ssh-upload: prevent buffer overrun
[gitweb.git] / unpack-trees.c
index 20251822e4a8dd410b0232b5e0b13d436758df85..2e2232cbb07e61de3be74302fba67142a58a857b 100644 (file)
@@ -1,10 +1,12 @@
-#include <signal.h>
-#include <sys/time.h>
 #include "cache.h"
+#include "dir.h"
 #include "tree.h"
 #include "tree-walk.h"
+#include "cache-tree.h"
 #include "unpack-trees.h"
 
+#define DBRT_DEBUG 1
+
 struct tree_entry_list {
        struct tree_entry_list *next;
        unsigned directory : 1;
@@ -22,6 +24,9 @@ static struct tree_entry_list *create_tree_entry_list(struct tree *tree)
        struct tree_entry_list *ret = NULL;
        struct tree_entry_list **list_p = &ret;
 
+       if (!tree->object.parsed)
+               parse_tree(tree);
+
        desc.buf = tree->buffer;
        desc.size = tree->size;
 
@@ -71,6 +76,12 @@ static int unpack_trees_rec(struct tree_entry_list **posns, int len,
 {
        int baselen = strlen(base);
        int src_size = len + 1;
+       int i_stk = i_stk;
+       int retval = 0;
+
+       if (o->dir)
+               i_stk = push_exclude_per_directory(o->dir, base, strlen(base));
+
        do {
                int i;
                const char *first;
@@ -137,7 +148,7 @@ static int unpack_trees_rec(struct tree_entry_list **posns, int len,
                }
                /* No name means we're done */
                if (!first)
-                       return 0;
+                       goto leave_directory;
 
                pathlen = strlen(first);
                ce_size = cache_entry_size(baselen + pathlen);
@@ -194,7 +205,7 @@ static int unpack_trees_rec(struct tree_entry_list **posns, int len,
 
                        any_files = 1;
 
-                       memcpy(ce->sha1, posns[i]->sha1, 20);
+                       hashcpy(ce->sha1, posns[i]->sha1);
                        src[i + o->merge] = ce;
                        subposns[i] = df_conflict_list;
                        posns[i] = posns[i]->next;
@@ -234,13 +245,20 @@ static int unpack_trees_rec(struct tree_entry_list **posns, int len,
                        newbase[baselen + pathlen] = '/';
                        newbase[baselen + pathlen + 1] = '\0';
                        if (unpack_trees_rec(subposns, len, newbase, o,
-                                            indpos, df_conflict_list))
-                               return -1;
+                                            indpos, df_conflict_list)) {
+                               retval = -1;
+                               goto leave_directory;
+                       }
                        free(newbase);
                }
                free(subposns);
                free(src);
        } while (1);
+
+ leave_directory:
+       if (o->dir)
+               pop_exclude_per_directory(o->dir, i_stk);
+       return retval;
 }
 
 /* Unlink the last component and attempt to remove leading
@@ -272,7 +290,7 @@ static void unlink_entry(char *name)
        }
 }
 
-static volatile int progress_update = 0;
+static volatile sig_atomic_t progress_update;
 
 static void progress_interval(int signum)
 {
@@ -364,17 +382,21 @@ int unpack_trees(struct object_list *trees, struct unpack_trees_options *o)
        int i;
        struct object_list *posn = trees;
        struct tree_entry_list df_conflict_list;
-       struct cache_entry df_conflict_entry;
+       static struct cache_entry *dfc;
 
        memset(&df_conflict_list, 0, sizeof(df_conflict_list));
        df_conflict_list.next = &df_conflict_list;
+       memset(&state, 0, sizeof(state));
        state.base_dir = "";
        state.force = 1;
        state.quiet = 1;
        state.refresh_cache = 1;
 
        o->merge_size = len;
-       o->df_conflict_entry = &df_conflict_entry;
+
+       if (!dfc)
+               dfc = xcalloc(1, sizeof(struct cache_entry) + 1);
+       o->df_conflict_entry = dfc;
 
