# [PATCH 1/4] lib-t6000.sh: write tag using git-update-ref 13 messages from 2020-06-30 to 2020-07-10. Participants: Han-Wen Nienhuys via GitGitGadget, Junio C Hamano, Junio C Hamano via GitGitGadget. Thread: https://gitlist.dev/t/53774 ## Han-Wen Nienhuys via GitGitGadget, 2020-06-30 12:05 Subject: [PATCH 1/4] lib-t6000.sh: write tag using git-update-ref Message-ID: URL: https://gitlist.dev/e/b7b2ad8e79efe534c48947da5ac83835df6fb73d.1593518738.git.gitgitgadget%40gmail.com In-Reply-To: ``` From: Han-Wen Nienhuys Signed-off-by: Han-Wen Nienhuys --- t/lib-t6000.sh | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/t/lib-t6000.sh b/t/lib-t6000.sh index b0ed4767e3..fba6778ca3 100644 --- a/t/lib-t6000.sh +++ b/t/lib-t6000.sh @@ -1,7 +1,5 @@ : included from 6002 and others -mkdir -p .git/refs/tags - >sed.script # Answer the sha1 has associated with the tag. The tag must exist under refs/tags @@ -26,7 +24,8 @@ save_tag () { _tag=$1 test -n "$_tag" || error "usage: save_tag tag commit-args ..." shift 1 - "$@" >".git/refs/tags/$_tag" + + git update-ref "refs/tags/$_tag" $("$@") echo "s/$(tag $_tag)/$_tag/g" >sed.script.tmp cat sed.script >>sed.script.tmp -- gitgitgadget ``` ## Han-Wen Nienhuys via GitGitGadget, 2020-06-30 12:05 Subject: [PATCH 0/4] Preliminary fixes for reftable support Message-ID: URL: https://gitlist.dev/e/pull.669.git.1593518738.gitgitgadget%40gmail.com ``` These are assorted small fixes split out from the reftable topic. Han-Wen Nienhuys (4): lib-t6000.sh: write tag using git-update-ref t3432: use git-reflog to inspect the reflog for HEAD checkout: add '\n' to reflog message Treat BISECT_HEAD as a pseudo ref builtin/bisect--helper.c | 3 +-- builtin/checkout.c | 5 +++-- git-bisect.sh | 4 ++-- t/lib-t6000.sh | 5 ++--- t/t3432-rebase-fast-forward.sh | 7 ++++--- 5 files changed, 12 insertions(+), 12 deletions(-) base-commit: f402ea68166bd77f09b176c96005ac7f8886e14b Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-669%2Fhanwen%2Fpreliminaries-v1 Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-669/hanwen/preliminaries-v1 Pull-Request: https://github.com/gitgitgadget/git/pull/669 -- gitgitgadget ``` ## Han-Wen Nienhuys via GitGitGadget, 2020-06-30 12:05 Subject: [PATCH 2/4] t3432: use git-reflog to inspect the reflog for HEAD Message-ID: URL: https://gitlist.dev/e/f238d1d7f8b59832eda48e47188479a2230a5927.1593518738.git.gitgitgadget%40gmail.com In-Reply-To: ``` From: Han-Wen Nienhuys Signed-off-by: Han-Wen Nienhuys --- t/t3432-rebase-fast-forward.sh | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/t/t3432-rebase-fast-forward.sh b/t/t3432-rebase-fast-forward.sh index 6f0452c0ea..22afeb8ccd 100755 --- a/t/t3432-rebase-fast-forward.sh +++ b/t/t3432-rebase-fast-forward.sh @@ -60,15 +60,16 @@ test_rebase_same_head_ () { fi && oldhead=\$(git rev-parse HEAD) && test_when_finished 'git reset --hard \$oldhead' && - cp .git/logs/HEAD expect && + git reflog HEAD > expect && git rebase$flag $* >stdout && + git reflog HEAD > actual && if test $what = work then old=\$(wc -l URL: https://gitlist.dev/e/8a62cc26686b1bb2f04cb01c79e336dca397a330.1593518738.git.gitgitgadget%40gmail.com In-Reply-To: ``` From: Han-Wen Nienhuys This makes check consistent with the rest of the git code, which tends to use a trailing '\n' when writing reflog messages. (The storage layer adds a '\n' when it is missing.) Signed-off-by: Han-Wen Nienhuys --- builtin/checkout.c | 5 +++-- 1 file changed, 3 insertions(+), 2 deletions(-) diff --git a/builtin/checkout.c b/builtin/checkout.c index af849c644f..bb11fcc4e9 100644 --- a/builtin/checkout.c +++ b/builtin/checkout.c @@ -884,8 +884,9 @@ static void update_refs_for_switch(const struct checkout_opts *opts, reflog_msg = getenv("GIT_REFLOG_ACTION"); if (!reflog_msg) - strbuf_addf(&msg, "checkout: moving from %s to %s", - old_desc ? old_desc : "(invalid)", new_branch_info->name); + strbuf_addf(&msg, "checkout: moving from %s to %s\n", + old_desc ? old_desc : "(invalid)", + new_branch_info->name); else strbuf_insertstr(&msg, 0, reflog_msg); -- gitgitgadget ``` ## Han-Wen Nienhuys via GitGitGadget, 2020-06-30 12:05 Subject: [PATCH 4/4] Treat BISECT_HEAD as a pseudo ref Message-ID: <2b7eb58c1556971d128fc7596bf8abf7a993432d.1593518738.git.gitgitgadget@gmail.com> URL: https://gitlist.dev/e/2b7eb58c1556971d128fc7596bf8abf7a993432d.1593518738.git.gitgitgadget%40gmail.com In-Reply-To: ``` From: Han-Wen Nienhuys Both the git-bisect.sh as bisect--helper inspected the file system directly. Signed-off-by: Han-Wen Nienhuys --- builtin/bisect--helper.c | 3 +-- git-bisect.sh | 4 ++-- 2 files changed, 3 insertions(+), 4 deletions(-) diff --git a/builtin/bisect--helper.c b/builtin/bisect--helper.c index ec4996282e..73f9324ad7 100644 --- a/builtin/bisect--helper.c +++ b/builtin/bisect--helper.c @@ -13,7 +13,6 @@ static GIT_PATH_FUNC(git_path_bisect_terms, "BISECT_TERMS") static GIT_PATH_FUNC(git_path_bisect_expected_rev, "BISECT_EXPECTED_REV") static GIT_PATH_FUNC(git_path_bisect_ancestors_ok, "BISECT_ANCESTORS_OK") static GIT_PATH_FUNC(git_path_bisect_start, "BISECT_START") -static GIT_PATH_FUNC(git_path_bisect_head, "BISECT_HEAD") static GIT_PATH_FUNC(git_path_bisect_log, "BISECT_LOG") static GIT_PATH_FUNC(git_path_head_name, "head-name") static GIT_PATH_FUNC(git_path_bisect_names, "BISECT_NAMES") @@ -164,7 +163,7 @@ static int bisect_reset(const char *commit) strbuf_addstr(&branch, commit); } - if (!file_exists(git_path_bisect_head())) { + if (!ref_exists("BISECT_HEAD")) { struct argv_array argv = ARGV_ARRAY_INIT; argv_array_pushl(&argv, "checkout", branch.buf, "--", NULL); diff --git a/git-bisect.sh b/git-bisect.sh index 08a6ed57dd..f03fbb18f0 100755 --- a/git-bisect.sh +++ b/git-bisect.sh @@ -41,7 +41,7 @@ TERM_GOOD=good bisect_head() { - if test -f "$GIT_DIR/BISECT_HEAD" + if git rev-parse --verify -q BISECT_HEAD > /dev/null then echo BISECT_HEAD else @@ -153,7 +153,7 @@ bisect_next() { git bisect--helper --bisect-next-check $TERM_GOOD $TERM_BAD $TERM_GOOD|| exit # Perform all bisection computation, display and checkout - git bisect--helper --next-all $(test -f "$GIT_DIR/BISECT_HEAD" && echo --no-checkout) + git bisect--helper --next-all $(git rev-parse --verify -q BISECT_HEAD > /dev/null && echo --no-checkout) res=$? # Check if we should exit because bisection is finished -- gitgitgadget ``` ## Junio C Hamano, 2020-07-07 04:40 Subject: Re: [PATCH 0/4] Preliminary fixes for reftable support Message-ID: URL: https://gitlist.dev/e/xmqqmu4c6iqg.fsf%40gitster.c.googlers.com In-Reply-To: ``` "Han-Wen Nienhuys via GitGitGadget" writes: > These are assorted small fixes split out from the reftable topic. > > Han-Wen Nienhuys (4): > lib-t6000.sh: write tag using git-update-ref > t3432: use git-reflog to inspect the reflog for HEAD > checkout: add '\n' to reflog message > Treat