Re: pcre 8.37 with latest patches won't build

[Michael Tremer <michael.tremer@ipfire.org>]

[-- Attachment #1: Type: text/plain, Size: 2662 bytes --]

Hi,

sorry, this is not your fault. I fucked that up.

I was sure that these patches built but clearly I did something wrong.

I reverted the commit so that you can pull and resume your build and
will then upload a new commit after I fixed this.

Thanks for the notice.

Best,
-Michael

On Wed, 2015-08-19 at 20:47 +0200, Matthias Fischer wrote:
> Hi,
> 
> yesterday I tried to build the latest 'next', but it failed. It
> couldn't 
> apply 'pcre-8.37-Fix-another-buffer-overflow.patch':
> 
> ***SNIP***
> Aug 18 19:03:54: Building pcre pcre-8.37.tar.gz checksum OK
> + cd /usr/src/lfs
> + make -f pcre LFS_BASEDIR=/usr/src install
> ====================================== Installing pcre-8.37 ...
> Install started; saving file list to /usr/src/lsalr ...
> cd /usr/src/pcre-8.37 && patch -Np1 < 
> /usr/src/src/patches/pcre-8.37-Fix-buffer-overflow-for-named
> -recursive-back-referen.patch
> patching file pcre_compile.c
> patching file testdata/testinput2
> patching file testdata/testoutput2
> cd /usr/src/pcre-8.37 && patch -Np1 < 
> /usr/src/src/patches/pcre-8.37-Fix-buffer-overflow-for-forward
> -reference-within-bac.patch
> patching file pcre_compile.c
> patching file testdata/testinput2
> patching file testdata/testoutput2
> cd /usr/src/pcre-8.37 && patch -Np1 < 
> /usr/src/src/patches/pcre-8.37-Fix-another-buffer-overflow.patch
> patching file pcre_compile.c
> Hunk #1 FAILED at 7210.
> 1 out of 1 hunk FAILED -- saving rejects to file pcre_compile.c.rej
> patching file testdata/testinput2
> patching file testdata/testoutput11-16
> patching file testdata/testoutput11-32
> patching file testdata/testoutput11-8
> patching file testdata/testoutput2
> make: *** [/usr/src/log/pcre-8.37] Error 1
> ***SNAP***
> 
> 'pcre_compile.c.rej' says:
> 
> ***SNIP***
> --- pcre_compile.c
> +++ pcre_compile.c
> @@ -7210,7 +7210,12 @@
>             real compile this will be picked up and the reference 
> wrapped with
>             OP_ONCE to make it atomic, so we must space in case this 
> occurs. */
> 
> -          if (recno == 0) *lengthptr += 2 + 2*LINK_SIZE;
> +          /* In fact, this can happen for a non-forward reference
> because
> +          another group with the same number might be created later.
> This
> +          issue is fixed "properly" in PCRE2. As PCRE1 is now in 
> maintenance
> +          only mode, we finesse the bug by allowing more memory
> always. */
> +
> +          /* if (recno == 0) */ *lengthptr += 2 + 2*LINK_SIZE;
>             }
> 
>           /* In the real compile, search the name table. We check the
> name
> ***SNAP***
> 
> Deleting this section didn't help. Anyone else or my fault?
> 
> Regards
> Matthias

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 819 bytes --]