict messages or record them as headers, not both"); + if (opt->mergeability_only) + BUG("Displaying conflict messages incompatible with mergeability-only checks"); trace2_region_enter("merge", "display messages", opt->repo); @@ -5171,10 +5193,12 @@ redo: result->path_messages = &opt->priv->conflicts; if (result->clean >= 0) { - result->tree = parse_tree_indirect(&working_tree_oid); - if (!result->tree) - die(_("unable to read tree (%s)"), - oid_to_hex(&working_tree_oid)); + if (!opt->mergeability_only) { + result->tree = parse_tree_indirect(&working_tree_oid); + if (!result->tree) + die(_("unable to read tree (%s)"), + oid_to_hex(&working_tree_oid)); + } /* existence of conflicted entries implies unclean */ result->clean &= strmap_empty(&opt->priv->conflicted); } diff --git a/merge-ort.h b/merge-ort.h index 30750c03962f..6045579825da 100644 --- a/merge-ort.h +++ b/merge-ort.h @@ -83,6 +83,7 @@ struct merge_options { /* miscellaneous control options */ const char *subtree_shift; unsigned renormalize : 1; + unsigned mergeability_only : 1; /* exit early, write fewer objects */ unsigned record_conflict_msgs_as_headers : 1; const char *msg_header_prefix; -- gitgitgadget ^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH v2 2/2] merge-tree: add a new --dry-run flag 2025-05-12 23:42 ` [PATCH v2 0/2] merge-tree: add new --dry-run option Elijah Newren via GitGitGadget 2025-05-12 23:42 ` [PATCH v2 1/2] merge-ort: add a new mergeability_only option Elijah Newren via GitGitGadget @ 2025-05-12 23:42 ` Elijah Newren via GitGitGadget 2025-05-13 7:15 ` Kristoffer Haugsbakk 2025-05-13 13:24 ` Junio C Hamano 2025-05-14 0:24 ` [PATCH v3 0/2] merge-tree: add new --dry-run option Elijah Newren via GitGitGadget 2 siblings, 2 replies; 29+ messages in thread From: Elijah Newren via GitGitGadget @ 2025-05-12 23:42 UTC (permalink / raw) To: git; +Cc: Elijah Newren, Elijah Newren, Elijah Newren From: Elijah Newren <newren@gmail.com> Git Forges may be interested in whether two branches can be merged while not being interested in what the resulting merge tree is nor which files conflicted. For such cases, add a new --dry-run flag which will make use of the new mergeability_only flag added to merge-ort in the previous commit. This option allows the merge machinery to, in the outer layer of the merge: * exit early when a conflict is detected * avoid writing (most) merged blobs/trees to the object store Signed-off-by: Elijah Newren <newren@gmail.com> --- Documentation/git-merge-tree.adoc | 6 +++++ builtin/merge-tree.c | 22 ++++++++++++++++++ t/t4301-merge-tree-write-tree.sh | 38 +++++++++++++++++++++++++++++++ 3 files changed, 66 insertions(+) diff --git a/Documentation/git-merge-tree.adoc b/Documentation/git-merge-tree.adoc index cf0578f9b5e8..7dcc17806191 100644 --- a/Documentation/git-merge-tree.adoc +++ b/Documentation/git-merge-tree.adoc @@ -65,6 +65,12 @@ OPTIONS default is to include these messages if there are merge conflicts, and to omit them otherwise. +--mergeability-only:: + Disable all output from the program. Useful when you are only + interested in the exit status. Allows merge-tree to exit + early on the first conflict it finds, and allows it to avoid + writing most objects created by merges. + --allow-unrelated-histories:: merge-tree will by default error out if the two branches specified share no common history. This flag can be given to override that diff --git a/builtin/merge-tree.c b/builtin/merge-tree.c index 4aafa73c6155..579e81d51844 100644 --- a/builtin/merge-tree.c +++ b/builtin/merge-tree.c @@ -490,6 +490,9 @@ static int real_merge(struct merge_tree_options *o, if (result.clean < 0) die(_("failure to merge")); + if (o->merge_options.mergeability_only) + goto cleanup; + if (show_messages == -1) show_messages = !result.clean; @@ -522,6 +525,8 @@ static int real_merge(struct merge_tree_options *o, } if (o->use_stdin) putchar(line_termination); + +cleanup: merge_finalize(&opt, &result); clear_merge_options(&opt); return !result.clean; /* result.clean < 0 handled above */ @@ -538,6 +543,7 @@ int cmd_merge_tree(int argc, int original_argc; const char *merge_base = NULL; int ret; + int dry_run = 0; const char * const merge_tree_usage[] = { N_("git merge-tree [--write-tree] [<options>] <branch1> <branch2>"), @@ -552,6 +558,10 @@ int cmd_merge_tree(int argc, N_("do a trivial merge only"), MODE_TRIVIAL), OPT_BOOL(0, "messages", &o.show_messages, N_("also show informational/conflict messages")), + OPT_BOOL_F(0, "dry-run", + &dry_run, + N_("suppress all output; only exit status wanted"), + PARSE_OPT_NONEG), OPT_SET_INT('z', NULL, &line_termination, N_("separate paths with the NUL character"), '\0'), OPT_BOOL_F(0, "name-only", @@ -583,6 +593,18 @@ int cmd_merge_tree(int argc, argc = parse_options(argc, argv, prefix, mt_options, merge_tree_usage, PARSE_OPT_STOP_AT_NON_OPTION); + if (dry_run && o.show_messages == -1) + o.show_messages = 0; + o.merge_options.mergeability_only = dry_run; + die_for_incompatible_opt2(dry_run, "--mergeability-only", + o.show_messages, "--messages"); + die_for_incompatible_opt2(dry_run, "--mergeability-only", + o.name_only, "--name-only"); + die_for_incompatible_opt2(dry_run, "--mergeability-only", + o.use_stdin, "--stdin"); + die_for_incompatible_opt2(dry_run, "--mergeability-only", + !line_termination, "-z"); + if (xopts.nr && o.mode == MODE_TRIVIAL) die(_("--trivial-merge is incompatible with all other options")); for (size_t x = 0; x < xopts.nr; x++) diff --git a/t/t4301-merge-tree-write-tree.sh b/t/t4301-merge-tree-write-tree.sh index f9c5883a7f7c..566a2b4ec737 100755 --- a/t/t4301-merge-tree-write-tree.sh +++ b/t/t4301-merge-tree-write-tree.sh @@ -54,6 +54,25 @@ test_expect_success setup ' git commit -m first-commit ' +test_expect_success '--dry-run on clean merge' ' + # Get rid of loose objects to start with + git gc && + echo "0 objects, 0 kilobytes" >expect && + git count-objects >actual && + test_cmp expect actual && + + # Ensure merge is successful (exit code of 0) + git merge-tree --write-tree --dry-run side1 side3 >output && + + # Ensure there is no output + test_must_be_empty output && + + # Ensure no loose objects written (all new objects written would have + # been in "outer layer" of the merge) + git count-objects >actual && + test_cmp expect actual +' + test_expect_success 'Clean merge' ' TREE_OID=$(git merge-tree --write-tree side1 side3) && q_to_tab <<-EOF >expect && @@ -72,6 +91,25 @@ test_expect_success 'Failed merge without rename detection' ' grep "CONFLICT (modify/delete): numbers deleted" out ' +test_expect_success '--dry-run on conflicted merge' ' + # Get rid of loose objects to start with + git gc && + echo "0 objects, 0 kilobytes" >expect && + git count-objects >actual && + test_cmp expect actual && + + # Ensure merge has conflict + test_expect_code 1 git merge-tree --write-tree --dry-run side1 side2 >output && + + # Ensure there is no output + test_must_be_empty output && + + # Ensure no loose objects written (all new objects written would have + # been in "outer layer" of the merge) + git count-objects >actual && + test_cmp expect actual +' + test_expect_success 'Content merge and a few conflicts' ' git checkout side1^0 && test_must_fail git merge side2 && -- gitgitgadget ^ permalink raw reply related [flat|nested] 29+ messages in thread
* Re: [PATCH v2 2/2] merge-tree: add a new --dry-run flag 2025-05-12 23:42 ` [PATCH v2 2/2] merge-tree: add a new --dry-run flag Elijah Newren via GitGitGadget @ 2025-05-13 7:15 ` Kristoffer Haugsbakk 2025-05-13 15:28 ` Elijah Newren 2025-05-13 13:24 ` Junio C Hamano 1 sibling, 1 reply; 29+ messages in thread From: Kristoffer Haugsbakk @ 2025-05-13 7:15 UTC (permalink / raw) To: Josh Soref, git; +Cc: Elijah Newren On Tue, May 13, 2025, at 01:42, Elijah Newren via GitGitGadget wrote: > From: Elijah Newren <newren@gmail.com> > +--mergeability-only:: > + Disable all output from the program. Useful when you are only > + interested in the exit status. Allows merge-tree to exit > + early on the first conflict it finds, and allows it to avoid > + writing most objects created by merges. The previous name is being used here. -- Kristoffer Haugsbakk ^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH v2 2/2] merge-tree: add a new --dry-run flag 2025-05-13 7:15 ` Kristoffer Haugsbakk @ 2025-05-13 15:28 ` Elijah Newren 0 siblings, 0 replies; 29+ messages in thread From: Elijah Newren @ 2025-05-13 15:28 UTC (permalink / raw) To: Kristoffer Haugsbakk; +Cc: Josh Soref, git On Tue, May 13, 2025 at 12:15 AM Kristoffer Haugsbakk <kristofferhaugsbakk@fastmail.com> wrote: > > On Tue, May 13, 2025, at 01:42, Elijah Newren via GitGitGadget wrote: > > From: Elijah Newren <newren@gmail.com> > > +--mergeability-only:: > > + Disable all output from the program. Useful when you are only > > + interested in the exit status. Allows merge-tree to exit > > + early on the first conflict it finds, and allows it to avoid > > + writing most objects created by merges. > > The previous name is being used here. Oops; thanks for catching that. I'll fix it up. ^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH v2 2/2] merge-tree: add a new --dry-run flag 2025-05-12 23:42 ` [PATCH v2 2/2] merge-tree: add a new --dry-run flag Elijah Newren via GitGitGadget 2025-05-13 7:15 ` Kristoffer Haugsbakk @ 2025-05-13 13:24 ` Junio C Hamano 2025-05-13 15:30 ` Elijah Newren 1 sibling, 1 reply; 29+ messages in thread From: Junio C Hamano @ 2025-05-13 13:24 UTC (permalink / raw) To: Elijah Newren via GitGitGadget; +Cc: git, Elijah Newren "Elijah Newren via GitGitGadget" <gitgitgadget@gmail.com> writes: > From: