git-fetch, git-branch: Support local --track via a special remote `.'
PATCH 1/1 git-fetch, git-branch: Support local --track via a special remote `.'
This patch adds support for a dummy remote `.' to avoid having to declare a fake remote like
[remote "local"]
url = .
fetch = refs/heads/*:refs/heads/*
Such a builtin remote simplifies the operation of "git-fetch", which will populate FETCH_HEAD but will not pretend that two repositories are in use, will not create a thin pack, and will not perform any useless remapping of names. The speed improvement is around 20%, and it should improve more if "git-fetch" is converted to a builtin.
To this end, git-parse-remote is grown with a new kind of remote, `builtin'. In git-fetch.sh, we treat the builtin remote specially in that it needs no pack/store operations. In fact, doing git-fetch on a builtin remote will simply populate FETCH_HEAD appropriately.
The patch also improves of the --track/--no-track support, extending it so that branch.<name>.remote items referring `.' can be created. Finally, it fixes a typo in git-checkout.sh.
--- Documentation/config.txt | 4 ++++ builtin-branch.c | 39 +++++++++++++++++++++++++-------------- git-checkout.sh | 2 +- git-fetch.sh | 29 +++++++++++++++++++---------- git-parse-remote.sh | 13 +++++++++++-- t/t3200-branch.sh | 6 ++++++ t/t5520-pull.sh | 24 ++++++++++++++++++++++++ 7 files changed, 90 insertions(+), 27 deletions(-)
This patch includes the testcase from Junio -- which now passes. Fixing it is easy but required a curious implementation:
since all we have to do is append fetched heads to FETCH_HEADS, it is simpler to put the handling of builtin `.' into fetch_dumb.
To make the name of fetch_dumb more consistent with the functionality, I renamed it to fetch_heads, as well as renaming fetch_native to fetch_packs.
The patch was regression tested and shows no regression. I added to t5520-pull.sh the testcase that was failing in the first respin.
Documentation/config.txt +4 -0
@@ -272,6 +272,10 @@ branch.<name>.merge:: `git fetch`) to lookup the default branch for merging. Without this option, `git pull` defaults to merge the first refspec fetched. Specify multiple values to get an octopus merge. + If you wish to setup `git pull` so that it merges into <name> from + another branch in the local repository, you can point + branch.<name>.merge to the desired branch, and use the special setting + `.` (a period) for branch.<name>.remote. color.branch:: A boolean to enable/disable color in the output of
builtin-branch.c +25 -14
@@ -372,9 +372,26 @@ static int get_remote_config(const char *key, const char *value)return 0 ; } -static void set_branch_defaults (const char *name,const char *real_ref) +static void set_branch_merge (const char *name,const char *config_remote, +const char *config_repo) {char key[1024 ]; +if (sizeof (key) <= +snprintf (key,sizeof (key),"branch.%s.remote" , name)) +die ("what a long branch name you have!" ); +git_config_set (key, config_remote); + +/* +* We do not have to check if we have enough space for +* the 'merge' key, since it's shorter than the +* previous 'remote' key, which we already checked. +*/ +snprintf (key,sizeof (key),"branch.%s.merge" , name); +git_config_set (key, config_repo); +} + +static void set_branch_defaults (const char *name,const char *real_ref) +{const char *slash =strrchr (real_ref,'/' );if (!slash)
@@ -384,21 +401,15 @@ static void set_branch_defaults(const char *name, const char *real_ref) start_len =strlen (real_ref); base_len = slash - real_ref;git_config (get_remote_config); +if (!config_repo && !config_remote && + !prefixcmp (real_ref,"refs/heads/" )) { +set_branch_merge (name,"." , real_ref); +printf ("Branch %s set up to track local branch %s.\n" , + name, real_ref); + }if (config_repo && config_remote) { -if (sizeof (key) <= -snprintf (key,sizeof (key),"branch.%s.remote" , name)) -die ("what a long branch name you have!" ); -git_config_set (key, config_remote); - -/* -* We do not have to check if we have enough space for -* the 'merge' key, since it's shorter than the -* previous 'remote' key, which we already checked. -*/ -snprintf (key,sizeof (key),"branch.%s.merge" , name); -git_config_set (key, config_repo); - +set_branch_merge (name, config_remote, config_repo);printf ("Branch %s set up to track remote branch %s.\n" , name, real_ref); }
