PCRE-7.6

[Randy McMurchy]

Version increment to 7.4

​ftp://ftp.csx.cam.ac.uk/pub/software/programming/pcre/

[Ag. Hatzimanikas]

Version increment to 7.6

[Ag. Hatzimanikas]

Please check this bug (1) first before updating.

Working patch also in comment 3.

​http://bugs.exim.org/show_bug.cgi?id=664

[Ag. Hatzimanikas]

Glib can use the system pcre instead of its internal pcre. But in order to do that it needs both unicode related switches to be enabled during pcre configuration/installation:

--enable-utf8 --enable-unicode-properties

Another important thing to mention is that its possible for pcregrep to read compressed files (.gz and .bz2) by using zlib and bz2 libraries (both are present in a base LFS system therefore there is no need for additional dependencies).

To add this kind of support two configure switches needed:

--enable-pcregrep-libz --enable-pcregrep-libbz2

Another Note: Pcre now can be built using Cmake.

And a last one. Pcre versions prior to 7.6 are vulnerable because of a buffer overflow problem when run in UTF-8 mode.

Many distributions using the pcre library (glib, zsh, and Python at least). Of all of these three, glib and Python needed to upgraded to newer versions because of the vulnerability, and that's because both of them are using an internal (old) pcre.

[Ag. Hatzimanikas]

Patch submitted. I also raised up severity/priority. I can update the package if I have another opinion about the switches than need to added as default or as optional.

My opinion is to enable all of them by default, since they add value to the end result and costs nothing.

[Ag. Hatzimanikas]

Fixed in r7208.