BISECT_HEAD as a pseudo ref > > builtin/bisect--helper.c | 3 +-- > builtin/checkout.c | 5 +++-- > git-bisect.sh | 4 ++-- > t/lib-t6000.sh | 5 ++--- > t/t3432-rebase-fast-forward.sh | 7 ++++--- > 5 files changed, 12 insertions(+), 12 deletions(-) Thanks for splitting these out, and sorry for not being able to get to them earlier. Will queue. ``` ## Han-Wen Nienhuys via GitGitGadget, 2020-07-10 17:19 Subject: [PATCH v2 0/4] Preliminary fixes for reftable support Message-ID: URL: https://gitlist.dev/e/pull.669.v2.git.1594401593.gitgitgadget%40gmail.com In-Reply-To: ``` These are assorted small fixes split out from the reftable topic. Han-Wen Nienhuys (3): lib-t6000.sh: write tag using git-update-ref t3432: use git-reflog to inspect the reflog for HEAD Treat BISECT_HEAD as a pseudo ref Junio C Hamano (1): Cleanse reflog messages in the refs.c layer builtin/bisect--helper.c | 3 +- git-bisect.sh | 4 +-- refs.c | 50 ++++++++++++++++++++++++++++------ refs/files-backend.c | 2 +- refs/refs-internal.h | 6 ---- t/lib-t6000.sh | 5 ++-- t/t3432-rebase-fast-forward.sh | 7 +++-- 7 files changed, 51 insertions(+), 26 deletions(-) base-commit: bd42bbe1a46c0fe486fc33e82969275e27e4dc19 Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-669%2Fhanwen%2Fpreliminaries-v2 Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-669/hanwen/preliminaries-v2 Pull-Request: https://github.com/gitgitgadget/git/pull/669 Range-diff vs v1: 1: b7b2ad8e79 = 1: 0f244ed6cb lib-t6000.sh: write tag using git-update-ref 2: f238d1d7f8 ! 2: 123d246edf t3432: use git-reflog to inspect the reflog for HEAD @@ t/t3432-rebase-fast-forward.sh: test_rebase_same_head_ () { oldhead=\$(git rev-parse HEAD) && test_when_finished 'git reset --hard \$oldhead' && - cp .git/logs/HEAD expect && -+ git reflog HEAD > expect && ++ git reflog HEAD >expect && git rebase$flag $* >stdout && -+ git reflog HEAD > actual && ++ git reflog HEAD >actual && if test $what = work then old=\$(wc -l 4: 6ca5b99c8d Cleanse reflog messages in the refs.c layer -- gitgitgadget ``` ## Han-Wen Nienhuys via GitGitGadget, 2020-07-10 17:19 Subject: [PATCH v2 1/4] lib-t6000.sh: write tag using git-update-ref Message-ID: <0f244ed6cbddd0673653349b78c3ede31f9bbf71.1594401593.git.gitgitgadget@gmail.com> URL: https://gitlist.dev/e/0f244ed6cbddd0673653349b78c3ede31f9bbf71.1594401593.git.gitgitgadget%40gmail.com In-Reply-To: ``` From: Han-Wen Nienhuys Signed-off-by: Han-Wen Nienhuys --- t/lib-t6000.sh | 5 ++--- 1 file changed, 2 insertions(+), 3 deletions(-) diff --git a/t/lib-t6000.sh b/t/lib-t6000.sh index b0ed4767e3..fba6778ca3 100644 --- a/t/lib-t6000.sh +++ b/t/lib-t6000.sh @@ -1,7 +1,5 @@ : included from 6002 and others -mkdir -p .git/refs/tags - >sed.script # Answer the sha1 has associated with the tag. The tag must exist under refs/tags @@ -26,7 +24,8 @@ save_tag () { _tag=$1 test -n "$_tag" || error "usage: save_tag tag commit-args ..." shift 1 - "$@" >".git/refs/tags/$_tag" + + git update-ref "refs/tags/$_tag" $("$@") echo "s/$(tag $_tag)/$_tag/g" >sed.script.tmp cat sed.script >>sed.script.tmp -- gitgitgadget ``` ## Han-Wen Nienhuys via GitGitGadget, 2020-07-10 17:19 Subject: [PATCH v2 3/4] Treat BISECT_HEAD as a pseudo ref Message-ID: URL: https://gitlist.dev/e/d4007c2a5b785aac54d3ce7635cd6a50225c1361.1594401593.git.gitgitgadget%40gmail.com In-Reply-To: ``` From: Han-Wen Nienhuys Both the git-bisect.sh as bisect--helper inspected