From: Taylor Blau <me@ttaylorr.com>
To: Jeff King <peff@peff.net>
Cc: git@vger.kernel.org
Subject: Re: [PATCH 0/6] better handling of gigantic config files
Date: Sun, 12 Apr 2020 18:49:27 -0600 [thread overview]
Message-ID: <20200413004927.GC55122@syl.local> (raw)
In-Reply-To: <20200410194211.GA1363484@coredump.intra.peff.net>
Hi Peff,
On Fri, Apr 10, 2020 at 03:42:11PM -0400, Jeff King wrote:
> The fact that parse_config_key() requires its callers to use an "int"
> for a string length has bugged me for a while, and it re-bugged me when
> looking at it today. So I finally decided to do something about it,
> which led to an odyssey of other small fixes and cleanups.
>
> In particular, I was curious what kinds of bad behavior you could
> provoke by having a key name larger than 2GB (especially because we use
> the same parser for .gitmodules files, which might not be trusted). It
> turns out: basically none, because the config parser chokes immediately
> dues to its own int/size_t confusion.
>
> After patch 5, the config system _can_ actually handle stupidly-sized
> config keys, but in the end I decided to explicitly disallow them.
> There's downstream code that would be impossible to fix, and nobody
> actually cares about this case working anyway. See patch 6 for more
> discussion. I do still think the other patches are worth having as a
> cleanup; the more code that is safe from unexpected integer truncation
> the better.
>
> [1/6]: remote: drop auto-strlen behavior of make_branch() and make_rewrite()
> [2/6]: parse_config_key(): return subsection len as size_t
> [3/6]: config: drop useless length variable in write_pair()
> [4/6]: git_config_parse_key(): return baselen as size_t
> [5/6]: config: use size_t to store parsed variable baselen
> [6/6]: config: reject parsing of files over INT_MAX
>
> archive-tar.c | 4 ++--
> builtin/help.c | 2 +-
> builtin/reflog.c | 2 +-
> config.c | 42 +++++++++++++++++++++++++++++-------------
> config.h | 4 ++--
> convert.c | 2 +-
> fsck.c | 2 +-
> ll-merge.c | 2 +-
> promisor-remote.c | 2 +-
> remote.c | 37 +++++++++++++------------------------
> submodule-config.c | 3 ++-
> userdiff.c | 4 ++--
> 12 files changed, 56 insertions(+), 50 deletions(-)
>
> -Peff
Thanks for doing this. I knew that it rang a bell for some reason, but
it was because of the upload-pack changes to limit the set of allowed
object filter choices that I'd sent as an RFC somewhere.
I was using 'parse_config_key()', and I think that you noted somewhere
that it was odd that it filled an int and not a size_t. So, thanks very
much for fixing that case.
This series looks very good and straightforward to me. So,
Reviewed-by: Taylor Blau <me@ttaylorr.com>
Thanks,
Taylor
next prev parent reply other threads:[~2020-04-13 0:49 UTC|newest]
Thread overview: 16+ messages / expand[flat|nested] mbox.gz Atom feed top
2020-04-10 19:42 [PATCH 0/6] better handling of gigantic config files Jeff King
2020-04-10 19:43 ` [PATCH 1/6] remote: drop auto-strlen behavior of make_branch() and make_rewrite() Jeff King
2020-04-10 21:44 ` Junio C Hamano
2020-04-10 19:44 ` [PATCH 2/6] parse_config_key(): return subsection len as size_t Jeff King
2020-04-10 19:44 ` [PATCH 3/6] config: drop useless length variable in write_pair() Jeff King
2020-04-10 21:51 ` Junio C Hamano
2020-04-10 19:46 ` [PATCH 4/6] git_config_parse_key(): return baselen as size_t Jeff King
2020-04-10 19:47 ` [PATCH 5/6] config: use size_t to store parsed variable baselen Jeff King
2020-04-10 19:50 ` [PATCH 6/6] config: reject parsing of files over INT_MAX Jeff King
2020-04-10 22:04 ` Junio C Hamano
2020-04-10 22:15 ` Jeff King
2020-04-13 0:47 ` Taylor Blau
2020-04-13 22:14 ` Junio C Hamano
2020-04-13 0:49 ` Taylor Blau [this message]
2020-04-13 17:20 ` [PATCH 0/6] better handling of gigantic config files Jeff King
2020-04-13 17:23 ` Taylor Blau
Reply instructions:
You may reply publicly to this message via plain-text email
using any one of the following methods:
* Save the following mbox file, import it into your mail client,
and reply-to-all from there: mbox
Avoid top-posting and favor interleaved quoting:
https://en.wikipedia.org/wiki/Posting_style#Interleaved_style
* Reply using the --to, --cc, and --in-reply-to
switches of git-send-email(1):
git send-email \
--in-reply-to=20200413004927.GC55122@syl.local \
--to=me@ttaylorr.com \
--cc=git@vger.kernel.org \
--cc=peff@peff.net \
/path/to/YOUR_REPLY
https://kernel.org/pub/software/scm/git/docs/git-send-email.html
* If your mail client supports setting the In-Reply-To header
via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line
before the message body.
This is an external index of several public inboxes,
see mirroring instructions on how to clone and mirror
all data and code used by this external index.