Elijah Newren <newren@gmail.com> > > Git Forges may be interested in whether two branches can be merged while > not being interested in what the resulting merge tree is nor which files > conflicted. For such cases, add a new --dry-run flag which > will make use of the new mergeability_only flag added to merge-ort in > the previous commit. This option allows the merge machinery to, in the The first three lines are almost identical to [1/2] here, modulo the internal name still being mergeability-only while the external name is now dry-run, which is perfectly fine as long as that is done consistently. > diff --git a/Documentation/git-merge-tree.adoc b/Documentation/git-merge-tree.adoc > index cf0578f9b5e8..7dcc17806191 100644 > --- a/Documentation/git-merge-tree.adoc > +++ b/Documentation/git-merge-tree.adoc > @@ -65,6 +65,12 @@ OPTIONS > default is to include these messages if there are merge > conflicts, and to omit them otherwise. > > +--mergeability-only:: But is this internal or external name? > + Disable all output from the program. Useful when you are only > + interested in the exit status. Allows merge-tree to exit > + early on the first conflict it finds, and allows it to avoid > + writing most objects created by merges. OK. ^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH v2 2/2] merge-tree: add a new --dry-run flag 2025-05-13 13:24 ` Junio C Hamano @ 2025-05-13 15:30 ` Elijah Newren 2025-05-14 14:08 ` Junio C Hamano 0 siblings, 1 reply; 29+ messages in thread From: Elijah Newren @ 2025-05-13 15:30 UTC (permalink / raw) To: Junio C Hamano; +Cc: Elijah Newren via GitGitGadget, git On Tue, May 13, 2025 at 6:24 AM Junio C Hamano <gitster@pobox.com> wrote: > > "Elijah Newren via GitGitGadget" <gitgitgadget@gmail.com> writes: > > > From: Elijah Newren <newren@gmail.com> > > > > Git Forges may be interested in whether two branches can be merged while > > not being interested in what the resulting merge tree is nor which files > > conflicted. For such cases, add a new --dry-run flag which > > will make use of the new mergeability_only flag added to merge-ort in > > the previous commit. This option allows the merge machinery to, in the > > The first three lines are almost identical to [1/2] here, modulo the > internal name still being mergeability-only while the external name > is now dry-run, which is perfectly fine as long as that is done > consistently. > > > diff --git a/Documentation/git-merge-tree.adoc b/Documentation/git-merge-tree.adoc > > index cf0578f9b5e8..7dcc17806191 100644 > > --- a/Documentation/git-merge-tree.adoc > > +++ b/Documentation/git-merge-tree.adoc > > @@ -65,6 +65,12 @@ OPTIONS > > default is to include these messages if there are merge > > conflicts, and to omit them otherwise. > > > > +--mergeability-only:: > > But is this internal or external name? external; Kristoffer caught this oversight too. I'll send a re-roll with this fixed. (I apparently also missed it in the die-if-incompatible-options too, not sure how I missed all of these, but I'll ensure they're all fixed up). ^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH v2 2/2] merge-tree: add a new --dry-run flag 2025-05-13 15:30 ` Elijah Newren @ 2025-05-14 14:08 ` Junio C Hamano 0 siblings, 0 replies; 29+ messages in thread From: Junio C Hamano @ 2025-05-14 14:08 UTC (permalink / raw) To: Elijah Newren; +Cc: Elijah Newren via GitGitGadget, git Elijah Newren <newren@gmail.com> writes: > external; Kristoffer caught this oversight too. I'll send a re-roll > with this fixed. (I apparently also missed it in the > die-if-incompatible-options too, not sure how I missed all of these, > but I'll ensure they're all fixed up). Thanks. Will hold. ^ permalink raw reply [flat|nested] 29+ messages in thread
* [PATCH v3 0/2] merge-tree: add new --dry-run option 2025-05-12 23:42 ` [PATCH v2 0/2] merge-tree: add new --dry-run option Elijah Newren via GitGitGadget 2025-05-12 23:42 ` [PATCH v2 1/2] merge-ort: add a new mergeability_only option Elijah Newren via GitGitGadget 2025-05-12 23:42 ` [PATCH v2 2/2] merge-tree: add a new --dry-run flag Elijah Newren via GitGitGadget @ 2025-05-14 0:24 ` Elijah Newren via GitGitGadget 2025-05-14 0:24 ` [PATCH v3 1/2] merge-ort: add a new mergeability_only option Elijah Newren via GitGitGadget ` (3 more replies) 2 siblings, 4 replies; 29+ messages in thread From: Elijah Newren via GitGitGadget @ 2025-05-14 0:24 UTC (permalink / raw) To: git; +Cc: Elijah Newren, Kristoffer Haugsbakk, Elijah Newren Changes since v2: * Converted locations missed in v1 in changing --mergeability-only -> --dry-run Changes since v1: * Renamed --mergeability-only flag to --dry-run, as per suggestion from Junio * added some commit message clarifications This adds a new flag, --dry-run, to git merge-tree, which suppresses all output and leaves only the exit status (reflecting successful merge or conflict). This is useful for Git Forges in cases where they are only interested in whether two branches can be merged, without needing the actual merge result or conflict details. The advantage of the flag is two fold: * The merge machinery can exit once it detects a conflict, instead of continuing to compute merge result information * The merge machinery can avoid writing merged blobs and trees to the object store when in the outer layer of the merging process (more details in the first commit message). Elijah Newren (2): merge-ort: add a new mergeability_only option merge-tree: add a new --dry-run flag Documentation/git-merge-tree.adoc | 6 +++++ builtin/merge-tree.c | 22 ++++++++++++++++++ merge-ort.c | 38 +++++++++++++++++++++++++------ merge-ort.h | 1 + t/t4301-merge-tree-write-tree.sh | 38 +++++++++++++++++++++++++++++++ 5 files changed, 98 insertions(+), 7 deletions(-) base-commit: 6c0bd1fc70efaf053abe4e57c976afdc72d15377 Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-1920%2Fnewren%2Fmergeability-only-v3 Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-1920/newren/mergeability-only-v3 Pull-Request: https://github.com/gitgitgadget/git/pull/1920 Range-diff vs v2: 1: 4757c4810d3 = 1: 4757c4810d3 merge-ort: add a new mergeability_only option 2: 1d18ab7feb8 ! 2: f11824317a8 merge-tree: add a new --dry-run flag @@ Documentation/git-merge-tree.adoc: OPTIONS default is to include these messages if there are merge conflicts, and to omit them otherwise. -+--mergeability-only:: ++--dry-run:: + Disable all output from the program. Useful when you are only + interested in the exit status. Allows merge-tree to exit -+ early on the first conflict it finds, and allows it to avoid -+ writing most objects created by merges. ++ early when it finds a conflict, and allows it to avoid writing ++ most objects created by merges. + --allow-unrelated-histories:: merge-tree will by default error out if the two branches specified @@ builtin/merge-tree.c: int cmd_merge_tree(int argc, + if (dry_run && o.show_messages == -1) + o.show_messages = 0; + o.merge_options.mergeability_only = dry_run; -+ die_for_incompatible_opt2(dry_run, "--mergeability-only", ++ die_for_incompatible_opt2(dry_run, "--dry-run", + o.show_messages, "--messages"); -+ die_for_incompatible_opt2(dry_run, "--mergeability-only", ++ die_for_incompatible_opt2(dry_run, "--dry-run", + o.name_only, "--name-only"); -+ die_for_incompatible_opt2(dry_run, "--mergeability-only", ++ die_for_incompatible_opt2(dry_run, "--dry-run", + o.use_stdin, "--stdin"); -+ die_for_incompatible_opt2(dry_run, "--mergeability-only", ++ die_for_incompatible_opt2(dry_run, "--dry-run", + !line_termination, "-z"); + if (xopts.nr && o.mode == MODE_TRIVIAL) -- gitgitgadget ^ permalink raw reply [flat|nested] 29+ messages in thread
* [PATCH v3 1/2] merge-ort: add a new mergeability_only option 2025-05-14 0:24 ` [PATCH v3 0/2] merge-tree: add new --dry-run option Elijah Newren via GitGitGadget @ 2025-05-14 0:24 ` Elijah Newren via GitGitGadget 2025-05-14 0:24 ` [PATCH v3 2/2] merge-tree: add a new --dry-run flag Elijah Newren via GitGitGadget ` (2 subsequent siblings) 3 siblings, 0 replies; 29+ messages in thread From: Elijah Newren via GitGitGadget @ 2025-05-14 0:24 UTC (permalink / raw) To: git; +Cc: Elijah Newren, Kristoffer Haugsbakk, Elijah Newren, Elijah Newren From: Elijah Newren <newren@gmail.com> Git Forges may be interested in whether two branches can be merged while not being interested in what the resulting merge tree is nor which files conflicted. For such cases, add a new mergeability_only option. This option allows the merge machinery to, in the "outer layer" of the merge: * exit upon first[-ish] conflict * avoid (not prevent) writing merged blobs/trees to the object store I have a number of qualifiers there, so let me explain each: "outer layer": Note that since the recursive merge of merge bases (corresponding to call_depth > 0) can conflict without the outer final merge (corresponding to call_depth == 0) conflicting, we can't short-circuit nor avoid writing merged blobs/trees to the object store during those inner merges. "first-ish conflict": The current patch only exits early from process_entries() on the first conflict it detects, but conflicts could have been detected in a previous function call, namely detect_and_process_renames(). However: * conflicts detected by detect_and_process_renames() are quite rare conflict types * the detection would still come after regular rename detection (which is the expensive part of detect_and_process_renames()), so it is not saving us much in computation time given that process_entries() directly follows detect_and_process_renames() * [this overlaps with the next bullet point] process_entries() is