the file system directly. Signed-off-by: Han-Wen Nienhuys --- builtin/bisect--helper.c | 3 +-- git-bisect.sh | 4 ++-- 2 files changed, 3 insertions(+), 4 deletions(-) diff --git a/builtin/bisect--helper.c b/builtin/bisect--helper.c index ec4996282e..73f9324ad7 100644 --- a/builtin/bisect--helper.c +++ b/builtin/bisect--helper.c @@ -13,7 +13,6 @@ static GIT_PATH_FUNC(git_path_bisect_terms, "BISECT_TERMS") static GIT_PATH_FUNC(git_path_bisect_expected_rev, "BISECT_EXPECTED_REV") static GIT_PATH_FUNC(git_path_bisect_ancestors_ok, "BISECT_ANCESTORS_OK") static GIT_PATH_FUNC(git_path_bisect_start, "BISECT_START") -static GIT_PATH_FUNC(git_path_bisect_head, "BISECT_HEAD") static GIT_PATH_FUNC(git_path_bisect_log, "BISECT_LOG") static GIT_PATH_FUNC(git_path_head_name, "head-name") static GIT_PATH_FUNC(git_path_bisect_names, "BISECT_NAMES") @@ -164,7 +163,7 @@ static int bisect_reset(const char *commit) strbuf_addstr(&branch, commit); } - if (!file_exists(git_path_bisect_head())) { + if (!ref_exists("BISECT_HEAD")) { struct argv_array argv = ARGV_ARRAY_INIT; argv_array_pushl(&argv, "checkout", branch.buf, "--", NULL); diff --git a/git-bisect.sh b/git-bisect.sh index 08a6ed57dd..f03fbb18f0 100755 --- a/git-bisect.sh +++ b/git-bisect.sh @@ -41,7 +41,7 @@ TERM_GOOD=good bisect_head() { - if test -f "$GIT_DIR/BISECT_HEAD" + if git rev-parse --verify -q BISECT_HEAD > /dev/null then echo BISECT_HEAD else @@ -153,7 +153,7 @@ bisect_next() { git bisect--helper --bisect-next-check $TERM_GOOD $TERM_BAD $TERM_GOOD|| exit # Perform all bisection computation, display and checkout - git bisect--helper --next-all $(test -f "$GIT_DIR/BISECT_HEAD" && echo --no-checkout) + git bisect--helper --next-all $(git rev-parse --verify -q BISECT_HEAD > /dev/null && echo --no-checkout) res=$? # Check if we should exit because bisection is finished -- gitgitgadget ``` ## Han-Wen Nienhuys via GitGitGadget, 2020-07-10 17:19 Subject: [PATCH v2 2/4] t3432: use git-reflog to inspect the reflog for HEAD Message-ID: <123d246edf9e3e40c321cfaff81fe96ae3a0c015.1594401593.git.gitgitgadget@gmail.com> URL: https://gitlist.dev/e/123d246edf9e3e40c321cfaff81fe96ae3a0c015.1594401593.git.gitgitgadget%40gmail.com In-Reply-To: ``` From: Han-Wen Nienhuys Signed-off-by: Han-Wen Nienhuys --- t/t3432-rebase-fast-forward.sh | 7 ++++--- 1 file changed, 4 insertions(+), 3 deletions(-) diff --git a/t/t3432-rebase-fast-forward.sh b/t/t3432-rebase-fast-forward.sh index 6f0452c0ea..a29eda87e9 100755 --- a/t/t3432-rebase-fast-forward.sh +++ b/t/t3432-rebase-fast-forward.sh @@ -60,15 +60,16 @@ test_rebase_same_head_ () { fi && oldhead=\$(git rev-parse HEAD) && test_when_finished 'git reset --hard \$oldhead' && - cp .git/logs/HEAD expect && + git reflog HEAD >expect && git rebase$flag $* >stdout && + git reflog HEAD >actual && if test $what = work then old=\$(wc -l URL: https://gitlist.dev/e/6ca5b99c8d7af1a3f35b3a7d25db284c879a2f10.1594401593.git.gitgitgadget%40gmail.com In-Reply-To: ``` From: Junio C Hamano Regarding reflog messages: - We expect that a reflog message consists of a single line. The file format used by the files backend may add a LF after the message as a delimiter, and output by commands like "git log -g" may complete such an incomplete line by adding a LF at the end, but philosophically, the terminating LF is not a part of the message. - We however allow callers of refs API to supply a random