Message ID | 20230422135001.GA3942563@coredump.intra.peff.net (mailing list archive) |
---|---|
State | Superseded |
Headers | show |
Series | fixing some parse_commit() timestamp corner cases | expand |
Am 22.04.23 um 15:50 schrieb Jeff King: > diff --git a/commit.c b/commit.c > index ede810ac1c..56877322d3 100644 > --- a/commit.c > +++ b/commit.c > @@ -120,6 +120,16 @@ static timestamp_t parse_commit_date(const char *buf, const char *tail) > if (dateptr == buf || dateptr == eol) > return 0; > > + /* > + * trim leading whitespace; parse_timestamp() will do this itself, but > + * it will walk past the newline at eol while doing so. So we insist > + * that there is at least one digit here. > + */ > + while (dateptr < eol && isspace(*dateptr)) > + dateptr++; > + if (!strchr("0123456789", *dateptr)) You could use (our own) isdigit() here instead. It's more concise and efficient. René
On Sat, Apr 22, 2023 at 05:53:10PM +0200, René Scharfe wrote: > Am 22.04.23 um 15:50 schrieb Jeff King: > > diff --git a/commit.c b/commit.c > > index ede810ac1c..56877322d3 100644 > > --- a/commit.c > > +++ b/commit.c > > @@ -120,6 +120,16 @@ static timestamp_t parse_commit_date(const char *buf, const char *tail) > > if (dateptr == buf || dateptr == eol) > > return 0; > > > > + /* > > + * trim leading whitespace; parse_timestamp() will do this itself, but > > + * it will walk past the newline at eol while doing so. So we insist > > + * that there is at least one digit here. > > + */ > > + while (dateptr < eol && isspace(*dateptr)) > > + dateptr++; > > + if (!strchr("0123456789", *dateptr)) > > You could use (our own) isdigit() here instead. It's more concise and > efficient. Heh, yes, that is much better. I had strspn() on the mind since that is what split_ident_line() uses. I think it could even just be: if (dateptr != eol) which implies that we found some non-whitespace character, and then we rely on parse_timestamp() to return 0 (which is what the current code is effectively doing). -Peff
Jeff King <peff@peff.net> writes: > The comment in parse_commit_date() claims that parse_timestamp() will > not walk past the end of the buffer we've been given, since it will hit > the newline at "eol" and stop. This is usually true, when dateptr > contains actual numbers to parse. But with a line like: > > committer name <email> \n I was wondering of this case while reading [2/3] ;-) > ... > In practice this can't cause us to walk off the end of an array, because > we always add an extra NUL byte to the end of objects we load from disk > (as a defense against exactly this kind of bug). However, you can see > the behavior in action when "committer" is the final header (which it > usually is, unless there's an encoding) and the subject line can be > parsed as an integer. We walk right past the newline on the committer > line, as well as the "\n\n" separator, and mistake the subject for the > timestamp. > + /* > + * trim leading whitespace; parse_timestamp() will do this itself, but > + * it will walk past the newline at eol while doing so. So we insist > + * that there is at least one digit here. > + */ "one digit" -> "one non-whitespace". > + while (dateptr < eol && isspace(*dateptr)) > + dateptr++; This is an expected change, but > + if (!strchr("0123456789", *dateptr)) > + return 0; this is not. Isn't the only problematic case that dateptr being at eol? That is what the proposed log message argued. > /* dateptr < eol && *eol == '\n', so parsing will stop at eol */ This comment is slightly stale. dateptr < eol, *eol == '\n', and we know the string starting at dateptr is not a run of whitespace and that is what makes the parsing stop at eol. > diff --git a/t/t4212-log-corrupt.sh b/t/t4212-log-corrupt.sh > index af4b35ff56..d4ef48d646 100755 > --- a/t/t4212-log-corrupt.sh > +++ b/t/t4212-log-corrupt.sh > @@ -92,4 +92,33 @@ test_expect_success 'absurdly far-in-future date' ' > git log -1 --format=%ad $commit > ' > > +test_expect_success 'create commit with whitespace committer date' ' > + # It is important that this subject line is numeric, since we want to > + # be sure we are not confused by skipping whitespace and accidentally > + # parsing the subject as a timestamp. Nice. > + # Do not use munge_author_date here. Besides not hitting the committer > + # line, it leaves the timezone intact, and we want nothing but > + # whitespace. > + test_commit 1234567890 && > + git cat-file commit HEAD >commit.orig && > + sed "s/>.*/> /" <commit.orig >commit.munge && > + ws_commit=$(git hash-object --literally -w -t commit commit.munge) > +' > + > +test_expect_success '--until treats whitespace date as sentinel' ' > + echo $ws_commit >expect && > + git rev-list --until=1980-01-01 $ws_commit >actual && > + test_cmp expect actual > +' > + > +test_expect_success 'pretty-printer handles whitespace date' ' > + # as with the %ad test above, we will show these as the empty string, > + # not the 1970 epoch date. This is intentional; see 7d9a281941 (t4212: > + # test bogus timestamps with git-log, 2014-02-24) for more discussion. > + echo : >expect && > + git log -1 --format="%at:%ct" $ws_commit >actual && > + test_cmp expect actual > +' > + > test_done
