diff options
Diffstat (limited to 'commit.c')
-rw-r--r-- | commit.c | 58 |
1 files changed, 50 insertions, 8 deletions
@@ -26,6 +26,7 @@ #include "run-command.h" #include "setup.h" #include "shallow.h" +#include "tree.h" #include "hook.h" static struct commit_extra_header *read_commit_extra_header_lines(const char *buf, size_t len, const char **); @@ -96,6 +97,7 @@ struct commit *lookup_commit_reference_by_name(const char *name) static timestamp_t parse_commit_date(const char *buf, const char *tail) { const char *dateptr; + const char *eol; if (buf + 6 >= tail) return 0; @@ -107,16 +109,56 @@ static timestamp_t parse_commit_date(const char *buf, const char *tail) return 0; if (memcmp(buf, "committer", 9)) return 0; - while (buf < tail && *buf++ != '>') - /* nada */; - if (buf >= tail) + + /* + * Jump to end-of-line so that we can walk backwards to find the + * end-of-email ">". This is more forgiving of malformed cases + * because unexpected characters tend to be in the name and email + * fields. + */ + eol = memchr(buf, '\n', tail - buf); + if (!eol) return 0; - dateptr = buf; - while (buf < tail && *buf++ != '\n') - /* nada */; - if (buf >= tail) + dateptr = eol; + while (dateptr > buf && dateptr[-1] != '>') + dateptr--; + if (dateptr == buf) return 0; - /* dateptr < buf && buf[-1] == '\n', so parsing will stop at buf-1 */ + + /* + * Trim leading whitespace, but make sure we have at least one + * non-whitespace character, as parse_timestamp() will otherwise walk + * right past the newline we found in "eol" when skipping whitespace + * itself. + * + * In theory it would be sufficient to allow any character not matched + * by isspace(), but there's a catch: our isspace() does not + * necessarily match the behavior of parse_timestamp(), as the latter + * is implemented by system routines which match more exotic control + * codes, or even locale-dependent sequences. + * + * Since we expect the timestamp to be a number, we can check for that. + * Anything else (e.g., a non-numeric token like "foo") would just + * cause parse_timestamp() to return 0 anyway. + */ + while (dateptr < eol && isspace(*dateptr)) + dateptr++; + if (!isdigit(*dateptr) && *dateptr != '-') + return 0; + + /* + * We know there is at least one digit (or dash), so we'll begin + * parsing there and stop at worst case at eol. + * + * Note that we may feed parse_timestamp() extra characters here if the + * commit is malformed, and it will parse as far as it can. For + * example, "123foo456" would return "123". That might be questionable + * (versus returning "0"), but it would help in a hypothetical case + * like "123456+0100", where the whitespace from the timezone is + * missing. Since such syntactic errors may be baked into history and + * hard to correct now, let's err on trying to make our best guess + * here, rather than insist on perfect syntax. + */ return parse_timestamp(dateptr, NULL, 10); } |