Exercises from Peter van der Linden's Expert C Programming.

A revisiting of a book I'd purchased some time ago but had never gotten around to really reading through.

TBD. Although the book is somewhat dated, it's fun to read and still useful. In fact, I actually started reading and doing the book's programming exercises during my 2023 year-end vacation, so at least to me there is both educational and entertainment value to be found. Can't say no to having fun while learning.

For Make, one can build a debugging, i.e. no optimization build, with

make

To build with optimization, one can specify CONFIG=Release, e.g.

make CONFIG=Release

To enable AddressSanitizer and specify a different compiler, in this case Clang, one can use

make CC=clang ENABLE_ASAN=1

To build the project's libraries as static instead of shared, one can use

make BUILD_SHARED=

BUILD_SHARED is set by default and results in shared libraries being built.

By default, if a C++ compiler is available and if Google Test >=1.10.0 is locatable via pkg-config, unit tests will also be built. If one of these components is missing, no tests will be built. One can also disable test building with

make BUILD_TESTS=

Using the provided build.sh script, one can build a debugging build with

./build.sh

To build a release build, use

./build.sh -c Release

To enable AddressSanitizer and specify a different C compiler, one can use

./build.sh -Ca -DCMAKE_C_COMPILER=clang -DENABLE_ASAN=ON

To build the project's libraries as static instead of shared, one can use

./build.sh -Ca -DBUILD_SHARED_LIBS=OFF

BUILD_SHARED_LIBS is set by default and results in shared libraries being built.

If a C++ compiler and Google Test >=1.10.0 are available, unit tests will be built by default. If one of these components is missing, no tests will be built. One can also explicitly choose to disable test building with

./build.sh -Ca -DBUILD_TESTS=OFF

This project also includes some GDB pretty printers for the pdxcp_fruit C++ library that is built if a compiler with C++17 support is available. Since the "fruit" classes are simple, so are the pretty printer Python classes, which makes the printers in this project a nice and clean example of how to write GDB pretty printers in Python.

To load the pretty printers, start GDB in the top-level directory of this repo. If auto-loading of .gdbinit is declined due to GDB auto-load safe path settings, you can source the commands youself with

source .gdbinit

This will load the pdxcp_fruit pretty printers so that e.g. when a pdxcp::apple is printed, one sees

(gdb) print a
$1 = pdxcp::apple(weight=5.2800000000000002)

Contrast this with a raw printout when derived type printing using the vtable is off (see GDB print settings):

(gdb) print a
$1 = {<pdxcp::fruit> = {
    _vptr.fruit = 0x7fffff7aed38 <vtable for pdxcp::apple+16>,
    weight_ = 5.2800000000000002, kcal_per_oz_ = 15}, <No data fields>}

Of course, with .gdbinit commands sourced, one can recover the raw printout with vtable info using:

(gdb) print /r a
$2 = (pdxcp::apple) {<pdxcp::fruit> = {
    _vptr.fruit = 0x7fffff7aed38 <vtable for pdxcp::apple+16>,
    weight_ = 5.2800000000000002, kcal_per_oz_ = 15}, <No data fields>}