sequence of NUL terminated bytes. We cleanse caller-supplied message by squashing a run of whitespaces into a SP, and by trimming trailing whitespace, before storing the message. This is how we tolerate, instead of erring out, a message with LF in it (be it at the end, in the middle, or both). Currently, the cleansing of the reflog message is done by the files backend, before the log is written out. This is sufficient with the current code, as that is the only backend that writes reflogs. But new backends can be added that write reflogs, and we'd want the resulting log message we would read out of "log -g" the same no matter what backend is used. An added benefit is that the "cleansing" function could be updated later, independent from individual backends, to e.g. allow multi-line log messages if we wanted to, and when that happens, it would help a lot to ensure we covered all bases if the cleansing function (which would be updated) is called from the generic layer. Side note: I am not interested in supporting multi-line reflog messages right at the moment (nobody is asking for it), but I envision that instead of the "squash a run of whitespaces into a SP and rtrim" cleansing, we can %urlencode problematic bytes in the message *AND* append a SP at the end, when a new version of Git that supports multi-line and/or verbatim reflog messages writes a reflog record. The reading side can detect the presense of SP at the end (which should have been rtrimmed out if it were written by existing versions of Git) as a signal that decoding %urlencode recovers the original reflog message. Signed-off-by: Han-Wen Nienhuys --- refs.c | 50 ++++++++++++++++++++++++++++++++++++-------- refs/files-backend.c | 2 +- refs/refs-internal.h | 6 ------ 3 files changed, 42 insertions(+), 16 deletions(-) diff --git a/refs.c b/refs.c index 639cba93b4..89814c7be4 100644 --- a/refs.c +++ b/refs.c @@ -902,7 +902,7 @@ int delete_ref(const char *msg, const char *refname, old_oid, flags); } -void copy_reflog_msg(struct strbuf *sb, const char *msg) +static void copy_reflog_msg(struct strbuf *sb, const char *msg) { char c; int wasspace = 1; @@ -919,6 +919,15 @@ void copy_reflog_msg(struct strbuf *sb, const char *msg) strbuf_rtrim(sb); } +static char *normalize_reflog_message(const char *msg) +{ + struct strbuf sb = STRBUF_INIT; + + if (msg && *msg) + copy_reflog_msg(&sb, msg); + return strbuf_detach(&sb, NULL); +} + int should_autocreate_reflog(const char *refname) { switch (log_all_ref_updates) { @@ -1124,7 +1133,7 @@ struct ref_update *ref_transaction_add_update( oidcpy(&update->new_oid, new_oid); if (flags & REF_HAVE_OLD) oidcpy(&update->old_oid, old_oid); - update->msg = xstrdup_or_null(msg); + update->msg = normalize_reflog_message(msg); return update; } @@ -1983,9 +1992,14 @@ int refs_create_symref(struct ref_store *refs, const char *refs_heads_master, const char *logmsg) { - return refs->be->create_symref(refs, ref_target, - refs_heads_master, - logmsg); + char *msg; + int retval; + + msg = normalize_reflog_message(logmsg); + retval = refs->be->create_symref(refs, ref_target, refs_heads_master, + msg); + free(msg); + return retval; } int create_symref(const char *ref_target, const char *refs_heads_master, @@ -2370,10 +2384,16 @@ int initial_ref_transaction_commit(struct ref_transaction *transaction, return refs->be->initial_transaction_commit(refs, transaction, err); } -int