        if (len) {
                posns = xmalloc(len * sizeof(struct tree_entry_list *));
@@ -393,3 +415,397 @@ int unpack_trees(struct object_list *trees, struct unpack_trees_options *o)
        check_updates(active_cache, active_nr, o);
        return 0;
 }
+
+/* Here come the merge functions */
+
+static void reject_merge(struct cache_entry *ce)
+{
+       die("Entry '%s' would be overwritten by merge. Cannot merge.",
+           ce->name);
+}
+
+static int same(struct cache_entry *a, struct cache_entry *b)
+{
+       if (!!a != !!b)
+               return 0;
+       if (!a && !b)
+               return 1;
+       return a->ce_mode == b->ce_mode &&
+              !hashcmp(a->sha1, b->sha1);
+}
+
+
+/*
+ * When a CE gets turned into an unmerged entry, we
+ * want it to be up-to-date
+ */
+static void verify_uptodate(struct cache_entry *ce,
+               struct unpack_trees_options *o)
+{
+       struct stat st;
+
+       if (o->index_only || o->reset)
+               return;
+
+       if (!lstat(ce->name, &st)) {
+               unsigned changed = ce_match_stat(ce, &st, 1);
+               if (!changed)
+                       return;
+               errno = 0;
+       }
+       if (o->reset) {
+               ce->ce_flags |= htons(CE_UPDATE);
+               return;
+       }
+       if (errno == ENOENT)
+               return;
+       die("Entry '%s' not uptodate. Cannot merge.", ce->name);
+}
+
+static void invalidate_ce_path(struct cache_entry *ce)
+{
+       if (ce)
+               cache_tree_invalidate_path(active_cache_tree, ce->name);
+}
+
+/*
+ * We do not want to remove or overwrite a working tree file that
+ * is not tracked, unless it is ignored.
+ */
+static void verify_absent(const char *path, const char *action,
+               struct unpack_trees_options *o)
+{
+       struct stat st;
+
+       if (o->index_only || o->reset || !o->update)
+               return;
+       if (!lstat(path, &st) && !(o->dir && excluded(o->dir, path)))
+               die("Untracked working tree file '%s' "
+                   "would be %s by merge.", path, action);
+}
+
+static int merged_entry(struct cache_entry *merge, struct cache_entry *old,
+               struct unpack_trees_options *o)
+{
+       merge->ce_flags |= htons(CE_UPDATE);
+       if (old) {
+               /*
+                * See if we can re-use the old CE directly?
+                * That way we get the uptodate stat info.
+                *
+                * This also removes the UPDATE flag on
+                * a match.
+                */
+               if (same(old, merge)) {
+                       *merge = *old;
+               } else {
+                       verify_uptodate(old, o);
+                       invalidate_ce_path(old);
+               }
+       }
+       else {
+               verify_absent(merge->name, "overwritten", o);
+               invalidate_ce_path(merge);
+       }
+
+       merge->ce_flags &= ~htons(CE_STAGEMASK);
+       add_cache_entry(merge, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE);
+       return 1;
+}
+
+static int deleted_entry(struct cache_entry *ce, struct cache_entry *old,
+               struct unpack_trees_options *o)
+{
+       if (old)
+               verify_uptodate(old, o);
+       else
+               verify_absent(ce->name, "removed", o);
+       ce->ce_mode = 0;
+       add_cache_entry(ce, ADD_CACHE_OK_TO_ADD|ADD_CACHE_OK_TO_REPLACE);
+       invalidate_ce_path(ce);
+       return 1;
+}
+
+static int keep_entry(struct cache_entry *ce)
+{
+       add_cache_entry(ce, ADD_CACHE_OK_TO_ADD);
+       return 1;
+}
+
+#if DBRT_DEBUG
+static void show_stage_entry(FILE *o,
+                            const char *label, const struct cache_entry *ce)
+{
+       if (!ce)
+               fprintf(o, "%s (missing)\n", label);
+       else
+               fprintf(o, "%s%06o %s %d\t%s\n",
+                       label,
+                       ntohl(ce->ce_mode),
+                       sha1_to_hex(ce->sha1),
+                       ce_stage(ce),
+                       ce->name);
+}
+#endif
+
+int threeway_merge(struct cache_entry **stages,
+               struct unpack_trees_options *o)
+{
+       struct cache_entry *index;
+       struct cache_entry *head;
+       struct cache_entry *remote = stages[o->head_idx + 1];
+       int count;