the place where virtually all object writing occurs (object writing is sometimes more of a concern for Forges than computation time), so exiting early here isn't saving us much in object writes either * the code changes needed to handle an earlier exit are slightly more invasive in detect_and_process_renames() than for process_entries(). Given the rareness of the even earlier conflicts, the limited savings we'd get from exiting even earlier, and in an attempt to keep this patch simpler, we don't guarantee that we actually exit on the first conflict detected. We can always revisit this decision later if we decide that a further micro-optimization to exit slightly earlier in rare cases is worthwhile. "avoid (not prevent) writing objects": The detect_and_process_renames() call can also write objects to the object store, when rename/rename conflicts involve one (or more) files that have also been modified on both sides. Because of this alternate call path leading to handle_content_merges(), our "early exit" does not prevent writing objects entirely, even within the "outer layer" (i.e. even within call_depth == 0). I figure that's fine though, since we're already writing objects for the inner merges (i.e. for call_depth > 0), which are likely going to represent vastly more objects than files involved in rename/rename+modify/modify cases in the outer merge, on average. Signed-off-by: Elijah Newren <newren@gmail.com> --- merge-ort.c | 38 +++++++++++++++++++++++++++++++------- merge-ort.h | 1 + 2 files changed, 32 insertions(+), 7 deletions(-) diff --git a/merge-ort.c b/merge-ort.c index 77310a4a52c9..47b3d1730ece 100644 --- a/merge-ort.c +++ b/merge-ort.c @@ -2127,6 +2127,7 @@ static int handle_content_merge(struct merge_options *opt, const struct version_info *b, const char *pathnames[3], const int extra_marker_size, + const int record_object, struct version_info *result) { /* @@ -2214,7 +2215,7 @@ static int handle_content_merge(struct merge_options *opt, ret = -1; } - if (!ret && + if (!ret && record_object && write_object_file(result_buf.ptr, result_buf.size, OBJ_BLOB, &result->oid)) { path_msg(opt, ERROR_OBJECT_WRITE_FAILED, 0, @@ -2897,6 +2898,7 @@ static int process_renames(struct merge_options *opt, struct version_info merged; struct conflict_info *base, *side1, *side2; unsigned was_binary_blob = 0; + const int record_object = true; pathnames[0] = oldpath; pathnames[1] = newpath; @@ -2947,6 +2949,7 @@ static int process_renames(struct merge_options *opt, &side2->stages[2], pathnames, 1 + 2 * opt->priv->call_depth, + record_object, &merged); if (clean_merge < 0) return -1; @@ -3061,6 +3064,7 @@ static int process_renames(struct merge_options *opt, struct conflict_info *base, *side1, *side2; int clean; + const int record_object = true; pathnames[0] = oldpath; pathnames[other_source_index] = oldpath; @@ -3080,6 +3084,7 @@ static int process_renames(struct merge_options *opt, &side2->stages[2], pathnames, 1 + 2 * opt->priv->call_depth, + record_object, &merged); if (clean < 0) return -1; @@ -3931,9 +3936,12 @@ static int write_completed_directory(struct merge_options *opt, * Write out the tree to the git object directory, and also * record the mode and oid in dir_info->result. */ + int record_tree = (!opt->mergeability_only || + opt->priv->call_depth); dir_info->is_null = 0; dir_info->result.mode = S_IFDIR; - if (write_tree(&dir_info->result.oid, &info->versions, offset, + if (record_tree && + write_tree(&dir_info->result.oid, &info->versions, offset, opt->repo->hash_algo->rawsz) < 0) ret = -1; } @@ -4231,10 +4239,13 @@ static int process_entry(struct merge_options *opt, struct version_info *o = &ci->stages[0]; struct version_info *a = &ci->stages[1]; struct version_info *b = &ci->stages[2]; + int record_object = (!opt->mergeability_only || + opt->priv->call_depth); clean_merge = handle_content_merge(opt, path, o, a, b, ci->pathnames, opt->priv->call_depth * 2, + record_object, &merged_file); if (clean_merge < 0) return -1; @@ -4395,6 +4406,8 @@ static int process_entries(struct merge_options *opt, STRING_LIST_INIT_NODUP, NULL, 0 }; int ret = 0; + const int record_tree = (!opt->mergeability_only || + opt->priv->call_depth); trace2_region_enter("merge", "process_entries setup", opt->repo); if (strmap_empty(&opt->priv->paths)) { @@ -4454,6 +4467,12 @@ static int process_entries(struct merge_options *opt, ret = -1; goto cleanup; }; + if (!ci->merged.clean && opt->mergeability_only && + !opt->priv->call_depth) { + ret = 0; + goto cleanup; + } + } } trace2_region_leave("merge", "processing", opt->repo); @@ -4468,7 +4487,8 @@ static int process_entries(struct merge_options *opt, fflush(stdout); BUG("dir_metadata accounting completely off; shouldn't happen"); } - if (write_tree(result_oid, &dir_metadata.versions, 0, + if (record_tree && + write_tree(result_oid, &dir_metadata.versions, 0, opt->repo->hash_algo->rawsz) < 0) ret = -1; cleanup: @@ -4715,6 +4735,8 @@ void merge_display_update_messages(struct merge_options *opt, if (opt->record_conflict_msgs_as_headers) BUG("Either display conflict messages or record them as headers, not both"); + if (opt->mergeability_only) + BUG("Displaying conflict messages incompatible with mergeability-only checks"); trace2_region_enter("merge", "display messages", opt->repo); @@ -5171,10 +5193,12 @@ redo: result->path_messages = &opt->priv->conflicts; if (result->clean >= 0) { - result->tree = parse_tree_indirect(&working_tree_oid); - if (!result->tree) - die(_("unable to read tree (%s)"), - oid_to_hex(&working_tree_oid)); + if (!opt->mergeability_only) { + result->tree = parse_tree_indirect(&working_tree_oid); + if (!result->tree) + die(_("unable to read tree (%s)"), + oid_to_hex(&working_tree_oid)); + } /* existence of conflicted entries implies unclean */ result->clean &= strmap_empty(&opt->priv->conflicted); } diff --git a/merge-ort.h b/merge-ort.h index 30750c03962f..6045579825da 100644 --- a/merge-ort.h +++ b/merge-ort.h @@ -83,6 +83,7 @@ struct merge_options { /* miscellaneous control options */ const char *subtree_shift; unsigned renormalize : 1; + unsigned mergeability_only : 1; /* exit early, write fewer objects */ unsigned record_conflict_msgs_as_headers : 1; const char *msg_header_prefix; -- gitgitgadget ^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH v3 2/2] merge-tree: add a new --dry-run flag 2025-05-14 0:24 ` [PATCH v3 0/2] merge-tree: add new --dry-run option Elijah Newren via GitGitGadget 2025-05-14 0:24 ` [PATCH v3 1/2] merge-ort: add a new mergeability_only option Elijah Newren via GitGitGadget @ 2025-05-14 0:24 ` Elijah Newren via GitGitGadget 2025-05-15 13:07 ` Junio C Hamano 2025-05-16 13:18 ` Phillip Wood 2025-05-14 15:34 ` [PATCH v3 0/2] merge-tree: add new --dry-run option Kristoffer Haugsbakk 2025-05-16 20:04 ` [PATCH v4 0/2] merge-tree: add new --quiet option Elijah Newren via GitGitGadget 3 siblings, 2 replies; 29+ messages in thread From: Elijah Newren via GitGitGadget @ 2025-05-14 0:24 UTC (permalink / raw) To: git; +Cc: Elijah Newren, Kristoffer Haugsbakk, Elijah Newren, Elijah Newren From: Elijah Newren <newren@gmail.com> Git Forges may be interested in whether two branches can be merged while not being interested in what the resulting merge tree is nor which files conflicted. For such cases, add a new --dry-run flag which will make use of the new mergeability_only flag added to merge-ort in the previous commit. This option allows the merge machinery to, in the outer layer of the merge: * exit early when a conflict is detected * avoid writing (most) merged blobs/trees to the object store Signed-off-by: Elijah Newren <newren@gmail.com> --- Documentation/git-merge-tree.adoc | 6 +++++ builtin/merge-tree.c | 22 ++++++++++++++++++ t/t4301-merge-tree-write-tree.sh | 38 +++++++++++++++++++++++++++++++ 3 files changed, 66 insertions(+) diff --git a/Documentation/git-merge-tree.adoc b/Documentation/git-merge-tree.adoc index cf0578f9b5e8..74716b910192 100644 --- a/Documentation/git-merge-tree.adoc +++ b/Documentation/git-merge-tree.adoc @@ -65,6 +65,12 @@ OPTIONS default is to include these messages if there are merge conflicts, and to omit them otherwise. +--dry-run:: + Disable all output from the program. Useful when you are only + interested in the exit status. Allows merge-tree to exit + early when it finds a conflict, and allows it to avoid writing + most objects created by merges. + --allow-unrelated-histories:: merge-tree will by default error out if the two branches specified share no common history. This flag can be given to override that diff --git a/builtin/merge-tree.c b/builtin/merge-tree.c index 4aafa73c6155..273ec171e988 100644 --- a/builtin/merge-tree.c +++ b/builtin/merge-tree.c @@ -490,6 +490,9 @@ static int real_merge(struct merge_tree_options *o, if (result.clean < 0) die(_("failure to merge")); + if (o->merge_options.mergeability_only) + goto cleanup; + if (show_messages == -1) show_messages = !result.clean; @@ -522,6 +525,8 @@ static int real_merge(struct merge_tree_options *o, } if (o->use_stdin) putchar(line_termination); + +cleanup: merge_finalize(&opt, &result); clear_merge_options(&opt); return !result.clean; /* result.clean < 0 handled above */ @@ -538,6 +543,7 @@ int cmd_merge_tree(int argc, int original_argc; const char *merge_base = NULL; int ret; + int dry_run = 0; const char * const merge_tree_usage[] = { N_("git merge-tree [--write-tree] [<options>] <branch1> <branch2>"), @@ -552,6 +558,10 @@ int cmd_merge_tree(int argc, N_("do a trivial merge only"), MODE_TRIVIAL), OPT_BOOL(0, "messages", &o.show_messages, N_("also show informational/conflict