git-checkout.sh +1 -1
@@ -89,7 +89,7 @@ while [ "$#" != "0" ]; do esac done -case "$new_branch,$track" in +case "$newbranch,$track" in ,--*) die "git checkout: --track and --no-track require -b" esac
git-fetch.sh +19 -10
@@ -161,7 +161,7 @@ then fi fi -fetch_native () { +fetch_packs () { eval=$(echo "$1" | git-fetch--tool parse-reflist "-") eval "$eval"
@@ -192,7 +192,7 @@ fetch_native () { } -fetch_dumb () { +fetch_heads () { reflist="$1" refs= rref=
@@ -286,6 +286,10 @@ fetch_dumb () { rsync_slurped_objects=t } ;; + .) + local_name=$remote_name + head=$(git-rev-parse --verify "$local_name") + ;; esac append_fetch_head "$head" "$remote" \
@@ -296,14 +300,19 @@ fetch_dumb () { } fetch_main () { - case "$remote" in - http://* | https://* | ftp://* | rsync://* ) - fetch_dumb "$@" - ;; - *) - fetch_native "$@" - ;; - esac + data_source=$(get_data_source "$remote_nick") + if test "$data_source" = builtin; then + fetch_heads "$@" + else + case "$remote" in + http://* | https://* | ftp://* | rsync://* ) + fetch_heads "$@" + ;; + *) + fetch_packs "$@" + ;; + esac + fi } fetch_main "$reflist" || exit
git-parse-remote.sh +11 -2
@@ -9,6 +9,9 @@ get_data_source () { */*) echo '' ;; + .) + echo builtin + ;; *) if test "$(git-config --get "remote.$1.url")" then
@@ -31,6 +34,9 @@ get_remote_url () { '') echo "$1" ;; + builtin) + echo "$1" + ;; config) git-config --get "remote.$1.url" ;;
@@ -57,7 +63,7 @@ get_default_remote () { get_remote_default_refs_for_push () { data_source=$(get_data_source "$1") case "$data_source" in - '' | branches) + '' | branches | builtin) ;; # no default push mapping, just send matching refs. config) git-config --get-all "remote.$1.push" ;;
@@ -163,6 +169,9 @@ get_remote_default_refs_for_fetch () { case "$data_source" in '') echo "HEAD:" ;; + builtin) + canon_refs_list_for_fetch -d "$1" \ + $(git-show-ref | sed -n 's,.*[ ]\(refs/.*\),\1:,p') ;; config) canon_refs_list_for_fetch -d "$1" \ $(git-config --get-all "remote.$1.fetch") ;;
@@ -177,7 +186,7 @@ get_remote_default_refs_for_fetch () { }' "$GIT_DIR/remotes/$1") ;; *) - die "internal error: get-remote-default-ref-for-push $1" ;; + die "internal error: get-remote-default-ref-for-fetch $1" ;; esac }
t/t3200-branch.sh +6 -0
@@ -145,9 +145,15 @@ test_expect_success 'test overriding tracking setup via --no-track' \ git-config remote.local.fetch refs/heads/*:refs/remotes/local/* && (git-show-ref -q refs/remotes/local/master || git-fetch local) && git-branch --no-track my2 local/master && + git-config branch.autosetupmerge false && ! test $(git-config branch.my2.remote) = local && ! test $(git-config branch.my2.merge) = refs/heads/master' +test_expect_success 'test local tracking setup' \ + 'git branch --track my6 s && + test $(git-config branch.my6.remote) = . && + test $(git-config branch.my6.merge) = refs/heads/s' + # Keep this test last, as it changes the current branch cat >expect <<EOF 0000000000000000000000000000000000000000 $HEAD $GIT_COMMITTER_NAME <$GIT_COMMITTER_EMAIL> 1117150200 +0000 branch: Created from master
t/t5520-pull.sh +24 -0
@@ -29,5 +29,29 @@ test_expect_success 'checking the results' ' diff file cloned/file ' +test_expect_success 'test . as a remote' ' + + git branch copy master && + git config branch.copy.remote . && + git config branch.copy.merge refs/heads/master && + echo updated >file && + git commit -a -m updated && + git checkout copy && + test `cat file` = file && + git pull && + test `cat file` = updated +' + +test_expect_success 'the default remote . should not break explicit pull' ' + git checkout -b second master^ && + echo modified >file && + git commit -a -m modified && + git checkout copy && + git reset --hard HEAD^ && + test `cat file` = file && + git pull . second && + test `cat file` = modified +' + test_done
Quoted 3 lines · Paolo Bonzini
To make the name of
fetch_dumbmore consistent with the functionality, I renamed it tofetch_heads, as well as renamingfetch_nativetofetch_packs.