+       int head_match = 0;
+       int remote_match = 0;
+       const char *path = NULL;
+
+       int df_conflict_head = 0;
+       int df_conflict_remote = 0;
+
+       int any_anc_missing = 0;
+       int no_anc_exists = 1;
+       int i;
+
+       for (i = 1; i < o->head_idx; i++) {
+               if (!stages[i])
+                       any_anc_missing = 1;
+               else {
+                       if (!path)
+                               path = stages[i]->name;
+                       no_anc_exists = 0;
+               }
+       }
+
+       index = stages[0];
+       head = stages[o->head_idx];
+
+       if (head == o->df_conflict_entry) {
+               df_conflict_head = 1;
+               head = NULL;
+       }
+
+       if (remote == o->df_conflict_entry) {
+               df_conflict_remote = 1;
+               remote = NULL;
+       }
+
+       if (!path && index)
+               path = index->name;
+       if (!path && head)
+               path = head->name;
+       if (!path && remote)
+               path = remote->name;
+
+       /* First, if there's a #16 situation, note that to prevent #13
+        * and #14.
+        */
+       if (!same(remote, head)) {
+               for (i = 1; i < o->head_idx; i++) {
+                       if (same(stages[i], head)) {
+                               head_match = i;
+                       }
+                       if (same(stages[i], remote)) {
+                               remote_match = i;
+                       }
+               }
+       }
+
+       /* We start with cases where the index is allowed to match
+        * something other than the head: #14(ALT) and #2ALT, where it
+        * is permitted to match the result instead.
+        */
+       /* #14, #14ALT, #2ALT */
+       if (remote && !df_conflict_head && head_match && !remote_match) {
+               if (index && !same(index, remote) && !same(index, head))
+                       reject_merge(index);
+               return merged_entry(remote, index, o);
+       }
+       /*
+        * If we have an entry in the index cache, then we want to
+        * make sure that it matches head.
+        */
+       if (index && !same(index, head)) {
+               reject_merge(index);
+       }
+
+       if (head) {
+               /* #5ALT, #15 */
+               if (same(head, remote))
+                       return merged_entry(head, index, o);
+               /* #13, #3ALT */
+               if (!df_conflict_remote && remote_match && !head_match)
+                       return merged_entry(head, index, o);
+       }
+
+       /* #1 */
+       if (!head && !remote && any_anc_missing)
+               return 0;
+
+       /* Under the new "aggressive" rule, we resolve mostly trivial
+        * cases that we historically had git-merge-one-file resolve.
+        */
+       if (o->aggressive) {
+               int head_deleted = !head && !df_conflict_head;
+               int remote_deleted = !remote && !df_conflict_remote;
+               /*
+                * Deleted in both.
+                * Deleted in one and unchanged in the other.
+                */
+               if ((head_deleted && remote_deleted) ||
+                   (head_deleted && remote && remote_match) ||
+                   (remote_deleted && head && head_match)) {
+                       if (index)
+                               return deleted_entry(index, index, o);
+                       else if (path && !head_deleted)
+                               verify_absent(path, "removed", o);
+                       return 0;
+               }
+               /*
+                * Added in both, identically.
+                */
+               if (no_anc_exists && head && remote && same(head, remote))
+                       return merged_entry(head, index, o);
+
+       }
+
+       /* Below are "no merge" cases, which require that the index be
+        * up-to-date to avoid the files getting overwritten with
+        * conflict resolution files.
+        */
+       if (index) {
+               verify_uptodate(index, o);
+       }
+
+       o->nontrivial_merge = 1;
+
+       /* #2, #3, #4, #6, #7, #9, #11. */
+       count = 0;
+       if (!head_match || !remote_match) {
+               for (i = 1; i < o->head_idx; i++) {
+                       if (stages[i]) {