messages")), + OPT_BOOL_F(0, "dry-run", + &dry_run, + N_("suppress all output; only exit status wanted"), + PARSE_OPT_NONEG), OPT_SET_INT('z', NULL, &line_termination, N_("separate paths with the NUL character"), '\0'), OPT_BOOL_F(0, "name-only", @@ -583,6 +593,18 @@ int cmd_merge_tree(int argc, argc = parse_options(argc, argv, prefix, mt_options, merge_tree_usage, PARSE_OPT_STOP_AT_NON_OPTION); + if (dry_run && o.show_messages == -1) + o.show_messages = 0; + o.merge_options.mergeability_only = dry_run; + die_for_incompatible_opt2(dry_run, "--dry-run", + o.show_messages, "--messages"); + die_for_incompatible_opt2(dry_run, "--dry-run", + o.name_only, "--name-only"); + die_for_incompatible_opt2(dry_run, "--dry-run", + o.use_stdin, "--stdin"); + die_for_incompatible_opt2(dry_run, "--dry-run", + !line_termination, "-z"); + if (xopts.nr && o.mode == MODE_TRIVIAL) die(_("--trivial-merge is incompatible with all other options")); for (size_t x = 0; x < xopts.nr; x++) diff --git a/t/t4301-merge-tree-write-tree.sh b/t/t4301-merge-tree-write-tree.sh index f9c5883a7f7c..566a2b4ec737 100755 --- a/t/t4301-merge-tree-write-tree.sh +++ b/t/t4301-merge-tree-write-tree.sh @@ -54,6 +54,25 @@ test_expect_success setup ' git commit -m first-commit ' +test_expect_success '--dry-run on clean merge' ' + # Get rid of loose objects to start with + git gc && + echo "0 objects, 0 kilobytes" >expect && + git count-objects >actual && + test_cmp expect actual && + + # Ensure merge is successful (exit code of 0) + git merge-tree --write-tree --dry-run side1 side3 >output && + + # Ensure there is no output + test_must_be_empty output && + + # Ensure no loose objects written (all new objects written would have + # been in "outer layer" of the merge) + git count-objects >actual && + test_cmp expect actual +' + test_expect_success 'Clean merge' ' TREE_OID=$(git merge-tree --write-tree side1 side3) && q_to_tab <<-EOF >expect && @@ -72,6 +91,25 @@ test_expect_success 'Failed merge without rename detection' ' grep "CONFLICT (modify/delete): numbers deleted" out ' +test_expect_success '--dry-run on conflicted merge' ' + # Get rid of loose objects to start with + git gc && + echo "0 objects, 0 kilobytes" >expect && + git count-objects >actual && + test_cmp expect actual && + + # Ensure merge has conflict + test_expect_code 1 git merge-tree --write-tree --dry-run side1 side2 >output && + + # Ensure there is no output + test_must_be_empty output && + + # Ensure no loose objects written (all new objects written would have + # been in "outer layer" of the merge) + git count-objects >actual && + test_cmp expect actual +' + test_expect_success 'Content merge and a few conflicts' ' git checkout side1^0 && test_must_fail git merge side2 && -- gitgitgadget ^ permalink raw reply related [flat|nested] 29+ messages in thread
* Re: [PATCH v3 2/2] merge-tree: add a new --dry-run flag 2025-05-14 0:24 ` [PATCH v3 2/2] merge-tree: add a new --dry-run flag Elijah Newren via GitGitGadget @ 2025-05-15 13:07 ` Junio C Hamano 2025-05-16 13:18 ` Phillip Wood 1 sibling, 0 replies; 29+ messages in thread From: Junio C Hamano @ 2025-05-15 13:07 UTC (permalink / raw) To: Elijah Newren via GitGitGadget; +Cc: git, Elijah Newren, Kristoffer Haugsbakk "Elijah Newren via GitGitGadget" <gitgitgadget@gmail.com> writes: > +--dry-run:: > + Disable all output from the program. Useful when you are only > + interested in the exit status. Allows merge-tree to exit > + early when it finds a conflict, and allows it to avoid writing > + most objects created by merges. > + OK. > @@ -583,6 +593,18 @@ int cmd_merge_tree(int argc, > argc = parse_options(argc, argv, prefix, mt_options, > merge_tree_usage, PARSE_OPT_STOP_AT_NON_OPTION); > > + if (dry_run && o.show_messages == -1) > + o.show_messages = 0; > + o.merge_options.mergeability_only = dry_run; > + die_for_incompatible_opt2(dry_run, "--dry-run", > + o.show_messages, "--messages"); > + die_for_incompatible_opt2(dry_run, "--dry-run", > + o.name_only, "--name-only"); > + die_for_incompatible_opt2(dry_run, "--dry-run", > + o.use_stdin, "--stdin"); > + die_for_incompatible_opt2(dry_run, "--dry-run", > + !line_termination, "-z"); Ah, I missed this one. Would have been caught in the previous round if we had test coverage for incompatible options. This round looks quite sensible to me. Queued. Thanks. ^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH v3 2/2] merge-tree: add a new --dry-run flag 2025-05-14 0:24 ` [PATCH v3 2/2] merge-tree: add a new --dry-run flag Elijah Newren via GitGitGadget 2025-05-15 13:07 ` Junio C Hamano @ 2025-05-16 13:18 ` Phillip Wood 2025-05-16 16:03 ` Elijah Newren 1 sibling, 1 reply; 29+ messages in thread From: Phillip Wood @ 2025-05-16 13:18 UTC (permalink / raw) To: Elijah Newren via GitGitGadget, git; +Cc: Elijah Newren, Kristoffer Haugsbakk Hi Elijah On 14/05/2025 01:24, Elijah Newren via GitGitGadget wrote: > From: Elijah Newren <newren@gmail.com> > > +--dry-run:: > + Disable all output from the program. Useful when you are only > + interested in the exit status. Allows merge-tree to exit > + early when it finds a conflict, and allows it to avoid writing > + most objects created by merges. I don't want to bike-shed but to me this feels more like "git diff --quiet" than "git push --dry-run" or "git send-email --dry-run" which still print diagnostic messages. Best Wishes Phillip > --allow-unrelated-histories:: > merge-tree will by default error out if the two branches specified > share no common history. This flag can be given to override that > diff --git a/builtin/merge-tree.c b/builtin/merge-tree.c > index 4aafa73c6155..273ec171e988 100644 > --- a/builtin/merge-tree.c > +++ b/builtin/merge-tree.c > @@ -490,6 +490,9 @@ static int real_merge(struct merge_tree_options *o, > if (result.clean < 0) > die(_("failure to merge")); > > + if (o->merge_options.mergeability_only) > + goto cleanup; > + > if (show_messages == -1) > show_messages = !result.clean; > > @@ -522,6 +525,8 @@ static int real_merge(struct merge_tree_options *o, > } > if (o->use_stdin) > putchar(line_termination); > + > +cleanup: > merge_finalize(&opt, &result); > clear_merge_options(&opt); > return !result.clean; /* result.clean < 0 handled above */ > @@ -538,6 +543,7 @@ int cmd_merge_tree(int argc, > int original_argc; > const char *merge_base = NULL; > int ret; > + int dry_run = 0; > > const char * const merge_tree_usage[] = { > N_("git merge-tree [--write-tree] [<options>] <branch1> <branch2>"), > @@ -552,6 +558,10 @@ int cmd_merge_tree(int argc, > N_("do a trivial merge only"), MODE_TRIVIAL), > OPT_BOOL(0, "messages", &o.show_messages, > N_("also show informational/conflict messages")), > + OPT_BOOL_F(0, "dry-run", > + &dry_run, > + N_("suppress all output; only exit status wanted"), > + PARSE_OPT_NONEG), > OPT_SET_INT('z', NULL, &line_termination, > N_("separate paths with the NUL character"), '\0'), > OPT_BOOL_F(0, "name-only", > @@ -583,6 +593,18 @@ int cmd_merge_tree(int argc, > argc = parse_options(argc, argv, prefix, mt_options, > merge_tree_usage, PARSE_OPT_STOP_AT_NON_OPTION); > > + if (dry_run && o.show_messages == -1) > + o.show_messages = 0; > + o.merge_options.mergeability_only = dry_run; > + die_for_incompatible_opt2(dry_run, "--dry-run", > + o.show_messages, "--messages"); > + die_for_incompatible_opt2(dry_run, "--dry-run", > + o.name_only, "--name-only"); > + die_for_incompatible_opt2(dry_run, "--dry-run", > + o.use_stdin, "--stdin"); > + die_for_incompatible_opt2(dry_run, "--dry-run", > + !line_termination, "-z"); > + > if (xopts.nr && o.mode == MODE_TRIVIAL) > die(_("--trivial-merge is incompatible with all other options")); > for (size_t x = 0; x < xopts.nr; x++) > diff --git a/t/t4301-merge-tree-write-tree.sh b/t/t4301-merge-tree-write-tree.sh > index f9c5883a7f7c..566a2b4ec737 100755 > --- a/t/t4301-merge-tree-write-tree.sh > +++ b/t/t4301-merge-tree-write-tree.sh > @@ -54,6 +54,25 @@ test_expect_success setup ' > git commit -m first-commit > ' > > +test_expect_success '--dry-run on clean merge' ' > + # Get rid of loose objects to start with > + git gc && > + echo "0 objects, 0 kilobytes" >expect && > + git count-objects >actual && > + test_cmp expect actual && > + > + # Ensure merge is successful (exit code of 0) > + git merge-tree --write-tree --dry-run side1 side3 >output && > + > + # Ensure there is no output > + test_must_be_empty output && > + > + # Ensure no loose objects written (all new objects written would have > + # been in "outer layer" of the merge) > + git count-objects >actual && > + test_cmp expect actual > +' > + > test_expect_success 'Clean merge' ' > TREE_OID=$(git merge-tree --write-tree side1 side3) && > q_to_tab <<-EOF >expect && > @@ -72,6 +91,25 @@ test_expect_success 'Failed merge without rename detection' ' > grep "CONFLICT (modify/delete): numbers deleted" out > ' > > +test_expect_success '--dry-run on conflicted merge' ' > + # Get rid of loose objects to start with > + git gc && > + echo "0 objects, 0 kilobytes" >expect && > + git count-objects >actual && > + test_cmp expect actual && > + > + # Ensure merge has conflict > + test_expect_code 1 git merge-tree --write-tree --dry-run side1 side2 >output && > + > + # Ensure there is no output > + test_must_be_empty output && > + > + # Ensure no loose objects written (all new objects written would have > + # been in "outer layer" of the merge) > + git count-objects >actual && > + test_cmp expect actual > +' > + > test_expect_success 'Content merge and a few conflicts' ' > git checkout side1^0 && > test_must_fail git merge side2 && ^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH v3 2/2] merge-tree: add a new --dry-run flag 2025-05-16 13:18 ` Phillip Wood @ 2025-05-16 16:03 ` Elijah Newren 0 siblings, 0 replies; 29+ messages in thread From: Elijah Newren @ 2025-05-16 16:03 UTC (permalink / raw) To: phillip.wood; +Cc: Elijah Newren via GitGitGadget, git, Kristoffer Haugsbakk On Fri, May 16, 2025 at 6:18 AM Phillip Wood <phillip.wood123@gmail.com> wrote: > > Hi Elijah > > On 14/05/2025 01:24, Elijah Newren via GitGitGadget wrote: > > From: Elijah Newren <newren@gmail.com> > > > > +--dry-run:: > > + Disable all output from the program. Useful when you are only > > + interested in the exit status. Allows merge-tree to exit > > + early when it finds a conflict, and allows it to avoid writing > > + most objects created by merges. > > I don't want to bike-shed but to me this feels more like "git diff > --quiet" than "git push --dry-run" or "git send-email --dry-run" which > still print diagnostic messages. > > Best Wishes > > Phillip --quiet seems fine to me. I'll make the change and then expect someone to comment on it with an even better name... ;-) ^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH v3 0/2] merge-tree: add new --dry-run option 2025-05-14 0:24 ` [PATCH v3 0/2] merge-tree: add new --dry-run option Elijah Newren via GitGitGadget 2025-05-14 0:24 ` [PATCH v3 1/2] merge-ort: add a new mergeability_only option Elijah Newren via GitGitGadget 2025-05-14 0:24 ` [PATCH v3 2/2] merge-tree: add a new --dry-run flag Elijah Newren via GitGitGadget @ 2025-05-14 15:34 ` Kristoffer Haugsbakk 2025-05-16 20:04 ` [PATCH v4 0/2] merge-tree: add new --quiet option Elijah Newren via GitGitGadget 3 siblings, 0 replies; 29+ messages in thread From: Kristoffer Haugsbakk @ 2025-05-14 15:34 UTC (permalink / raw) To: Josh Soref, git; +Cc: Elijah Newren On Wed, May 14, 2025, at 02:24, Elijah Newren via GitGitGadget wrote: > Changes since v2: > > * Converted locations missed in v1 in changing --mergeability-only -> > --dry-run > > Changes since v1: > > * Renamed --mergeability-only flag to --dry-run, as per suggestion from > Junio > * added some commit message clarifications > > This adds a new flag, --dry-run, to git merge-tree, which suppresses all > output and leaves only the exit status (reflecting successful merge or > conflict). This is useful for Git Forges in cases where they are only > interested in whether two branches can be merged, without needing the actual > merge result or conflict details. > > The advantage of the flag is two fold: > > * The merge machinery can exit once it detects a conflict, instead of > continuing to compute merge result information > * The merge machinery can avoid writing merged blobs and trees to the > object store when in the outer layer of the merging process (more details > in the first commit message). > > Elijah Newren (2): > merge-ort: add a new mergeability_only option > merge-tree: add a new --dry-run flag All I can say is that this looks good considering the comments on v2. Interdiff: ``` diff --git a/Documentation/git-merge-tree.adoc b/Documentation/git-merge-tree.adoc index 7dcc1780619..74716b91019 100644 --- a/Documentation/git-merge-tree.adoc +++ b/Documentation/git-merge-tree.adoc @@ -65,11 +65,11 @@ OPTIONS default is to include these messages if there are merge conflicts, and to omit them otherwise. ---mergeability-only:: +--dry-run:: Disable all output from the program. Useful when you are only interested in the exit status. Allows merge-tree to exit - early on the first conflict it finds, and allows it to avoid - writing most objects created by merges. + early when it finds a conflict, and allows it to avoid writing + most objects created by merges. --allow-unrelated-histories:: merge-tree will by default error out if the two branches specified diff --git a/builtin/merge-tree.c b/builtin/merge-tree.c index 579e81d5184..273ec171e98 100644 --- a/builtin/merge-tree.c +++ b/builtin/merge-tree.c @@ -596,13 +596,13 @@ int cmd_merge_tree(int argc, if (dry_run && o.show_messages == -1) o.show_messages = 0; o.merge_options.mergeability_only = dry_run; - die_for_incompatible_opt2(dry_run, "--mergeability-only", + die_for_incompatible_opt2(dry_run, "--dry-run", o.show_messages, "--messages"); - die_for_incompatible_opt2(dry_run, "--mergeability-only", + die_for_incompatible_opt2(dry_run, "--dry-run", o.name_only, "--name-only"); - die_for_incompatible_opt2(dry_run, "--mergeability-only", + die_for_incompatible_opt2(dry_run, "--dry-run", o.use_stdin, "--stdin"); - die_for_incompatible_opt2(dry_run, "--mergeability-only", + die_for_incompatible_opt2(dry_run, "--dry-run", !line_termination, "-z"); if (xopts.nr && o.mode == MODE_TRIVIAL) ``` ^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH v4 0/2] merge-tree: add new --quiet option 2025-05-14 0:24 ` [PATCH v3 0/2] merge-tree: add new --dry-run option Elijah Newren via GitGitGadget ` (2 preceding siblings ...) 2025-05-14 15:34 ` [PATCH v3 0/2] merge-tree: add new --dry-run option Kristoffer Haugsbakk @ 2025-05-16 20:04 ` Elijah Newren via GitGitGadget 2025-05-16 20:04 ` [PATCH v4 1/2] merge-ort: add a new mergeability_only option Elijah Newren via GitGitGadget ` (2 more replies) 3 siblings, 3 replies; 29+ messages in thread From: Elijah Newren via GitGitGadget @ 2025-05-16 20:04 UTC (permalink / raw) To: git; +Cc: Elijah Newren, Kristoffer Haugsbakk, Phillip Wood, Elijah Newren Changes since v3: * Renamed --dry-run -> --quiet . Any further naming suggestions? Changes since v2: * Converted locations missed in v1 in changing --mergeability-only -> --dry-run Changes since v1: * Renamed --mergeability-only flag to --dry-run, as per suggestion from Junio * added some commit message clarifications This adds a new flag, --dry-run, to git merge-tree, which suppresses all output and leaves only the exit status (reflecting successful merge or conflict). This is useful for Git Forges in cases where they are only interested in whether two branches can be merged, without needing the actual merge result or conflict details. The advantage of the flag is two fold: * The merge machinery can exit once it detects a conflict, instead of continuing to compute merge result information * The merge machinery can avoid writing merged blobs and trees to the object store when in the outer layer of the merging process (more details in the first commit message). Elijah Newren (2): merge-ort: add a new mergeability_only option merge-tree: add a new --quiet flag Documentation/git-merge-tree.adoc | 6 +++++ builtin/merge-tree.c | 18 +++++++++++++++ merge-ort.c | 38 +++++++++++++++++++++++++------ merge-ort.h | 1 + t/t4301-merge-tree-write-tree.sh | 38 +++++++++++++++++++++++++++++++ 5 files changed, 94 insertions(+), 7 deletions(-) base-commit: 6c0bd1fc70efaf053abe4e57c976afdc72d15377 Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-1920%2Fnewren%2Fmergeability-only-v4 Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-1920/newren/mergeability-only-v4 Pull-Request: https://github.com/gitgitgadget/git/pull/1920 Range-diff vs v3: 1: 4757c4810d3 = 1: 4757c4810d3 merge-ort: add a new mergeability_only option 2: f11824317a8 ! 2: 7c40d3c9216 merge-tree: add a new --dry-run flag @@ Metadata Author: Elijah Newren <newren@gmail.com> ## Commit message ## - merge-tree: add a new --dry-run flag + merge-tree: add a new --quiet flag Git Forges may be interested in whether two branches can be merged while not being interested in what the resulting merge tree is nor which files - conflicted. For such cases, add a new --dry-run flag which + conflicted. For such cases, add a new --quiet flag which will make use of the new mergeability_only flag added to merge-ort in the previous commit. This option allows the merge machinery to, in the outer layer of the merge: @@ Documentation/git-merge-tree.adoc: OPTIONS default is to include these messages if there are merge conflicts, and to omit them otherwise. -+--dry-run:: ++--quiet:: + Disable all output from the program. Useful when you are only + interested in the exit status. Allows merge-tree to exit + early when it finds a conflict, and allows it to avoid writing @@ builtin/merge-tree.c: int cmd_merge_tree(int argc, int original_argc; const char *merge_base = NULL; int ret; -+ int dry_run = 0; ++ int quiet = 0; const char * const merge_tree_usage[] = { N_("git merge-tree [--write-tree] [<options>] <branch1> <branch2>"), @@ builtin/merge-tree.c: int cmd_merge_tree(int argc, N_("do a trivial merge only"), MODE_TRIVIAL), OPT_BOOL(0, "messages", &o.show_messages, N_("also show informational/conflict messages")), -+ OPT_BOOL_F(0, "dry-run", -+ &dry_run, ++ OPT_BOOL_F(0, "quiet", ++ &quiet, + N_("suppress all output; only exit status wanted"), + PARSE_OPT_NONEG), OPT_SET_INT('z', NULL, &line_termination, @@ builtin/merge-tree.c: int cmd_merge_tree(int argc, argc = parse_options(argc, argv, prefix, mt_options, merge_tree_usage, PARSE_OPT_STOP_AT_NON_OPTION); -+ if (dry_run && o.show_messages == -1) ++ if (quiet && o.show_messages == -1) + o.show_messages = 0; -+ o.merge_options.mergeability_only = dry_run; -+ die_for_incompatible_opt2(dry_run, "--dry-run", -+ o.show_messages, "--messages"); -+ die_for_incompatible_opt2(dry_run, "--dry-run", -+ o.name_only, "--name-only"); -+ die_for_incompatible_opt2(dry_run, "--dry-run", -+ o.use_stdin, "--stdin"); -+ die_for_incompatible_opt2(dry_run, "--dry-run", -+ !line_termination, "-z"); ++ o.merge_options.mergeability_only = quiet; ++ die_for_incompatible_opt2(quiet, "--quiet", o.show_messages, "--messages"); ++ die_for_incompatible_opt2(quiet, "--quiet", o.name_only, "--name-only"); ++ die_for_incompatible_opt2(quiet, "--quiet", o.use_stdin, "--stdin"); ++ die_for_incompatible_opt2(quiet, "--quiet", !line_termination, "-z"); + if (xopts.nr && o.mode == MODE_TRIVIAL) die(_("--trivial-merge is incompatible with all other options")); @@ t/t4301-merge-tree-write-tree.sh: test_expect_success setup ' git commit -m first-commit ' -+test_expect_success '--dry-run on clean merge' ' ++test_expect_success '--quiet on clean merge' ' + # Get rid of loose objects to start with + git gc && + echo "0 objects, 0 kilobytes" >expect && @@ t/t4301-merge-tree-write-tree.sh: test_expect_success setup ' + test_cmp expect actual && + + # Ensure merge is successful (exit code of 0) -+ git merge-tree --write-tree --dry-run side1 side3 >output && ++ git merge-tree --write-tree --quiet side1 side3 >output && + + # Ensure there is no output + test_must_be_empty output && @@ t/t4301-merge-tree-write-tree.sh: test_expect_success 'Failed merge without rena grep "CONFLICT (modify/delete): numbers deleted" out ' -+test_expect_success '--dry-run on conflicted merge' ' ++test_expect_success '--quiet on conflicted merge' ' + # Get rid of loose objects to start with + git gc && + echo "0 objects, 0 kilobytes" >expect && @@ t/t4301-merge-tree-write-tree.sh: test_expect_success 'Failed merge without rena + test_cmp expect actual && + + # Ensure merge has conflict -+ test_expect_code 1 git merge-tree --write-tree --dry-run side1 side2 >output && ++ test_expect_code 1 git merge-tree --write-tree --quiet side1 side2 >output && + + # Ensure there is no output + test_must_be_empty output && -- gitgitgadget ^ permalink raw reply [flat|nested] 29+ messages in thread
* [PATCH v4 1/2] merge-ort: add a new mergeability_only option 2025-05-16 20:04 ` [PATCH v4 0/2] merge-tree: add new --quiet option Elijah Newren via GitGitGadget @ 2025-05-16 20:04 ` Elijah Newren via GitGitGadget 2025-05-16 20:04 ` [PATCH v4 2/2] merge-tree: add a new --quiet flag Elijah Newren via GitGitGadget 2025-05-19 9:05 ` [PATCH v4 0/2] merge-tree: add new --quiet option Phillip Wood 2 siblings, 0 replies; 29+ messages in thread From: Elijah Newren via GitGitGadget @ 2025-05-16 20:04 UTC (permalink / raw) To: git Cc: Elijah Newren, Kristoffer Haugsbakk, Phillip Wood, Elijah Newren, Elijah Newren From: Elijah Newren <newren@gmail.com> Git Forges may be interested in whether two branches can be merged while not being interested in what the resulting merge tree is nor which files conflicted. For such cases, add a new mergeability_only option. This option allows the merge machinery to, in the "outer layer" of the merge: * exit upon first[-ish] conflict * avoid (not prevent) writing merged blobs/trees to the object store I have a number of qualifiers there, so let me explain each: "outer layer": Note that since the recursive merge of merge bases (corresponding to call_depth > 0) can conflict without the outer final merge (corresponding to call_depth == 0) conflicting, we can't short-circuit nor avoid writing merged blobs/trees to the object store during those inner merges. "first-ish conflict": The current patch only exits early from process_entries() on the first conflict it detects, but conflicts could have been detected in a previous function call, namely detect_and_process_renames(). However: * conflicts detected by detect_and_process_renames() are quite rare conflict types * the detection would still come after regular rename detection (which is the expensive part of detect_and_process_renames()), so it is not saving us much in computation time given that process_entries() directly follows detect_and_process_renames() * [this overlaps with the next bullet point] process_entries() is the place where virtually all object writing occurs (object writing is sometimes more of a concern for Forges than computation time), so exiting early here isn't saving us much in object writes either * the code changes needed to handle an earlier exit are slightly more invasive in detect_and_process_renames() than for process_entries(). Given the rareness of the even earlier conflicts, the limited savings we'd get from exiting even earlier, and in an attempt to keep this patch simpler, we don't guarantee that we actually exit on the first conflict detected. We can always revisit this decision later if we decide that a further micro-optimization to exit slightly earlier in rare cases is worthwhile. "avoid (not prevent) writing objects": The detect_and_process_renames() call can also write objects to the object store, when rename/rename conflicts involve one (or more) files that have also been modified on both sides. Because of this alternate call path leading to handle_content_merges(), our "early exit" does not prevent writing objects entirely, even within the "outer layer" (i.e. even within call_depth == 0). I figure that's fine though, since we're already writing objects for the inner merges (i.e. for call_depth > 0), which are likely going to represent vastly more objects than files involved in rename/rename+modify/modify cases in the outer merge, on average. Signed-off-by: Elijah Newren <newren@gmail.com> --- merge-ort.c | 38 +++++++++++++++++++++++++++++++------- merge-ort.h | 1 + 2 files changed, 32 insertions(+), 7 deletions(-) diff --git a/merge-ort.c b/merge-ort.c index 77310a4a52c9..47b3d1730ece 100644 --- a/merge-ort.c +++ b/merge-ort.c @@ -2127,6 +2127,7 @@ static int handle_content_merge(struct merge_options *opt, const struct version_info *b, const char *pathnames[3], const int extra_marker_size, + const int record_object, struct version_info *result) { /* @@ -2214,7 +2215,7 @@ static int handle_content_merge(struct merge_options *opt, ret = -1; } - if (!ret && + if (!ret && record_object && write_object_file(result_buf.ptr, result_buf.size, OBJ_BLOB, &result->oid)) { path_msg(opt, ERROR_OBJECT_WRITE_FAILED, 0, @@ -2897,6 +2898,7 @@ static int process_renames(struct merge_options *opt, struct version_info merged; struct conflict_info *base, *side1, *side2; unsigned was_binary_blob = 0; + const int record_object = true; pathnames[0] = oldpath; pathnames[1] = newpath; @@ -2947,6 +2949,7 @@ static int process_renames(struct merge_options *opt, &side2->stages[2], pathnames, 1 + 2 * opt->priv->call_depth, + record_object, &merged); if (clean_merge < 0) return -1; @@ -3061,6 +3064,7 @@ static int process_renames(struct merge_options *opt, struct conflict_info *base, *side1, *side2; int clean; + const int record_object = true; pathnames[0] = oldpath; pathnames[other_source_index] = oldpath; @@ -3080,6 +3084,7 @@ static int process_renames(struct merge_options *opt, &side2->stages[2], pathnames, 1 + 2 * opt->priv->call_depth, + record_object, &merged); if (clean < 0) return -1; @@ -3931,9 +3936,12 @@ static int write_completed_directory(struct merge_options *opt, * Write out the tree to the git object directory, and also * record the mode and oid in dir_info->result. */ + int record_tree = (!opt->mergeability_only || + opt->priv->call_depth); dir_info->is_null = 0; dir_info->result.mode = S_IFDIR; - if (write_tree(&dir_info->result.oid, &info->versions, offset, + if (record_tree && + write_tree(&dir_info->result.oid, &info->versions, offset, opt->repo->hash_algo->rawsz) < 0) ret = -1; } @@ -4231,10 +4239,13 @@ static int process_entry(struct merge_options *opt, struct version_info *o = &ci->stages[0]; struct version_info *a = &ci->stages[1]; struct version_info *b = &ci->stages[2]; + int record_object = (!opt->mergeability_only || + opt->priv->call_depth); clean_merge = handle_content_merge(opt, path, o, a, b, ci->pathnames, opt->priv->call_depth * 2, + record_object, &merged_file); if (clean_merge < 0) return -1; @@ -4395,6 +4406,8 @@ static int process_entries(struct merge_options *opt, STRING_LIST_INIT_NODUP, NULL, 0 }; int ret = 0; + const int record_tree = (!opt->mergeability_only || + opt->priv->call_depth); trace2_region_enter("merge", "process_entries setup", opt->repo); if (strmap_empty(&opt->priv->paths)) { @@ -4454,6 +4467,12 @@ static int process_entries(struct merge_options *opt, ret = -1; goto cleanup; }; + if (!ci->merged.clean && opt->mergeability_only && + !opt->priv->call_depth) { + ret = 0; + goto cleanup; + } + } } trace2_region_leave("merge", "processing", opt->repo); @@ -4468,7 +4487,8 @@ static int process_entries(struct merge_options *opt, fflush(stdout); BUG("dir_metadata accounting completely off; shouldn't happen"); } - if (write_tree(result_oid, &dir_metadata.versions, 0, + if (record_tree && + write_tree(result_oid, &dir_metadata.versions, 0, opt->repo->hash_algo->rawsz) < 0) ret = -1; cleanup: @@ -4715,6 +4735,8 @@ void merge_display_update_messages(struct merge_options *opt, if (opt->record_conflict_msgs_as_headers) BUG("Either display conflict messages or record them as headers, not both"); + if (opt->mergeability_only) + BUG("Displaying conflict messages incompatible with mergeability-only checks"); trace2_region_enter("merge", "display messages", opt->repo); @@ -5171,10 +5193,12 @@ redo: result->path_messages = &opt->priv->conflicts; if (result->clean >= 0) { - result->tree = parse_tree_indirect(&working_tree_oid); - if (!result->tree) - die(_("unable to read tree (%s)"), - oid_to_hex(&working_tree_oid)); + if (!opt->mergeability_only) { + result->tree = parse_tree_indirect(&working_tree_oid); + if (!result->tree) + die(_("unable to read tree (%s)"), + oid_to_hex(&working_tree_oid)); + } /* existence of conflicted entries implies unclean */ result->clean &= strmap_empty(&opt->priv->conflicted); } diff --git a/merge-ort.h b/merge-ort.h index 30750c03962f..6045579825da 100644 --- a/merge-ort.h +++ b/merge-ort.h @@ -83,6 +83,7 @@ struct merge_options { /* miscellaneous control options */ const char *subtree_shift; unsigned renormalize : 1; + unsigned mergeability_only : 1; /* exit early, write fewer objects */ unsigned record_conflict_msgs_as_headers : 1; const char *msg_header_prefix; -- gitgitgadget ^ permalink raw reply related [flat|nested] 29+ messages in thread