On Mon, Apr 24, 2023 at 11:01:26AM -0700, Junio C Hamano wrote: > > + /* > > + * trim leading whitespace; parse_timestamp() will do this itself, but > > + * it will walk past the newline at eol while doing so. So we insist > > + * that there is at least one digit here. > > + */ > > "one digit" -> "one non-whitespace". > > > + while (dateptr < eol && isspace(*dateptr)) > > + dateptr++; > > This is an expected change, but > > > + if (!strchr("0123456789", *dateptr)) > > + return 0; > > this is not. Isn't the only problematic case that dateptr being at > eol? That is what the proposed log message argued. Yes, that would be sufficient. I was moving things slightly closer to what split_ident_line() does by actually checking for numbers. But that led to the final paragraph in the commit message explaining how it all ends up the same either way. So I'll swap this out for: if (dateptr == eol) which I think requires less explanation, as it leaves the function more like it was originally (and the behavior is the same either way). > > /* dateptr < eol && *eol == '\n', so parsing will stop at eol */ > > This comment is slightly stale. dateptr < eol, *eol == '\n', and we > know the string starting at dateptr is not a run of whitespace and > that is what makes the parsing stop at eol. Yeah, I hoped the extra context of the earlier comment would be enough. ;) But it is probably better to spell it out by expanding this comment. The code is certainly tricky enough. -Peff
On Sat, Apr 22, 2023 at 08:37:15PM -0400, Jeff King wrote: > > You could use (our own) isdigit() here instead. It's more concise and > > efficient. > > Heh, yes, that is much better. I had strspn() on the mind since that is > what split_ident_line() uses. > > I think it could even just be: > > if (dateptr != eol) > > which implies that we found some non-whitespace character, and then we > rely on parse_timestamp() to return 0 (which is what the current code is > effectively doing). This should be "dateptr == eol" of course, because the body of the conditional is "return 0" to signal an error. It's correct in the v2 of the series I just sent out. -Peff
diff --git a/commit.c b/commit.c index ede810ac1c..56877322d3 100644 --- a/commit.c +++ b/commit.c @@ -120,6 +120,16 @@ static timestamp_t parse_commit_date(const char *buf, const char *tail) if (dateptr == buf || dateptr == eol) return 0; + /* + * trim leading whitespace; parse_timestamp() will do this itself, but + * it will walk past the newline at eol while doing so. So we insist + * that there is at least one digit here. + */ + while (dateptr < eol && isspace(*dateptr)) + dateptr++; + if (!strchr("0123456789", *dateptr)) + return 0; + /* dateptr < eol && *eol == '\n', so parsing will stop at eol */ return parse_timestamp(dateptr, NULL, 10); } diff --git a/t/t4212-log-corrupt.sh b/t/t4212-log-corrupt.sh index af4b35ff56..d4ef48d646 100755 --- a/t/t4212-log-corrupt.sh +++ b/t/t4212-log-corrupt.sh @@ -92,4 +92,33 @@ test_expect_success 'absurdly far-in-future date' ' git log -1 --format=%ad $commit ' +test_expect_success 'create commit with whitespace committer date' ' + # It is important that this subject line is numeric, since we want to + # be sure we are not confused by skipping whitespace and accidentally + # parsing the subject as a timestamp. + # + # Do not use munge_author_date here. Besides not hitting the committer + # line, it leaves the timezone intact, and we want nothing but + # whitespace. + test_commit 1234567890 && + git cat-file commit HEAD >commit.orig && + sed "s/>.*/> /" <commit.orig >commit.munge && + ws_commit=$(git hash-object --literally -w -t commit commit.munge) +' + +test_expect_success '--until treats whitespace date as sentinel' ' + echo $ws_commit >expect && + git rev-list --until=1980-01-01 $ws_commit >actual && + test_cmp expect actual +' + +test_expect_success 'pretty-printer handles whitespace date' ' + # as with the %ad test above, we will show these as the empty string, + # not the 1970 epoch date. This is intentional; see 7d9a281941 (t4212: + # test bogus timestamps with git-log, 2014-02-24) for more discussion. + echo : >expect && + git log -1 --format="%at:%ct" $ws_commit >actual && + test_cmp expect actual +' + test_done
The comment in parse_commit_date() claims that parse_timestamp() will not walk past the end of the buffer we've been given, since it will hit the newline at "eol" and stop. This is usually true, when dateptr contains actual numbers to parse. But with a line like: committer name <email> \n with just whitespace, and no numbers, parse_timestamp() will consume that newline as part of the leading whitespace, and we may walk past our "tail" pointer (which itself is set from the "size" parameter passed in to parse_commit_buffer()). In practice this can't cause us to walk off the end of an array, because we always add an extra NUL byte to the end of objects we load from disk (as a defense against exactly this kind of bug). However, you can see the behavior in action when "committer" is the final header (which it usually is, unless there's an encoding) and the subject line can be parsed as an integer. We walk right past the newline on the committer line, as well as the "\n\n" separator, and mistake the subject for the timestamp. The new test demonstrates such a case. I also added a test to check this case against the pretty-print formatter, which uses split_ident_line(). It's not subject to the same bug, because it insists that there be one or more digits in the timestamp. We can use the same logic here. If there's a non-whitespace but non-digit value (say "committer name <email> foo"), then parse_timestamp() would already have returned 0 anyway. So the only change should be for this "whitespace only" case. Signed-off-by: Jeff King <peff@peff.net> --- commit.c | 10 ++++++++++ t/t4212-log-corrupt.sh | 29 +++++++++++++++++++++++++++++ 2 files changed, 39 insertions(+)