refs_delete_refs(struct ref_store *refs, const char *msg, +int refs_delete_refs(struct ref_store *refs, const char *logmsg, struct string_list *refnames, unsigned int flags) { - return refs->be->delete_refs(refs, msg, refnames, flags); + char *msg; + int retval; + + msg = normalize_reflog_message(logmsg); + retval = refs->be->delete_refs(refs, msg, refnames, flags); + free(msg); + return retval; } int delete_refs(const char *msg, struct string_list *refnames, @@ -2385,7 +2405,13 @@ int delete_refs(const char *msg, struct string_list *refnames, int refs_rename_ref(struct ref_store *refs, const char *oldref, const char *newref, const char *logmsg) { - return refs->be->rename_ref(refs, oldref, newref, logmsg); + char *msg; + int retval; + + msg = normalize_reflog_message(logmsg); + retval = refs->be->rename_ref(refs, oldref, newref, msg); + free(msg); + return retval; } int rename_ref(const char *oldref, const char *newref, const char *logmsg) @@ -2396,7 +2422,13 @@ int rename_ref(const char *oldref, const char *newref, const char *logmsg) int refs_copy_existing_ref(struct ref_store *refs, const char *oldref, const char *newref, const char *logmsg) { - return refs->be->copy_ref(refs, oldref, newref, logmsg); + char *msg; + int retval; + + msg = normalize_reflog_message(logmsg); + retval = refs->be->copy_ref(refs, oldref, newref, msg); + free(msg); + return retval; } int copy_existing_ref(const char *oldref, const char *newref, const char *logmsg) diff --git a/refs/files-backend.c b/refs/files-backend.c index 6516c7bc8c..e0aba23eb2 100644 --- a/refs/files-backend.c +++ b/refs/files-backend.c @@ -1629,7 +1629,7 @@ static int log_ref_write_fd(int fd, const struct object_id *old_oid, strbuf_addf(&sb, "%s %s %s", oid_to_hex(old_oid), oid_to_hex(new_oid), committer); if (msg && *msg) - copy_reflog_msg(&sb, msg); + strbuf_addstr(&sb, msg); strbuf_addch(&sb, '\n'); if (write_in_full(fd, sb.buf, sb.len) < 0) ret = -1; diff --git a/refs/refs-internal.h b/refs/refs-internal.h index 4271362d26..357359a0be 100644 --- a/refs/refs-internal.h +++ b/refs/refs-internal.h @@ -96,12 +96,6 @@ enum peel_status { */ enum peel_status peel_object(const struct object_id *name, struct object_id *oid); -/* - * Copy the reflog message msg to sb while cleaning up the whitespaces. - * Especially, convert LF to space, because reflog file is one line per entry. - */ -void copy_reflog_msg(struct strbuf *sb, const char *msg); - /** * Information needed for a single ref update. Set new_oid to the new * value or to null_oid to delete the ref. To check the old value -- gitgitgadget ``` ## Junio C Hamano, 2020-07-10 20:57 Subject: Re: [PATCH v2 0/4] Preliminary fixes for reftable support Message-ID: URL: https://gitlist.dev/e/xmqqy2nrrsuy.fsf%40gitster.c.googlers.com In-Reply-To: ``` "Han-Wen Nienhuys via GitGitGadget" writes: > These are assorted small fixes split out from the reftable topic. Will replace. I'll retitle the 3/4 and 4/4 to bisect: treat BISECT_HEAD as a pseudo ref reflgo: cleanse messages in the refs.c layer while queuing. Thanks. > Han-Wen Nienhuys (3): > lib-t6000.sh: write tag using git-update-ref > t3432: use git-reflog to inspect the reflog for HEAD > Treat BISECT_HEAD as a pseudo ref > > Junio C Hamano (1): > Cleanse reflog messages in the refs.c layer > > builtin/bisect--helper.c | 3 +- > git-bisect.sh | 4 +-- > refs.c | 50 ++++++++++++++++++++++++++++------ > refs/files-backend.c | 2 +- > refs/refs-internal.h | 6 ---- > t/lib-t6000.sh | 5 ++-- > t/t3432-rebase-fast-forward.sh | 7 +++-- > 7 files changed, 51 insertions(+), 26 deletions(-) > > > base-commit: bd42bbe1a46c0fe486fc33e82969275e27e4dc19 > Published-As: https://github.com/gitgitgadget/git/releases/tag/pr-669%2Fhanwen%2Fpreliminaries-v2 > Fetch-It-Via: git fetch https://github.com/gitgitgadget/git pr-669/hanwen/preliminaries-v2 > Pull-Request: https://github.com/gitgitgadget/git/pull/669 > > Range-diff vs v1: > > 1: b7b2ad8e79 = 1: 0f244ed6cb lib-t6000.sh: write tag using git-update-ref > 2: f238d1d7f8 ! 2: 123d246edf t3432: use git-reflog to inspect the reflog for HEAD > @@ t/t3432-rebase-fast-forward.sh: test_rebase_same_head_ () { > oldhead=\$(git rev-parse HEAD) && > test_when_finished 'git reset --hard \$oldhead' && > - cp .git/logs/HEAD expect && > -+ git reflog HEAD > expect && > ++ git reflog HEAD >expect && > git rebase$flag $* >stdout && > -+ git reflog HEAD > actual && > ++ git reflog HEAD >actual && > if test $what = work > then > old=\$(wc -l 3: 8a62cc2668 < -: ---------- checkout: add '\n' to reflog message > 4: 2b7eb58c15 = 3: d4007c2a5b Treat BISECT_HEAD as a pseudo ref > -: ---------- > 4: 6ca5b99c8d Cleanse reflog messages in the refs.c layer ``` ## Junio C Hamano, 2020-07-10 21:06 Subject: Re: [PATCH v2 4/4] Cleanse reflog messages in the refs.c layer Message-ID: URL: https://gitlist.dev/e/xmqqtuyfrsg0.fsf%40gitster.c.googlers.com In-Reply-To: <6ca5b99c8d7af1a3f35b3a7d25db284c879a2f10.1594401593.git.gitgitgadget@gmail.com> ``` "Junio C Hamano via GitGitGadget" writes: > Currently, the cleansing of the reflog message is done by the files > backend, before the log is written out. This is sufficient with the > current code, as that is the only backend that writes reflogs. But > new backends can be added that write reflogs, and we'd want the > resulting log message we would read out of "log -g" the same no > matter what backend is used. I _think_ there is no correctness impact, but this slightly changes the semantics. We used to take whatever random string from the caller and at the lowest layer of the files backend, i.e. > strbuf_addf(&sb, "%s %s %s", oid_to_hex(old_oid), oid_to_hex(new_oid), committer); > if (msg && *msg) > - copy_reflog_msg(&sb, msg); > + strbuf_addstr(&sb, msg); cleansed and added the message to strbuf. With the updated code, normalize_reflog_message() is defined like so > +static char *normalize_reflog_message(const char *msg) > +{ > + struct strbuf sb = STRBUF_INIT; > + > + if (msg && *msg) > + copy_reflog_msg(&sb, msg); > + return strbuf_detach(&sb, NULL); > +} to always return a non-NULL but an empty string, even if the caller gave us a NULL pointer. We always pass a non-NULL string around, and the updated low-level code avoids calling strbuf_addstr() because msg is not NULL, but *msg is '\0'. We _might_ be able to optimize by tweaking the normalizer a bit further, perhaps like so: static char *normalize_reflog_message(const char *msg) { struct strbuf sb = STRBUF_INIT; if (msg && *msg) copy_reflog_msg(&sb, msg); if (sb.len) { return strbuf_detach(&sb, NULL); } else { strbuf_release(&sb); return NULL; } } to avoid allocating 1 byte '\0' on the heap by strbuf_detach() when sb here ends up being truly empty. But it would be a premature optimization to worry about such a thing at this point, I think. Thanks. ```