This package contains a test suite for libffi.

This test suite can be compiled with a C compiler. No need for ‘expect’ or some other package that is often not installed.

The test suite consists of 81 C functions, each with a different signature.

  • test-call verifies that calling each function directly produces the same results as calling the function indirectly through ‘ffi_call’.

  • test-callback verifies that calling each function directly produces the same results as calling a function that is a callback (object build by ‘ffi_prep_closure_loc’) and simulates the original function.

Each direct or indirect invocation should produce one line of output to stdout. A correct output consists of paired lines, such as

void f(void): void f(void): int f(void):->99 int f(void):->99 int f(int):(1)->2 int f(int):(1)->2 int f(2*int):(1,2)->3 int f(2*int):(1,2)->3 …

The Makefile then creates two files:

  • failed-call, which consists of the non-paired lines of output of ‘test-call’,

  • failed-callback, which consists of the non-paired lines of output of ‘test-callback’.

The test suite passes if both failed-call and failed-callback come out as empty.

How to use the test suite


  1. Modify the Makefile’s variables prefix = the directory in which libffi was installed CC = the C compiler, often with options such as “-m32” or “-m64”

    that enforce a certain ABI,

    CFLAGS = optimization options (need to change them only for non-GCC

    compilers)
  2. Run “make”. If it fails already in “test-call”, run also “make check-callback”.

  3. If this failed, inspect the output files.

How to interpret the results


The failed-call and failed-callback files consist of paired lines: The first line is the result of the direct invocation. The second line is the result of invocation through libffi.

For example, this output

uchar f(uchar,ushort,uint,ulong):(97,2,3,4)->255 uchar f(uchar,ushort,uint,ulong):(97,2,3,4)->0

indicates that the arguments were passed correctly, but the return value came out wrong.

And this output

float f(17*float,3*int,L):(0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.1,1.2,1.3,1.4,1.5,1.6,1.7,1.8,6,7,8,561,1105,1729,2465,2821,6601)->15319.1 float f(17*float,3*int,L):(0.1,0.2,0.3,0.4,0.5,0.6,0.7,0.8,0.9,1.1,1.2,1.3,1.4,1.5,1.6,1.7,1.8,-140443648,10,268042216,-72537980,-140443648,-140443648,-140443648,-140443648,-140443648)->-6.47158e+08

indicates that integer arguments that come after 17 floating-point arguments were not passed correctly.

Credits


The test suite is based on the one of GNU libffcall-2.0. Authors: Bill Triggs, Bruno Haible