git/list[1] front-page[2] threads[3] people[4] search[5] about
 

[PATCH v6 03/32] files-backend: break out ref reading

From
David Turner <dturner@twopensource.com>
Date
Feb 24, 2016, 22:58 UTC
Message-ID
<1456354744-8022-4-git-send-email-dturner@twopensource.com>
In-Reply-To
<1456354744-8022-1-git-send-email-dturner@twopensource.com>

Refactor resolve_ref_1 in terms of a new function read_raw_ref, which is responsible for reading ref data from the ref storage.

Later, we will make read_raw_ref a pluggable backend function, and make resolve_ref_unsafe common.

Testing done: Hacked in code to run both old and new version of resolve_ref_1 and compare all outputs, failing dramatically if outputs differed. Ran test suite.

Signed-off-by: David Turner <dturner@twopensource.com>
Helped-by: Duy Nguyen <pclouds@gmail.com>
---
 refs/files-backend.c | 265 ++++++++++++++++++++++++++++++---------------------
 1 file changed, 159 insertions(+), 106 deletions(-)
diff --git a/refs/files-backend.c b/refs/files-backend.c
index fd664d6..ef5f28d 100644
--- a/refs/files-backend.c
+++ b/refs/files-backend.c
@@ -1377,10 +1377,9 @@ static struct ref_entry *get_packed_ref(const char *refname)
 
 /*
  * A loose ref file doesn't exist; check for a packed ref.  The
- * options are forwarded from resolve_safe_unsafe().
+ * options are forwarded from resolve_ref_unsafe().
  */
 static int resolve_missing_loose_ref(const char *refname,
-				     int resolve_flags,
 				     unsigned char *sha1,
 				     int *flags)
 {
@@ -1398,64 +1397,48 @@ static int resolve_missing_loose_ref(const char *refname,
 		return 0;
 	}
 	/* The reference is not a packed reference, either. */
-	if (resolve_flags & RESOLVE_REF_READING) {
-		errno = ENOENT;
-		return -1;
-	} else {
-		hashclr(sha1);
-		return 0;
-	}
+	errno = ENOENT;
+	return -1;
 }
 
-/* This function needs to return a meaningful errno on failure */
-static const char *resolve_ref_1(const char *refname,
-				 int resolve_flags,
-				 unsigned char *sha1,
-				 int *flags,
-				 struct strbuf *sb_refname,
-				 struct strbuf *sb_path,
-				 struct strbuf *sb_contents)
+/*
+ * Read a raw ref from the filesystem or packed refs file.
+ *
+ * If the ref is a sha1, fill in sha1 and return 0.
+ *
+ * If the ref is symbolic, fill in *symref with the referrent
+ * (e.g. "refs/heads/master") and return 0.  The caller is responsible
+ * for validating the referrent.  Set REF_ISSYMREF in flags.
+ *
+ * If the ref is neither a symbolic ref nor a sha1, it is broken.  Set
+ * REF_ISBROKEN in flags, set errno to EINVAL, and return -1.
+ *
+ * If the ref doesn't exist, set errno to ENOENT and return -1.
+ *
+ * If there is another error reading the ref, set errno appropriately and
+ * return -1.
+ *
+ * Backend-specific flags might be set in flags as well, regardless of
+ * outcome.
+ *
+ * sb_path is workspace: the caller should allocate and free it.
+ */
+static int read_raw_ref(const char *refname, unsigned char *sha1,
+			struct strbuf *symref, struct strbuf *sb_path,
+			unsigned int *flags)
 {
-	int depth = MAXDEPTH;
-	int bad_name = 0;
-
-	if (flags)
-		*flags = 0;
+	struct strbuf sb_contents = STRBUF_INIT;
+	int ret = -1;
+	const char *path;
+	const char *buf;
 
-	if (check_refname_format(refname, REFNAME_ALLOW_ONELEVEL)) {
-		if (flags)
-			*flags |= REF_BAD_NAME;
+	strbuf_reset(sb_path);
+	strbuf_git_path(sb_path, "%s", refname);
+	path = sb_path->buf;
 