* [PATCH v4 2/2] merge-tree: add a new --quiet flag 2025-05-16 20:04 ` [PATCH v4 0/2] merge-tree: add new --quiet option Elijah Newren via GitGitGadget 2025-05-16 20:04 ` [PATCH v4 1/2] merge-ort: add a new mergeability_only option Elijah Newren via GitGitGadget @ 2025-05-16 20:04 ` Elijah Newren via GitGitGadget 2025-05-17 19:52 ` Kristoffer Haugsbakk 2025-05-19 9:05 ` [PATCH v4 0/2] merge-tree: add new --quiet option Phillip Wood 2 siblings, 1 reply; 29+ messages in thread From: Elijah Newren via GitGitGadget @ 2025-05-16 20:04 UTC (permalink / raw) To: git Cc: Elijah Newren, Kristoffer Haugsbakk, Phillip Wood, Elijah Newren, Elijah Newren From: Elijah Newren <newren@gmail.com> Git Forges may be interested in whether two branches can be merged while not being interested in what the resulting merge tree is nor which files conflicted. For such cases, add a new --quiet flag which will make use of the new mergeability_only flag added to merge-ort in the previous commit. This option allows the merge machinery to, in the outer layer of the merge: * exit early when a conflict is detected * avoid writing (most) merged blobs/trees to the object store Signed-off-by: Elijah Newren <newren@gmail.com> --- Documentation/git-merge-tree.adoc | 6 +++++ builtin/merge-tree.c | 18 +++++++++++++++ t/t4301-merge-tree-write-tree.sh | 38 +++++++++++++++++++++++++++++++ 3 files changed, 62 insertions(+) diff --git a/Documentation/git-merge-tree.adoc b/Documentation/git-merge-tree.adoc index cf0578f9b5e8..f824eea61f1e 100644 --- a/Documentation/git-merge-tree.adoc +++ b/Documentation/git-merge-tree.adoc @@ -65,6 +65,12 @@ OPTIONS default is to include these messages if there are merge conflicts, and to omit them otherwise. +--quiet:: + Disable all output from the program. Useful when you are only + interested in the exit status. Allows merge-tree to exit + early when it finds a conflict, and allows it to avoid writing + most objects created by merges. + --allow-unrelated-histories:: merge-tree will by default error out if the two branches specified share no common history. This flag can be given to override that diff --git a/builtin/merge-tree.c b/builtin/merge-tree.c index 4aafa73c6155..7f41665dfd7e 100644 --- a/builtin/merge-tree.c +++ b/builtin/merge-tree.c @@ -490,6 +490,9 @@ static int real_merge(struct merge_tree_options *o, if (result.clean < 0) die(_("failure to merge")); + if (o->merge_options.mergeability_only) + goto cleanup; + if (show_messages == -1) show_messages = !result.clean; @@ -522,6 +525,8 @@ static int real_merge(struct merge_tree_options *o, } if (o->use_stdin) putchar(line_termination); + +cleanup: merge_finalize(&opt, &result); clear_merge_options(&opt); return !result.clean; /* result.clean < 0 handled above */ @@ -538,6 +543,7 @@ int cmd_merge_tree(int argc, int original_argc; const char *merge_base = NULL; int ret; + int quiet = 0; const char * const merge_tree_usage[] = { N_("git merge-tree [--write-tree] [<options>] <branch1> <branch2>"), @@ -552,6 +558,10 @@ int cmd_merge_tree(int argc, N_("do a trivial merge only"), MODE_TRIVIAL), OPT_BOOL(0, "messages", &o.show_messages, N_("also show informational/conflict messages")), + OPT_BOOL_F(0, "quiet", + &quiet, + N_("suppress all output; only exit status wanted"), + PARSE_OPT_NONEG), OPT_SET_INT('z', NULL, &line_termination, N_("separate paths with the NUL character"), '\0'), OPT_BOOL_F(0, "name-only", @@ -583,6 +593,14 @@ int cmd_merge_tree(int argc, argc = parse_options(argc, argv, prefix, mt_options, merge_tree_usage, PARSE_OPT_STOP_AT_NON_OPTION); + if (quiet && o.show_messages == -1) + o.show_messages = 0; + o.merge_options.mergeability_only = quiet; + die_for_incompatible_opt2(quiet, "--quiet", o.show_messages, "--messages"); + die_for_incompatible_opt2(quiet, "--quiet", o.name_only, "--name-only"); + die_for_incompatible_opt2(quiet, "--quiet", o.use_stdin, "--stdin"); + die_for_incompatible_opt2(quiet, "--quiet", !line_termination, "-z"); + if (xopts.nr && o.mode == MODE_TRIVIAL) die(_("--trivial-merge is incompatible with all other options")); for (size_t x = 0; x < xopts.nr; x++) diff --git a/t/t4301-merge-tree-write-tree.sh b/t/t4301-merge-tree-write-tree.sh index f9c5883a7f7c..6e117ee93c8b 100755 --- a/t/t4301-merge-tree-write-tree.sh +++ b/t/t4301-merge-tree-write-tree.sh @@ -54,6 +54,25 @@ test_expect_success setup ' git commit -m first-commit ' +test_expect_success '--quiet on clean merge' ' + # Get rid of loose objects to start with + git gc && + echo "0 objects, 0 kilobytes" >expect && + git count-objects >actual && + test_cmp expect actual && + + # Ensure merge is successful (exit code of 0) + git merge-tree --write-tree --quiet side1 side3 >output && + + # Ensure there is no output + test_must_be_empty output && + + # Ensure no loose objects written (all new objects written would have + # been in "outer layer" of the merge) + git count-objects >actual && + test_cmp expect actual +' + test_expect_success 'Clean merge' ' TREE_OID=$(git merge-tree --write-tree side1 side3) && q_to_tab <<-EOF >expect && @@ -72,6 +91,25 @@ test_expect_success 'Failed merge without rename detection' ' grep "CONFLICT (modify/delete): numbers deleted" out ' +test_expect_success '--quiet on conflicted merge' ' + # Get rid of loose objects to start with + git gc && + echo "0 objects, 0 kilobytes" >expect && + git count-objects >actual && + test_cmp expect actual && + + # Ensure merge has conflict + test_expect_code 1 git merge-tree --write-tree --quiet side1 side2 >output && + + # Ensure there is no output + test_must_be_empty output && + + # Ensure no loose objects written (all new objects written would have + # been in "outer layer" of the merge) + git count-objects >actual && + test_cmp expect actual +' + test_expect_success 'Content merge and a few conflicts' ' git checkout side1^0 && test_must_fail git merge side2 && -- gitgitgadget ^ permalink raw reply related [flat|nested] 29+ messages in thread
* Re: [PATCH v4 2/2] merge-tree: add a new --quiet flag 2025-05-16 20:04 ` [PATCH v4 2/2] merge-tree: add a new --quiet flag Elijah Newren via GitGitGadget @ 2025-05-17 19:52 ` Kristoffer Haugsbakk 2025-05-17 19:57 ` Kristoffer Haugsbakk 0 siblings, 1 reply; 29+ messages in thread From: Kristoffer Haugsbakk @ 2025-05-17 19:52 UTC (permalink / raw) To: Josh Soref, git; +Cc: Elijah Newren, Phillip Wood On Fri, May 16, 2025, at 22:04, Elijah Newren via GitGitGadget wrote: > From: Elijah Newren <newren@gmail.com> > + if (quiet && o.show_messages == -1) > + o.show_messages = 0; > + o.merge_options.mergeability_only = quiet; > + die_for_incompatible_opt2(quiet, "--quiet", o.show_messages, "--messages"); > + die_for_incompatible_opt2(quiet, "--quiet", o.name_only, "--name-only"); > + die_for_incompatible_opt2(quiet, "--quiet", o.use_stdin, "--stdin"); > + die_for_incompatible_opt2(quiet, "--quiet", !line_termination, "-z"); I’ve been using git-merge-tree(1) for some scripting but only today tried out `--stdin` for printing refs that conflict. ``` # Pipe in pairs merge_pairs=$(mktemp) tee $merge_pairs \ | git merge-tree --stdin --no-messages \ | tr '\0' '\n' \ | grep --extended-regexp '^(1|0)$' \ | paste -d' ' - $merge_pairs \ | grep '^0' \ | cut -d' ' -f2- ``` (Previously I called the command in a loop) I could imagine a `--format` option to just keep one of the arguments, which means the tee(1) (for cross-referencing the ref) and all the other things are gone: ``` git merge-tree --format='%(if)%(conflicted)%(then)oid2%(end)' --stdin ``` (But imagined options aside) `--stdin` is presumably for efficiency and `--quiet`/`--dry-run` definitely is. But `--quiet` can only be used in the mode where you can only do a single merge, not in the `--stdin` batch mode. `--quiet`/`--dry-run` with informational output (c.f. the above die-for-incompatible) would “break” the documented output format since conflicts haven’t been computed all the way and there are no OIDs for successful merges. But the user is opting into a new mode here, never seen before. Can’t they opt into a new informational mode where `--stdin --quiet` can co-exist? Then you can have dry-run batch mode. -- Kristoffer Haugsbakk ^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH v4 2/2] merge-tree: add a new --quiet flag 2025-05-17 19:52 ` Kristoffer Haugsbakk @ 2025-05-17 19:57 ` Kristoffer Haugsbakk 0 siblings, 0 replies; 29+ messages in thread From: Kristoffer Haugsbakk @ 2025-05-17 19:57 UTC (permalink / raw) To: Josh Soref, git; +Cc: Elijah Newren, Phillip Wood On Sat, May 17, 2025, at 21:52, Kristoffer Haugsbakk wrote: > Can’t they opt into a new informational mode where > `--stdin --quiet` can co-exist? Yes, I now immediately see the contradiction in the literal text: “informational mode” and `--quiet`. But I shall not weigh in on the naming matter. -- Kristoffer Haugsbakk ^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH v4 0/2] merge-tree: add new --quiet option 2025-05-16 20:04 ` [PATCH v4 0/2] merge-tree: add new --quiet option Elijah Newren via GitGitGadget 2025-05-16 20:04 ` [PATCH v4 1/2] merge-ort: add a new mergeability_only option Elijah Newren via GitGitGadget 2025-05-16 20:04 ` [PATCH v4 2/2] merge-tree: add a new --quiet flag Elijah Newren