On Tue, Feb 12, 2019 at 06:33:39PM +0100, René Scharfe wrote:
> Am 12.02.2019 um 08:20 schrieb René Scharfe:
> > That command currently prints the pax comment in tar archives if it
> > looks like a SHA1 hash based on its length. It should continue to do
> > so, and _also_ show longer hashes. Your change makes it _only_ show
> > those longer hashes.
> >
> > So it could check for all known valid hash lengths in turn, or accept
> > any payload length between 40 and the_hash_algo->hexsz, or loosen up
> > totally and show comments of any length.
>
> How about the following patch, as a preparation?
>
> -- >8 --
> From: Rene Scharfe <l.s.r@xxxxxx>
> Subject: [PATCH] get-tar-commit-id: parse comment record
>
> Parse pax comment records properly and get rid of magic numbers for
> acceptable comment length. This simplifies a later change to handle
> longer hashes.
>
> Signed-off-by: Rene Scharfe <l.s.r@xxxxxx>
> ---
> builtin/get-tar-commit-id.c | 13 +++++++++++--
> 1 file changed, 11 insertions(+), 2 deletions(-)
>
> diff --git a/builtin/get-tar-commit-id.c b/builtin/get-tar-commit-id.c
> index 2706fcfaf2..312e44ed05 100644
> --- a/builtin/get-tar-commit-id.c
> +++ b/builtin/get-tar-commit-id.c
> @@ -21,6 +21,8 @@ int cmd_get_tar_commit_id(int argc, const char **argv, const char *prefix)
> char *content = buffer + RECORDSIZE;
> const char *comment;
> ssize_t n;
> + long len;
> + char *end;
>
> if (argc != 1)
> usage(builtin_get_tar_commit_id_usage);
> @@ -32,10 +34,17 @@ int cmd_get_tar_commit_id(int argc, const char **argv, const char *prefix)
> die_errno("git get-tar-commit-id: EOF before reading tar header");
> if (header->typeflag[0] != 'g')
> return 1;
> - if (!skip_prefix(content, "52 comment=", &comment))
> +
> + len = strtol(content, &end, 10);
> + if (errno == ERANGE || end == content || len < 0)
> + return 1;
> + if (!skip_prefix(end, " comment=", &comment))
> + return 1;
> + len -= comment - content;
> + if (len != GIT_SHA1_HEXSZ + 1)
> return 1;
So it turns out I wrote a different patch for this in a later series,
but I like this style better. I'm going to squash my existing patches
together and then rework them such that they're based off yours.
My technique iterated over the entire header, comparing the entire
header, which is much less elegant than this solution.
--
brian m. carlson: Houston, Texas, US
OpenPGP: https://keybase.io/bk2204
Attachment:
signature.asc
Description: PGP signature