-		if (!(resolve_flags & RESOLVE_REF_ALLOW_BAD_NAME) ||
-		    !refname_is_safe(refname)) {
-			errno = EINVAL;
-			return NULL;
-		}
-		/*
-		 * dwim_ref() uses REF_ISBROKEN to distinguish between
-		 * missing refs and refs that were present but invalid,
-		 * to complain about the latter to stderr.
-		 *
-		 * We don't know whether the ref exists, so don't set
-		 * REF_ISBROKEN yet.
-		 */
-		bad_name = 1;
-	}
-	for (;;) {
-		const char *path;
+	for(;;) {
 		struct stat st;
-		char *buf;
 		int fd;
-
-		if (--depth < 0) {
-			errno = ELOOP;
-			return NULL;
-		}
-
-		strbuf_reset(sb_path);
-		strbuf_git_path(sb_path, "%s", refname);
-		path = sb_path->buf;
-
 		/*
 		 * We might have to loop back here to avoid a race
 		 * condition: first we lstat() the file, then we try
@@ -1465,49 +1448,45 @@ static const char *resolve_ref_1(const char *refname,
 		 * we don't want to report that as an error but rather
 		 * try again starting with the lstat().
 		 */
-	stat_ref:
+
 		if (lstat(path, &st) < 0) {
 			if (errno != ENOENT)
-				return NULL;
-			if (resolve_missing_loose_ref(refname, resolve_flags,
-						      sha1, flags))
-				return NULL;
-			if (bad_name) {
-				hashclr(sha1);
-				if (flags)
-					*flags |= REF_ISBROKEN;
-			}
-			return refname;
+				break;
+			if (resolve_missing_loose_ref(refname, sha1, flags)) 
+				break;
+			ret = 0;
+			break;
 		}
 
 		/* Follow "normalized" - ie "refs/.." symlinks by hand */
 		if (S_ISLNK(st.st_mode)) {
-			strbuf_reset(sb_contents);
-			if (strbuf_readlink(sb_contents, path, 0) < 0) {
+			strbuf_reset(&sb_contents);
+			if (strbuf_readlink(&sb_contents, path, 0) < 0) {
 				if (errno == ENOENT || errno == EINVAL)
 					/* inconsistent with lstat; retry */
-					goto stat_ref;
+					continue;
 				else
-					return NULL;
+					break;
 			}
-			if (starts_with(sb_contents->buf, "refs/") &&
-			    !check_refname_format(sb_contents->buf, 0)) {
-				strbuf_swap(sb_refname, sb_contents);
-				refname = sb_refname->buf;
+			if (starts_with(sb_contents.buf, "refs/") &&
+			    !check_refname_format(sb_contents.buf, 0)) {
+				strbuf_swap(&sb_contents, symref);
 				if (flags)
 					*flags |= REF_ISSYMREF;
-				if (resolve_flags & RESOLVE_REF_NO_RECURSE) {
-					hashclr(sha1);
-					return refname;
-				}
-				continue;
+				ret = 0;
+				break;
+			} else {
+				/* bogus symlink ref  */
+				if (flags)
+					*flags |= REF_ISBROKEN;
+				break;
 			}
 		}
 
 		/* Is it a directory? */
 		if (S_ISDIR(st.st_mode)) {
 			errno = EISDIR;
-			return NULL;
+			break;
 		}
 
 		/*
@@ -1518,35 +1497,110 @@ static const char *resolve_ref_1(const char *refname,
 		if (fd < 0) {
 			if (errno == ENOENT)
 				/* inconsistent with lstat; retry */
-				goto stat_ref;
+				continue;
 			else
-				return NULL;
+				break;
 		}
-		strbuf_reset(sb_contents);
-		if (strbuf_read(sb_contents, fd, 256) < 0) {
+		strbuf_reset(&sb_contents);
+		if (strbuf_read(&sb_contents, fd, 256) < 0) {
 			int save_errno = errno;
 			close(fd);
 			errno = save_errno;
-			return NULL;
+			break;
 		}
 		close(fd);
-		strbuf_rtrim(sb_contents);
+		strbuf_rtrim(&sb_contents);
+		buf = sb_contents.buf;
+		if (starts_with(buf, "ref:")) {
+			buf += 4;
+			while (isspace(*buf))
+				buf++;
+
+			strbuf_reset(symref);
+			strbuf_addstr(symref, buf);
+			if (flags)
+				*flags |= REF_ISSYMREF;
+			ret = 0;
+			break;
+		}
 