via GitGitGadget @ 2025-05-19 9:05 ` Phillip Wood 2025-05-19 15:59 ` Junio C Hamano 2 siblings, 1 reply; 29+ messages in thread From: Phillip Wood @ 2025-05-19 9:05 UTC (permalink / raw) To: Elijah Newren via GitGitGadget, git; +Cc: Elijah Newren, Kristoffer Haugsbakk Hi Elijah On 16/05/2025 21:04, Elijah Newren via GitGitGadget wrote: > Changes since v3: > > * Renamed --dry-run -> --quiet . Any further naming suggestions? Thanks for re-rolling this version looks good to me. Thanks Phillip > Changes since v2: > > * Converted locations missed in v1 in changing --mergeability-only -> > --dry-run > > Changes since v1: > > * Renamed --mergeability-only flag to --dry-run, as per suggestion from > Junio > * added some commit message clarifications > > This adds a new flag, --dry-run, to git merge-tree, which suppresses all > output and leaves only the exit status (reflecting successful merge or > conflict). This is useful for Git Forges in cases where they are only > interested in whether two branches can be merged, without needing the actual > merge result or conflict details. > > The advantage of the flag is two fold: > > * The merge machinery can exit once it detects a conflict, instead of > continuing to compute merge result information > * The merge machinery can avoid writing merged blobs and trees to the > object store when in the outer layer of the merging process (more details > in the first commit message). > > Elijah Newren (2): > merge-ort: add a new mergeability_only option > merge-tree: add a new --quiet flag > > Documentation/git-merge-tree.adoc | 6 +++++ > builtin/merge-tree.c | 18 +++++++++++++++ > merge-ort.c | 38 +++++++++++++++++++++++++------ > merge-ort.h | 1 + > t/t4301-merge-tree-write-tree.sh | 38 +++++++++++++++++++++++++++++++ > 5 files changed, 94 insertions(+), 7 deletions(-) > > > base-commit: 6c0bd1fc70efaf053abe4e57c976afdc72d15377 > Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-1920%2Fnewren%2Fmergeability-only-v4 > Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-1920/newren/mergeability-only-v4 > Pull-Request: https://github.com/gitgitgadget/git/pull/1920 > > Range-diff vs v3: > > 1: 4757c4810d3 = 1: 4757c4810d3 merge-ort: add a new mergeability_only option > 2: f11824317a8 ! 2: 7c40d3c9216 merge-tree: add a new --dry-run flag > @@ Metadata > Author: Elijah Newren <newren@gmail.com> > > ## Commit message ## > - merge-tree: add a new --dry-run flag > + merge-tree: add a new --quiet flag > > Git Forges may be interested in whether two branches can be merged while > not being interested in what the resulting merge tree is nor which files > - conflicted. For such cases, add a new --dry-run flag which > + conflicted. For such cases, add a new --quiet flag which > will make use of the new mergeability_only flag added to merge-ort in > the previous commit. This option allows the merge machinery to, in the > outer layer of the merge: > @@ Documentation/git-merge-tree.adoc: OPTIONS > default is to include these messages if there are merge > conflicts, and to omit them otherwise. > > -+--dry-run:: > ++--quiet:: > + Disable all output from the program. Useful when you are only > + interested in the exit status. Allows merge-tree to exit > + early when it finds a conflict, and allows it to avoid writing > @@ builtin/merge-tree.c: int cmd_merge_tree(int argc, > int original_argc; > const char *merge_base = NULL; > int ret; > -+ int dry_run = 0; > ++ int quiet = 0; > > const char * const merge_tree_usage[] = { > N_("git merge-tree [--write-tree] [<options>] <branch1> <branch2>"), > @@ builtin/merge-tree.c: int cmd_merge_tree(int argc, > N_("do a trivial merge only"), MODE_TRIVIAL), > OPT_BOOL(0, "messages", &o.show_messages, > N_("also show informational/conflict messages")), > -+ OPT_BOOL_F(0, "dry-run", > -+ &dry_run, > ++ OPT_BOOL_F(0, "quiet", > ++ &quiet, > + N_("suppress all output; only exit status wanted"), > + PARSE_OPT_NONEG), > OPT_SET_INT('z', NULL, &line_termination, > @@ builtin/merge-tree.c: int cmd_merge_tree(int argc, > argc = parse_options(argc, argv, prefix, mt_options, > merge_tree_usage, PARSE_OPT_STOP_AT_NON_OPTION); > > -+ if (dry_run && o.show_messages == -1) > ++ if (quiet && o.show_messages == -1) > + o.show_messages = 0; > -+ o.merge_options.mergeability_only = dry_run; > -+ die_for_incompatible_opt2(dry_run, "--dry-run", > -+ o.show_messages, "--messages"); > -+ die_for_incompatible_opt2(dry_run, "--dry-run", > -+ o.name_only, "--name-only"); > -+ die_for_incompatible_opt2(dry_run, "--dry-run", > -+ o.use_stdin, "--stdin"); > -+ die_for_incompatible_opt2(dry_run, "--dry-run", > -+ !line_termination, "-z"); > ++ o.merge_options.mergeability_only = quiet; > ++ die_for_incompatible_opt2(quiet, "--quiet", o.show_messages, "--messages"); > ++ die_for_incompatible_opt2(quiet, "--quiet", o.name_only, "--name-only"); > ++ die_for_incompatible_opt2(quiet, "--quiet", o.use_stdin, "--stdin"); > ++ die_for_incompatible_opt2(quiet, "--quiet", !line_termination, "-z"); > + > if (xopts.nr && o.mode == MODE_TRIVIAL) > die(_("--trivial-merge is incompatible with all other options")); > @@ t/t4301-merge-tree-write-tree.sh: test_expect_success setup ' > git commit -m first-commit > ' > > -+test_expect_success '--dry-run on clean merge' ' > ++test_expect_success '--quiet on clean merge' ' > + # Get rid of loose objects to start with > + git gc && > + echo "0 objects, 0 kilobytes" >expect && > @@ t/t4301-merge-tree-write-tree.sh: test_expect_success setup ' > + test_cmp expect actual && > + > + # Ensure merge is successful (exit code of 0) > -+ git merge-tree --write-tree --dry-run side1 side3 >output && > ++ git merge-tree --write-tree --quiet side1 side3 >output && > + > + # Ensure there is no output > + test_must_be_empty output && > @@ t/t4301-merge-tree-write-tree.sh: test_expect_success 'Failed merge without rena > grep "CONFLICT (modify/delete): numbers deleted" out > ' > > -+test_expect_success '--dry-run on conflicted merge' ' > ++test_expect_success '--quiet on conflicted merge' ' > + # Get rid of loose objects to start with > + git gc && > + echo "0 objects, 0 kilobytes" >expect && > @@ t/t4301-merge-tree-write-tree.sh: test_expect_success 'Failed merge without rena > + test_cmp expect actual && > + > + # Ensure merge has conflict > -+ test_expect_code 1 git merge-tree --write-tree --dry-run side1 side2 >output && > ++ test_expect_code 1 git merge-tree --write-tree --quiet side1 side2 >output && > + > + # Ensure there is no output > + test_must_be_empty output && > ^ permalink raw reply [flat|nested] 29+ messages in thread
* Re: [PATCH v4 0/2] merge-tree: add new --quiet option 2025-05-19 9:05 ` [PATCH v4 0/2] merge-tree: add new --quiet option Phillip Wood @ 2025-05-19 15:59 ` Junio C Hamano 0 siblings, 0 replies; 29+ messages in thread From: Junio C Hamano @ 2025-05-19 15:59 UTC (permalink / raw) To: Phillip Wood Cc: Elijah Newren via GitGitGadget, git, Elijah Newren, Kristoffer Haugsbakk Phillip Wood <phillip.wood123@gmail.com> writes: > Hi Elijah > > On 16/05/2025 21:04, Elijah Newren via GitGitGadget wrote: >> Changes since v3: >> * Renamed --dry-run -> --quiet . Any further naming suggestions? > > Thanks for re-rolling this version looks good to me. > > Thanks > > Phillip Thanks, all. Let's mark it for 'next' and merge it down. ^ permalink raw reply [flat|nested] 29+ messages in thread
end of thread, other threads:[~2025-05-19 15:59 UTC | newest] Thread overview: 29+ messages (download: mbox.gz follow: Atom feed -- links below jump to the message on this page -- 2025-05-10 22:02 [PATCH 0/2] merge-tree: add new --mergeability-only option Elijah Newren via GitGitGadget 2025-05-10 22:02 ` [PATCH 1/2] merge-ort: add a new mergeability_only option Elijah Newren via GitGitGadget 2025-05-10 22:02 ` [PATCH 2/2] merge-tree: add a new --mergeability-only flag Elijah Newren via GitGitGadget 2025-05-12 17:04 ` [PATCH 0/2] merge-tree: add new --mergeability-only option Junio C Hamano 2025-05-12 17:41 ` Elijah Newren 2025-05-12 18:27 ` Junio C Hamano 2025-05-12 18:37 ` Elijah Newren 2025-05-12 23:42 ` [PATCH v2 0/2] merge-tree: add new --dry-run option Elijah Newren via GitGitGadget 2025-05-12 23:42 ` [PATCH v2 1/2] merge-ort: add a new mergeability_only option Elijah Newren via GitGitGadget 2025-05-12 23:42 ` [PATCH v2 2/2] merge-tree: add a new --dry-run flag Elijah Newren via GitGitGadget 2025-05-13 7:15 ` Kristoffer Haugsbakk 2025-05-13 15:28 ` Elijah Newren 2025-05-13 13:24 ` Junio C Hamano 2025-05-13 15:30 ` Elijah Newren 2025-05-14 14:08 ` Junio C Hamano 2025-05-14 0:24 ` [PATCH v3 0/2] merge-tree: add new --dry-run option Elijah Newren via GitGitGadget 2025-05-14 0:24 ` [PATCH v3 1/2] merge-ort: add a new mergeability_only option Elijah Newren via GitGitGadget 2025-05-14 0:24 ` [PATCH v3 2/2] merge-tree: add a new --dry-run flag Elijah Newren via GitGitGadget 2025-05-15 13:07 ` Junio C Hamano 2025-05-16 13:18 ` Phillip Wood 2025-05-16 16:03 ` Elijah Newren 2025-05-14 15:34 ` [PATCH v3 0/2] merge-tree: add new --dry-run option Kristoffer Haugsbakk 2025-05-16 20:04 ` [PATCH v4 0/2] merge-tree: add new --quiet option Elijah Newren via GitGitGadget 2025-05-16 20:04 ` [PATCH v4 1/2] merge-ort: add a new mergeability_only option Elijah Newren via GitGitGadget 2025-05-16 20:04 ` [PATCH v4 2/2] merge-tree: add a new --quiet flag Elijah Newren via GitGitGadget 2025-05-17 19:52 ` Kristoffer Haugsbakk 2025-05-17 19:57 ` Kristoffer Haugsbakk 2025-05-19 9:05 ` [PATCH v4 0/2] merge-tree: add new --quiet option Phillip Wood 2025-05-19 15:59 ` Junio C Hamano
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).