I agree renaming fetch_dumb vs fetch_native to names that reflect their nature better might be a good idea. The former is to perform fetching objects and updating tracking ref for one ref at a time, while the latter is to do the fetching and updating all relevant tracking refs in one go. The 'per-ref' behaviour is exactly why the former is called "dumb".
If you really want to rename, perhaps making the former fetch_per_ref and the latter fetch_all_at_once would reflect their nature better. While I do not have strong preference either way, I think "fetch_head" is a slight improvement over "fetch_dumb", because it tells that "per-ref" nature a bit better, and "fetch_pack" is not any better than "fetch_native".
While talking about names, I think calling data_source 'builtin' feels very wrong, as we _might_ want to later add other kinds of repository the system knows about. I would have picked 'self', as this change is about referring the repository itself using a fake name '.' for it.
I think it would be a lot more efficient if you made the 'self' fetch piggyback on fetch_all_at_once side, like the attached patch.
This is on top of your patch to illustrate the difference, and does not have the renaming I suggested (i.e. dumb->per_ref, packs->all_at_once, and builtin->self).
git-fetch.sh +5 -5
@@ -170,6 +170,10 @@ fetch_packs () { die "shallow clone with bundle is not supported" git-bundle unbundle "$remote" $rref || echo failed "$remote" + elif test "." = "$remote" + then + git-show-ref $rref || + echo failed . else git-fetch-pack --thin $exec $keep $shallow_depth $no_progress \ "$remote" $rref ||
@@ -280,10 +284,6 @@ fetch_heads () { rsync_slurped_objects=t } ;; - .) - local_name=$remote_name - head=$(git-rev-parse --verify "$local_name") - ;; esac append_fetch_head "$head" "$remote" \
@@ -296,7 +296,7 @@ fetch_heads () { fetch_main () { data_source=$(get_data_source "$remote_nick") if test "$data_source" = builtin; then - fetch_heads "$@" + fetch_packs "$@" else case "$remote" in http://* | https://* | ftp://* | rsync://* )
PATCH 1/1 git-fetch, git-parse-remote: Cleanup implementation of '.'
As per the mailing list exchanges, this applies the following changes:
- renames the data source from 'builtin' to 'self'.
- renames fetch_packs/fetch_heads to fetch_per_ref and fetch_all_at_once
- processes any remote whose URL is '.' (not only the builtin one) using native-store (this is tested by t3200-branch.sh).
--- git-fetch.sh | 33 +++++++++++++-------------------- git-parse-remote.sh | 8 ++++---- 2 files changed, 17 insertions(+), 24 deletions(-)
The patch is a cleanup of the one you posted, and it was appropriately tested with no regression.