 		/*
-		 * Is it a symbolic ref?
+		 * Please note that FETCH_HEAD has additional
+		 * data after the sha.
 		 */
-		if (!starts_with(sb_contents->buf, "ref:")) {
-			/*
-			 * Please note that FETCH_HEAD has a second
-			 * line containing other data.
-			 */
-			if (get_sha1_hex(sb_contents->buf, sha1) ||
-			    (sb_contents->buf[40] != '\0' && !isspace(sb_contents->buf[40]))) {
+		if (get_sha1_hex(buf, sha1) ||
+		    (buf[40] != '\0' && !isspace(buf[40]))) {
+			if (flags)
+				*flags |= REF_ISBROKEN;
+			errno = EINVAL;
+			break;
+		}
+		ret = 0;
+		break;
+	}
+
+	strbuf_release(&sb_contents);
+	return ret;
+}
+
+/* This function needs to return a meaningful errno on failure */
+static const char *resolve_ref_1(const char *refname,
+				 int resolve_flags,
+				 unsigned char *sha1,
+				 int *flags,
+				 struct strbuf *sb_refname,
+				 struct strbuf *sb_path)
+{
+	int bad_name = 0;
+	int symref_count;
+
+	if (flags)
+		*flags = 0;
+
+	if (check_refname_format(refname, REFNAME_ALLOW_ONELEVEL)) {
+		if (flags)
+			*flags |= REF_BAD_NAME;
+
+		if (!(resolve_flags & RESOLVE_REF_ALLOW_BAD_NAME) ||
+		    !refname_is_safe(refname)) {
+			errno = EINVAL;
+			return NULL;
+		}
+		/*
+		 * dwim_ref() uses REF_ISBROKEN to distinguish between
+		 * missing refs and refs that were present but invalid,
+		 * to complain about the latter to stderr.
+		 *
+		 * We don't know whether the ref exists, so don't set
+		 * REF_ISBROKEN yet.
+		 */
+		bad_name = 1;
+	}
+
+	for (symref_count = 0; symref_count < MAXDEPTH; symref_count++) {
+		int read_flags = 0;
+
+		if (read_raw_ref(refname, sha1, sb_refname, sb_path, &read_flags)) {
+			int saved_errno = errno;
+			if (flags)
+				*flags |= read_flags;
+			errno = saved_errno;
+			if (bad_name) {
+				hashclr(sha1);
 				if (flags)
 					*flags |= REF_ISBROKEN;
-				errno = EINVAL;
+			}
+			if (resolve_flags & RESOLVE_REF_READING || errno != ENOENT) {
 				return NULL;
+			} else {
+				hashclr(sha1);
+				return refname;
 			}
+		}
+		if (flags)
+			*flags |= read_flags;
+
+		if (!(read_flags & REF_ISSYMREF)) {
 			if (bad_name) {
 				hashclr(sha1);
 				if (flags)
@@ -1554,44 +1608,43 @@ static const char *resolve_ref_1(const char *refname,
 			}
 			return refname;
 		}
-		if (flags)
-			*flags |= REF_ISSYMREF;
-		buf = sb_contents->buf + 4;
-		while (isspace(*buf))
-			buf++;
-		strbuf_reset(sb_refname);
-		strbuf_addstr(sb_refname, buf);
+
 		refname = sb_refname->buf;
 		if (resolve_flags & RESOLVE_REF_NO_RECURSE) {
 			hashclr(sha1);
+			if (bad_name && flags)
+				*flags |= REF_ISBROKEN;
 			return refname;
 		}
-		if (check_refname_format(buf, REFNAME_ALLOW_ONELEVEL)) {
+
+		if (check_refname_format(refname, REFNAME_ALLOW_ONELEVEL)) {
 			if (flags)
 				*flags |= REF_ISBROKEN;
-
 			if (!(resolve_flags & RESOLVE_REF_ALLOW_BAD_NAME) ||
-			    !refname_is_safe(buf)) {
+			    !refname_is_safe(refname)) {
 				errno = EINVAL;
 				return NULL;
 			}
 			bad_name = 1;
 		}
 	}
+
+	if (flags)
+		*flags |= REF_ISBROKEN;
+	return NULL;
 }
 
 const char *resolve_ref_unsafe(const char *refname, int resolve_flags,
 			       unsigned char *sha1, int *flags)
 {
 	static struct strbuf sb_refname = STRBUF_INIT;
-	struct strbuf sb_contents = STRBUF_INIT;
 	struct strbuf sb_path = STRBUF_INIT;
 	const char *ret;
 
 	ret = resolve_ref_1(refname, resolve_flags, sha1, flags,
-			    &sb_refname, &sb_path, &sb_contents);
+			    &sb_refname, &sb_path);
+
 	strbuf_release(&sb_path);
-	strbuf_release(&sb_contents);
 	return ret;
 }
 