+                               keep_entry(stages[i]);
+                               count++;
+                               break;
+                       }
+               }
+       }
+#if DBRT_DEBUG
+       else {
+               fprintf(stderr, "read-tree: warning #16 detected\n");
+               show_stage_entry(stderr, "head   ", stages[head_match]);
+               show_stage_entry(stderr, "remote ", stages[remote_match]);
+       }
+#endif
+       if (head) { count += keep_entry(head); }
+       if (remote) { count += keep_entry(remote); }
+       return count;
+}
+
+/*
+ * Two-way merge.
+ *
+ * The rule is to "carry forward" what is in the index without losing
+ * information across a "fast forward", favoring a successful merge
+ * over a merge failure when it makes sense.  For details of the
+ * "carry forward" rule, please see <Documentation/git-read-tree.txt>.
+ *
+ */
+int twoway_merge(struct cache_entry **src,
+               struct unpack_trees_options *o)
+{
+       struct cache_entry *current = src[0];
+       struct cache_entry *oldtree = src[1], *newtree = src[2];
+
+       if (o->merge_size != 2)
+               return error("Cannot do a twoway merge of %d trees",
+                            o->merge_size);
+
+       if (current) {
+               if ((!oldtree && !newtree) || /* 4 and 5 */
+                   (!oldtree && newtree &&
+                    same(current, newtree)) || /* 6 and 7 */
+                   (oldtree && newtree &&
+                    same(oldtree, newtree)) || /* 14 and 15 */
+                   (oldtree && newtree &&
+                    !same(oldtree, newtree) && /* 18 and 19*/
+                    same(current, newtree))) {
+                       return keep_entry(current);
+               }
+               else if (oldtree && !newtree && same(current, oldtree)) {
+                       /* 10 or 11 */
+                       return deleted_entry(oldtree, current, o);
+               }
+               else if (oldtree && newtree &&
+                        same(current, oldtree) && !same(current, newtree)) {
+                       /* 20 or 21 */
+                       return merged_entry(newtree, current, o);
+               }
+               else {
+                       /* all other failures */
+                       if (oldtree)
+                               reject_merge(oldtree);
+                       if (current)
+                               reject_merge(current);
+                       if (newtree)
+                               reject_merge(newtree);
+                       return -1;
+               }
+       }
+       else if (newtree)
+               return merged_entry(newtree, current, o);
+       else
+               return deleted_entry(oldtree, current, o);
+}
+
+/*
+ * Bind merge.
+ *
+ * Keep the index entries at stage0, collapse stage1 but make sure
+ * stage0 does not have anything there.
+ */
+int bind_merge(struct cache_entry **src,
+               struct unpack_trees_options *o)
+{
+       struct cache_entry *old = src[0];
+       struct cache_entry *a = src[1];
+
+       if (o->merge_size != 1)
+               return error("Cannot do a bind merge of %d trees\n",
+                            o->merge_size);
+       if (a && old)
+               die("Entry '%s' overlaps.  Cannot bind.", a->name);
+       if (!a)
+               return keep_entry(old);
+       else
+               return merged_entry(a, NULL, o);
+}
+
+/*
+ * One-way merge.
+ *
+ * The rule is:
+ * - take the stat information from stage0, take the data from stage1
+ */
+int oneway_merge(struct cache_entry **src,
+               struct unpack_trees_options *o)
+{
+       struct cache_entry *old = src[0];
+       struct cache_entry *a = src[1];
+
+       if (o->merge_size != 1)
+               return error("Cannot do a oneway merge of %d trees",
+                            o->merge_size);
+
+       if (!a)
+               return deleted_entry(old, old, o);
+       if (old && same(old, a)) {
+               if (o->reset) {
+                       struct stat st;
+                       if (lstat(old->name, &st) ||
+                           ce_match_stat(old, &st, 1))
+                               old->ce_flags |= htons(CE_UPDATE);
+               }
+               return keep_entry(old);
+       }
+       return merged_entry(a, old, o);
+}