Re: [pcre-dev] Here is pcre-7.1-RC2 for you to play with

On Tue, 20 Mar 2007, Sheri wrote:

> in test 2 there are numerous individual tests report "size disagreement".
>
> e.g.,
> Size disagreement: pcre_fullinfo=62 call to malloc for 0
>
> also some differences in "Memory allocation (code space)"

These are all the same. There is a static variable called "gotten_store"
in pcretest, which is set when PCRE calls the private function
new_malloc() to get memory. This static is used when generating those
messages - it seems incorrectly to be zero in your case. (The 0 in the
message above should be 62.)

Either: for some reason PCRE isn't calling pcretest's malloc function,
or the static variable isn't getting set correctly. This smacks of some
kind of linkage problem or variable scope problem, but I haven't any
idea as to what to suggest. Someone with Windows programming experience
is needed to figure out what is going on (and if anything can be done to
get rid of those errors).

> test 7
>
> 3 tests report "No Match" (all using callouts)

Hmm. Callouts are another case where the address of a function is passed
around.

> Below is my configure command, which I followed immediated by "make":
>
> CFLAGS='-O3 -s' ./configure --enable-utf8 --enable-newline-is-any
> --enable-unicode-properties --disable-stack-for-recursion
> --disable-static --disable-cpp

I wonder if the --disable-stack-for-recursion might be influencing the
malloc things? Have you tried without that?

Philip

--
Philip Hazel, University of Cambridge Computing Service.