I can see now what you meant by the split between fetch and merge logic making my patch more complicated than necessary.
git-fetch.sh +13 -20
@@ -161,7 +161,7 @@ then fi fi -fetch_packs () { +fetch_all_at_once () { eval=$(echo "$1" | git-fetch--tool parse-reflist "-") eval "$eval"
@@ -169,7 +169,9 @@ fetch_packs () { ( : subshell because we muck with IFS IFS=" $LF" ( - if test -f "$remote" ; then + if test "$remote" = . ; then + git-show-ref $rref || echo failed "$remote" + elif test -f "$remote" ; then test -n "$shallow_depth" && die "shallow clone with bundle is not supported" git-bundle unbundle "$remote" $rref ||
@@ -192,7 +194,7 @@ fetch_packs () { } -fetch_heads () { +fetch_per_ref () { reflist="$1" refs= rref=
@@ -286,10 +288,6 @@ fetch_heads () { rsync_slurped_objects=t } ;; - .) - local_name=$remote_name - head=$(git-rev-parse --verify "$local_name") - ;; esac append_fetch_head "$head" "$remote" \
@@ -300,19 +298,14 @@ fetch_heads () { } fetch_main () { - data_source=$(get_data_source "$remote_nick") - if test "$data_source" = builtin; then - fetch_heads "$@" - else - case "$remote" in - http://* | https://* | ftp://* | rsync://* ) - fetch_heads "$@" - ;; - *) - fetch_packs "$@" - ;; - esac - fi + case "$remote" in + http://* | https://* | ftp://* | rsync://* ) + fetch_per_ref "$@" + ;; + *) + fetch_all_at_once "$@" + ;; + esac } fetch_main "$reflist" || exit
git-parse-remote.sh +4 -4
@@ -10,7 +10,7 @@ get_data_source () { echo '' ;; .) - echo builtin + echo self ;; *) if test "$(git-config --get "remote.$1.url")"
@@ -34,7 +34,7 @@ get_remote_url () { '') echo "$1" ;; - builtin) + self) echo "$1" ;; config)
@@ -63,7 +63,7 @@ get_default_remote () { get_remote_default_refs_for_push () { data_source=$(get_data_source "$1") case "$data_source" in - '' | branches | builtin) + '' | branches | self) ;; # no default push mapping, just send matching refs. config) git-config --get-all "remote.$1.push" ;;
@@ -169,7 +169,7 @@ get_remote_default_refs_for_fetch () { case "$data_source" in '') echo "HEAD:" ;; - builtin) + self) canon_refs_list_for_fetch -d "$1" \ $(git-show-ref | sed -n 's,.*[ ]\(refs/.*\),\1:,p') ;; config)
Re: [PATCH] git-fetch, git-parse-remote: Cleanup implementation of '.'
Quoted 5 lines · Paolo Bonzini
The patch is a cleanup of the one you posted, and it was appropriately tested with no regression.
I can see now what you meant by the split between fetch and merge logic making my patch more complicated than necessary.
Thanks. I'll squash the two and apply.
Review
Quoted 3 lines
(quoted diff) +3
-0+ builtin) + canon_refs_list_for_fetch -d "$1" \ + $(git-show-ref | sed -n 's,.*[ ]\(refs/.*\),\1:,p') ;;
I may not be really thinking straight tonight (no I am not drunk, but just a tad sick), but I wonder if this is sufficient?
$(git-for-each-ref --format='%(refname):')
Shorter and one less process and pipe.
Re: [PATCH] git-fetch, git-parse-remote: Cleanup implementation of '.'
Quoted 7 lines
I may not be really thinking straight tonight (no I am not drunk, but just a tad sick), but I wonder if this is sufficient?
$(git-for-each-ref --format='%(refname):')Shorter and one less process and pipe.
It is. Would you take care of that?
Paolo
Re: [PATCH] git-fetch, git-parse-remote: Cleanup implementation of '.'
Quoted 8 lines · Paolo Bonzini
Quoted 6 lines
I may not be really thinking straight tonight (no I am not drunk, but just a tad sick), but I wonder if this is sufficient?
$(git-for-each-ref --format='%(refname):')Shorter and one less process and pipe.
It is. Would you take care of that?
Surely, and thanks for sanity checking.