diff options
| -rw-r--r-- | Documentation/git-rev-list.txt | 13 | ||||
| -rw-r--r-- | builtin-rev-list.c | 212 | ||||
| -rwxr-xr-x | t/t6002-rev-list-bisect.sh | 2 | ||||
| -rwxr-xr-x | t/t6004-rev-list-path-optim.sh | 34 |
4 files changed, 253 insertions, 8 deletions
diff --git a/Documentation/git-rev-list.txt b/Documentation/git-rev-list.txt index 4f145eaba4..3fa45b81cc 100644 --- a/Documentation/git-rev-list.txt +++ b/Documentation/git-rev-list.txt @@ -26,6 +26,7 @@ SYNOPSIS [ [\--objects | \--objects-edge] [ \--unpacked ] ] [ \--pretty | \--header ] [ \--bisect ] + [ \--bisect-vars ] [ \--merge ] [ \--reverse ] [ \--walk-reflogs ] @@ -249,6 +250,18 @@ introduces a regression is thus reduced to a binary search: repeatedly generate and test new 'midpoint's until the commit chain is of length one. +--bisect-vars:: + +This calculates the same as `--bisect`, but outputs text ready +to be eval'ed by the shell. These lines will assign the name of +the midpoint revision to the variable `bisect_rev`, and the +expected number of commits to be tested after `bisect_rev` is +tested to `bisect_nr`, the expected number of commits to be +tested if `bisect_rev` turns out to be good to `bisect_good`, +the expected number of commits to be tested if `bisect_rev` +turns out to be bad to `bisect_bad`, and the number of commits +we are bisecting right now to `bisect_all`. + -- Commit Ordering diff --git a/builtin-rev-list.c b/builtin-rev-list.c index 51858e3233..09e3a60bf6 100644 --- a/builtin-rev-list.c +++ b/builtin-rev-list.c @@ -36,7 +36,8 @@ static const char rev_list_usage[] = " --abbrev=nr | --no-abbrev\n" " --abbrev-commit\n" " special purpose:\n" -" --bisect" +" --bisect\n" +" --bisect-vars" ; static struct rev_info revs; @@ -168,7 +169,8 @@ static void clear_distance(struct commit_list *list) } } -static struct commit_list *find_bisection(struct commit_list *list) +static struct commit_list *find_bisection(struct commit_list *list, + int *reaches, int *all) { int nr, closest; struct commit_list *p, *best; @@ -182,19 +184,21 @@ static struct commit_list *find_bisection(struct commit_list *list) } closest = -1; best = list; + *all = nr; for (p = list; p; p = p->next) { - int distance; + int distance, reach; if (revs.prune_fn && !(p->item->object.flags & TREECHANGE)) continue; - distance = count_distance(p); + distance = reach = count_distance(p); clear_distance(list); if (nr - distance < distance) distance = nr - distance; if (distance > closest) { best = p; + *reaches = reach; closest = distance; } } @@ -203,6 +207,160 @@ static struct commit_list *find_bisection(struct commit_list *list) return best; } +static inline int commit_interesting(struct commit_list *elem) +{ + unsigned flags = elem->item->object.flags; + if (flags & UNINTERESTING) + return 0; + return (!revs.prune_fn || (flags & TREECHANGE)); +} + +static inline int weight(struct commit_list *elem) +{ + return *((int*)(elem->item->util)); +} + +static inline void weight_set(struct commit_list *elem, int weight) +{ + *((int*)(elem->item->util)) = weight; +} + +static int count_interesting_parents(struct commit_list *elem) +{ + int cnt = 0; + if (!elem->item->parents) + return cnt; + for (elem = elem->item->parents; elem; elem = elem->next) { + if (commit_interesting(elem)) + cnt++; + } + return cnt; +} + +static struct commit_list *find_bisection_2(struct commit_list *list, + int *reaches, int *all) +{ + int n, nr, counted, distance; + struct commit_list *p, *best; + int *weights; + + for (nr = 0, p = list; p; p = p->next) { + if (commit_interesting(p)) + nr++; + } + *all = nr; + weights = xcalloc(nr, sizeof(int*)); + counted = 0; + + for (n = 0, p = list; p; p = p->next) { + if (!commit_interesting(p)) + continue; + if (commit_interesting(p)) { + /* + * positive weight is the number of interesting + * commits it can reach, including itself. + * weight = 0 means it has one parent and + * its distance is unknown. + * weight < 0 means it has more than one + * parent and its distance is unknown. + */ + p->item->util = &weights[n++]; + switch (count_interesting_parents(p)) { + case 0: + weight_set(p, 1); + counted++; + break; + case 1: + weight_set(p, 0); + break; + default: + weight_set(p, -1); + break; + } + } + } + + /* + * If you have only one parent in the resulting set + * then you can reach one commit more than that parent + * can reach. So we do not have to run the expensive + * count_distance() for single strand of pearls. + * + * However, if you have more than one parents, you cannot + * just add their distance and one for yourself, since + * they usually reach the same ancestor and you would + * end up counting them twice that way. + * + * So we will first count distance of merges the usual + * way, and then fill the blanks using cheaper algorithm. + */ + for (p = list; p; p = p->next) { + if (!commit_interesting(p)) + continue; + n = weight(p); + if (0 <= n) + continue; + distance = count_distance(p); + clear_distance(p); + weight_set(p, distance); + + /* Does it happen to be at exactly half-way? */ + distance *= 2; + if (nr == distance || (nr+1) == distance) { + p->next = NULL; + *reaches = weight(p); + free(weights); + return p; + } + counted++; + } + + while (counted < nr) { + for (p = list; p; p = p->next) { + struct commit_list *q; + + if (!commit_interesting(p) || 0 < weight(p)) + continue; + for (q = p->item->parents; q; q = q->next) + if (commit_interesting(q) && 0 < weight(q)) + break; + if (!q) + continue; + weight_set(p, weight(q)+1); + counted++; + + /* Does it happen to be at exactly half-way? */ + distance = weight(p) * 2; + if (nr == distance || (nr+1) == distance) { + p->next = NULL; + *reaches = weight(p); + free(weights); + return p; + } + } + } + + /* Then find the best one */ + counted = 0; + best = list; + for (p = list; p; p = p->next) { + if (!commit_interesting(p)) + continue; + distance = weight(p); + if (nr - distance < distance) + distance = nr - distance; + if (distance > counted) { + best = p; + counted = distance; + *reaches = weight(p); + } + } + if (best) + best->next = NULL; + free(weights); + return best; +} + static void read_revisions_from_stdin(struct rev_info *revs) { char line[1000]; @@ -225,6 +383,7 @@ int cmd_rev_list(int argc, const char **argv, const char *prefix) struct commit_list *list; int i; int read_from_stdin = 0; + int bisect_show_vars = 0; git_config(git_default_config); init_revisions(&revs, prefix); @@ -247,6 +406,11 @@ int cmd_rev_list(int argc, const char **argv, const char *prefix) bisect_list = 1; continue; } + if (!strcmp(arg, "--bisect-vars")) { + bisect_list = 1; + bisect_show_vars = 1; + continue; + } if (!strcmp(arg, "--stdin")) { if (read_from_stdin++) die("--stdin given twice?"); @@ -285,8 +449,44 @@ int cmd_rev_list(int argc, const char **argv, const char *prefix) if (revs.tree_objects) mark_edges_uninteresting(revs.commits, &revs, show_edge); - if (bisect_list) - revs.commits = find_bisection(revs.commits); + if (bisect_list) { + int reaches = reaches, all = all; + + if (!revs.prune_fn) + revs.commits = find_bisection_2(revs.commits, + &reaches, &all); + else + revs.commits = find_bisection(revs.commits, + &reaches, &all); + if (bisect_show_vars) { + int cnt; + if (!revs.commits) + return 1; + /* + * revs.commits can reach "reaches" commits among + * "all" commits. If it is good, then there are + * (all-reaches) commits left to be bisected. + * On the other hand, if it is bad, then the set + * to bisect is "reaches". + * A bisect set of size N has (N-1) commits further + * to test, as we already know one bad one. + */ + cnt = all-reaches; + if (cnt < reaches) + cnt = reaches; + printf("bisect_rev=%s\n" + "bisect_nr=%d\n" + "bisect_good=%d\n" + "bisect_bad=%d\n" + "bisect_all=%d\n", + sha1_to_hex(revs.commits->item->object.sha1), + cnt - 1, + all - reaches - 1, + reaches - 1, + all); + return 0; + } + } traverse_commit_list(&revs, show_commit, show_object); diff --git a/t/t6002-rev-list-bisect.sh b/t/t6002-rev-list-bisect.sh index 7831e3461c..fcb3302764 100755 --- a/t/t6002-rev-list-bisect.sh +++ b/t/t6002-rev-list-bisect.sh @@ -163,7 +163,7 @@ test_sequence() # the bisection point is the head - this is the bad point. # -test_output_expect_success "--bisect l5 ^root" 'git-rev-list $_bisect_option l5 ^root' <<EOF +test_output_expect_success "$_bisect_option l5 ^root" 'git-rev-list $_bisect_option l5 ^root' <<EOF c3 EOF diff --git a/t/t6004-rev-list-path-optim.sh b/t/t6004-rev-list-path-optim.sh index 5182dbb158..761f09b1e5 100755 --- a/t/t6004-rev-list-path-optim.sh +++ b/t/t6004-rev-list-path-optim.sh @@ -7,7 +7,8 @@ test_description='git-rev-list trivial path optimization test' test_expect_success setup ' echo Hello > a && git add a && -git commit -m "Initial commit" a +git commit -m "Initial commit" a && +initial=$(git rev-parse --verify HEAD) ' test_expect_success path-optimization ' @@ -16,4 +17,35 @@ test_expect_success path-optimization ' test $(git-rev-list $commit -- . | wc -l) = 1 ' +test_expect_success 'further setup' ' + git checkout -b side && + echo Irrelevant >c && + git add c && + git commit -m "Side makes an irrelevant commit" && + echo "More Irrelevancy" >c && + git add c && + git commit -m "Side makes another irrelevant commit" && + echo Bye >a && + git add a && + git commit -m "Side touches a" && + side=$(git rev-parse --verify HEAD) && + echo "Yet more Irrelevancy" >c && + git add c && + git commit -m "Side makes yet another irrelevant commit" && + git checkout master && + echo Another >b && + git add b && + git commit -m "Master touches b" && + git merge side && + echo Touched >b && + git add b && + git commit -m "Master touches b again" +' + +test_expect_success 'path optimization 2' ' + ( echo "$side"; echo "$initial" ) >expected && + git rev-list HEAD -- a >actual && + diff -u expected actual +' + test_done |