-- 
2.4.2.767.g62658d5-twtrsrc
Previous: David TurnerNext: Junio C Hamano
Message 4 of 56 in “refs backend”
  1. 00/32 refs backendDavid Turner, Feb 24, 2016
  2. 01/32 refs: move head_ref{,_submodule} to the common codeDavid Turner, Feb 24, 2016
  3. 02/32 refs: move for_each_*ref* functions into common codeDavid Turner, Feb 24, 2016
  4. 03/32 files-backend: break out ref readingDavid Turner, Feb 24, 2016
  5. Junio C HamanoFeb 25, 2016
  6. Junio C HamanoFeb 25, 2016
  7. David TurnerFeb 25, 2016
  8. Junio C HamanoFeb 25, 2016
  9. 04/32 refs: move resolve_ref_unsafe into common codeDavid Turner, Feb 24, 2016
  10. Junio C HamanoFeb 25, 2016
  11. 05/32 refs: add a backend method structure with transaction functionsDavid Turner, Feb 24, 2016
  12. Jeff KingFeb 27, 2016
  13. David TurnerFeb 29, 2016
  14. 06/32 refs: add methods for misc ref operationsDavid Turner, Feb 24, 2016
  15. 07/32 refs: add method for do_for_each_refDavid Turner, Feb 24, 2016
  16. 08/32 refs: reduce the visibility of do_for_each_ref()David Turner, Feb 24, 2016
  17. 09/32 refs: add do_for_each_per_worktree_refDavid Turner, Feb 24, 2016
  18. 10/32 refs: add methods for reflogDavid Turner, Feb 24, 2016
  19. 11/32 refs: add method for initial ref transaction commitDavid Turner, Feb 24, 2016
  20. 12/32 refs: add method for delete_refsDavid Turner, Feb 24, 2016
  21. 13/32 refs: add methods to init refs dbDavid Turner, Feb 24, 2016
  22. 14/32 refs: add method to rename refsDavid Turner, Feb 24, 2016
  23. 15/32 refs: handle non-normal ref renamesDavid Turner, Feb 24, 2016
  24. 16/32 refs: make lock genericDavid Turner, Feb 24, 2016
  25. 17/32 refs: move duplicate check to common codeDavid Turner, Feb 24, 2016
  26. 18/32 refs: allow log-only updatesDavid Turner, Feb 24, 2016
  27. 19/32 refs: don't dereference on renameDavid Turner, Feb 24, 2016
  28. 20/32 refs: on symref reflog expire, lock symref not referrentDavid Turner, Feb 24, 2016
  29. 21/32 refs: resolve symbolic refs firstDavid Turner, Feb 24, 2016
  30. 22/32 refs: always handle non-normal refs in files backendDavid Turner, Feb 24, 2016
  31. 23/32 init: allow alternate ref strorage to be set for new reposDavid Turner, Feb 24, 2016
  32. 24/32 refs: check submodules' ref storage configDavid Turner, Feb 24, 2016
  33. 25/32 clone: allow ref storage backend to be set for cloneDavid Turner, Feb 24, 2016
  34. 26/32 svn: learn ref-storage argumentDavid Turner, Feb 24, 2016
  35. 27/32 refs: register ref storage backendsDavid Turner, Feb 24, 2016
  36. 28/32 config: read ref storage config on startupDavid Turner, Feb 24, 2016
  37. Jeff KingFeb 27, 2016
  38. Jeff KingFeb 27, 2016
  39. David TurnerFeb 29, 2016
  40. Jeff KingFeb 29, 2016
  41. David TurnerFeb 29, 2016
  42. Jeff KingFeb 29, 2016
  43. Jeff KingFeb 29, 2016
  44. David TurnerFeb 29, 2016
  45. 29/32 refs: break out resolve_ref_unsafe_submoduleDavid Turner, Feb 24, 2016
  46. 30/32 refs: add LMDB refs storage backendDavid Turner, Feb 24, 2016
  47. 31/32 refs: tests for lmdb backendDavid Turner, Feb 24, 2016
  48. 32/32 tests: add ref-storage argumentDavid Turner, Feb 24, 2016
  49. SZEDER GáborFeb 29, 2016
  50. David TurnerFeb 29, 2016
  51. Ramsay JonesFeb 25, 2016
  52. David TurnerFeb 25, 2016
  53. Duy NguyenFeb 25, 2016
  54. David TurnerFeb 25, 2016
  55. Stefan BellerFeb 25, 2016
  56. Duy NguyenFeb 25, 2016

Read the whole thread, see it on lore, or plain text.

$ cat FOOTERMessages come from the public archive at lore.kernel.org/git, fetched every hour. The front page is chosen and written each morning by an AI editor and can be wrong; the threads themselves are the record. About and API. For agents: an MCP server at https://gitlist.dev/mcp, and any thread, story or person page as Markdown by adding .md to its URL (or sending Accept: text/markdown). Details in /llms.txt.