public inbox for bzip2-devel@sourceware.org
 help / color / mirror / Atom feed
From: Julian Seward <jseward@acm.org>
To: Mark Wielaard <mark@klomp.org>,
	Federico Mena Quintero <federico@gnome.org>,
	bzip2-devel@sourceware.org
Subject: Re: Alternative nSelectors patch (Was: bzip2 1.0.7 released)
Date: Tue, 01 Jan 2019 00:00:00 -0000	[thread overview]
Message-ID: <4f434101-5ce3-d757-2f61-c9e419911e00@acm.org> (raw)
In-Reply-To: <308d9e82220760205ee673bf0505ee1815d48596.camel@klomp.org>


Hi Mark,

This seems to me like a better patch than my proposal, so I retract my
proposal and vote for this one instead.

The one thing that concerned me was that, it would be a disaster -- having
ignored all selectors above 18002 -- if subsequent decoding actually *did*
manage somehow to try to read more than 18002 selectors out of s->selectorMtf,
because we'd be reading uninitialised memory.  But this seems to me can't
happen because, after the selector-reading loop, you added

+      if (nSelectors > BZ_MAX_SELECTORS)
+        nSelectors = BZ_MAX_SELECTORS;

and the following loop:

       /*--- Undo the MTF values for the selectors. ---*/
       ...

is the only place that reads s->selectorMtf, and then only for the range
0 .. nSelectors-1.

So it seems good to me.  Does this sync with your analysis?

J


On 01/07/2019 01:36, Mark Wielaard wrote:
> Hi,
> 
> On Fri, 2019-06-28 at 13:10 +0200, Mark Wielaard wrote:
>>> It seems to me to be important to now split BZ_MAX_SELECTORS into these two
>>> parts so as to make it clear to everybody that we're accepting (decompressing)
>>> a slightly larger set of inputs than we create (a la that old saying about
>>> network protocol implementations), so as to tolerate other compressors.
>>
>> That seems good. The attached patch does this and makes it possible to
>> decode the problematic bz2 file.
> 
> Sorry, it is a bit too late here to properly document this patch and
> explain why I think it is a better one than the "split-max-selectors"
> fix. But hopefully the new testsuite example and the comment in the
> patch make clear what my thinking is.
> 
> This resolved both the issue with the large file reported as with the
> new test suite file (lbzip2/32767.bz2). The whole testsuite passes now,
> even under valgrind and with gcc -fsanitize=undefined.
> 
> Comments on the patch idea more than welcome.
> 
> Thanks,
> 
> Mark
> 

  reply	other threads:[~2019-07-02  6:34 UTC|newest]

Thread overview: 16+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2019-01-01  0:00 bzip2 1.0.7 released Mark Wielaard
2019-01-01  0:00 ` Mark Wielaard
2019-01-01  0:00   ` Federico Mena Quintero
2019-01-01  0:00   ` Mark Wielaard
2019-01-01  0:00     ` Federico Mena Quintero
2019-01-01  0:00       ` Julian Seward
2019-01-01  0:00         ` Mark Wielaard
2019-01-01  0:00           ` bzip2 test suite (Was: bzip2 1.0.7 released) Mark Wielaard
2019-01-01  0:00           ` bzip2 1.0.7 released Mark Wielaard
2019-01-01  0:00             ` Federico Mena Quintero
2019-01-01  0:00               ` Mark Wielaard
2019-01-01  0:00           ` Alternative nSelectors patch (Was: bzip2 1.0.7 released) Mark Wielaard
2019-01-01  0:00             ` Julian Seward [this message]
2019-01-01  0:00               ` Mark Wielaard
2019-01-01  0:00                 ` Mark Wielaard
2019-01-01  0:00   ` bzip2 1.0.7 released Jeffrey Walton

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=4f434101-5ce3-d757-2f61-c9e419911e00@acm.org \
    --to=jseward@acm.org \
    --cc=bzip2-devel@sourceware.org \
    --cc=federico@gnome.org \
    --cc=mark@klomp.org \
    /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 a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).