Skip to content

Latest commit

 

History

History
633 lines (458 loc) · 23.5 KB

README.md

File metadata and controls

633 lines (458 loc) · 23.5 KB

Cinch CMake Build Utilities

Cinch is a set of utilities and configuration options designed to make cmake builds easy to use and manage.

Installation

Cinch uses standard CMake install features. However, because Cinch depends on its own command-line tool (Cinch-Utils) to build its documentation, it must be installed in the stages documented in this section.

Install the Cinch-Utils Command-Line Tool

Directions for installing cinch-utils are here.

Install Documentation

To install the Cinch documentation, you should run cmake in the Cinch build directory with documentation enabled:

% cmake -DCMAKE_INSTALL_PREFIX=/path/to/install -DENABLE_DOCUMENTATION=ON ..
% make install

General Features

Recursive Sub-Project Structure

The Cinch build system is designed to make modular code development easy. By modular, we mean that subprojects can be incorporated into a Cinch-based top-level project, and they will be automatically added to the top-level project's build targets. This makes it easy to create new projects that combine the capabilities of a set of subprojects. This allows users to build up functionality and control the functionality of the top-level project.

Prevent In-Place Builds

Cinch prohibits users from creating in-place builds, i.e., builds that are rooted in the top-level project directory of a Cinch project. If the user attempts to configure such a build, cmake will exit with an error and instructions for how to clean up and create an out-of-source build.

Basic Structure

Cinch eases build system maintainence by imposing a specific structure on the project source layout.

project/
        app/ (optional application subdirectory)
        cinch/
        CMakeLists.txt -> cinch/cmake/ProjectLists.txt
        config/
               documentation.cmake
               packages.cmake
               project.cmake
        doc/
        src/ (optional library source subdirectory)
            CMakeLists.txt -> cinch/cmake/SourceLists.txt

You may also have any number of submodules under the project directory.

Description of Basic Structure

project

The project top-level directory.

app

An application target subdirectory. Application targets can be added using the cinch_add_application_directory documented below. This subdirectory should contain a CMakeLists.txt file that adds whatever cmake targets are needed for the specific application.

cinch

The Cinch subdirectory. This should be checked-out from the Cinch git server: 'git clone --recursive [email protected]:losalamos/cinch.git'.

CMakeLists.txt

Create a file, whichs sets cmake_minimum_required() and includes the Cinch ProjectLists.txt file.

config

The project configuration directory. This directory is covered in detail below.

doc

The documentation subdirectory. This subdirectory should contain configuration files for Cinch-generated guide documentation, and for doxygen interface documentation.

src

A library target source subdirectory. Library targets can be added using the cinch_add_library_target documented below.

Config Subdirectory

The config subdirectory must contain the following files that provide specialization of the project. Although all of the files must exist, the only file that is required to have content is the project.cmake file.

project.cmake

This file cannot be empty. At a minimum, it must specify the name of the top-level project by calling the CMake project function to set the name, version and enabled languages for the entire project. For more documentation, at a prompt on a machine with a valid CMake installation, type:

% cmake --help project

Additionally, this file may call the following Cinch function (They may also be left Null):

  • cinch_add_application_directory (documented here)

    Add a project-specific build directory that should be included by CMake when searching for list files. This directory should contain a valid CMakeLists.txt file that configures additional build targets.

  • cinch_add_library_target (documented here)

    Add a library target to build for this project.

  • cinch_add_subproject (documented here)

    Add a subproject to this project.

packages.cmake

This file is used to specify CMake find_package requirements for locating installed third-party packages. The content of this file can be any set of valid CMake commands. Values that are set in this file will be available to low-level CMakeLists.txt files for configuring source-level build options.

documentation.cmake

This file is used to add documentation targets with the cinch_add_doc interface (Doxygen documentation is handled separately).

Command-Line Options

Cinch provides various command-line options that may be passed on the cmake configuration line to affect its behavior.

Development Mode

CMake Option: ENABLE_CINCH_DEVELOPMENT (default OFF)

Put Cinch into development mode. This option effects some of the information that is generated by Cinch which is helpful for non-release candidates. If this option is enabled, it will turn on the following features:

  • Documentation Target Annotation
    Documentation targets will have colorized output indicating the source inputs for each section of the documentation.

Verbose Mode

CMake Option: ENABLE_CINCH_VERBOSE (default OFF)

Enable more detailed build output.

Guide Documentation

CMake Option: ENABLE_DOCUMENTATION (default OFF)

Cinch has a powerful documentation facility implemented using the Cinch command-line utility and Pandoc. To create documentation, define a configuration file for each document that should be created in the 'doc' subdirectory. Then, add markdown (.md) or latex (.tex) files to the source tree that document whichever aspects of the project should be included. The caveat is that these documentation fragments should have a special comment header at the beginning of each, of the form:

<!-- CINCHDOC DOCUMENT(Name of Document) SECTION(Name of Section) -->

This special header indicates for which document the fragment is intended and the section within which it should appear. Headers may span multiple lines provided that <!-- CINCHDOC begins the comment. If no attributes (DOCUMENT, SECTION, etc.) are specified, the utility will use a default document and section ('Default' and 'Default'). Multiple fragments intended for different documents and sections may be included within a single input file. For latex fragments, use a header of the form:

% CINCHDOC DOCUMENT(Name of Document) SECTION(Name of Section)

Latex-style CINCHDOC headers must be on a single line.

Build targets can be added to the documentation.cmake file in the config directory. Each target should be created by calling:

cinch_add_doc(target-name config.py top-level-search-directory output)

target-name The build target label, i.e., a make target will be created such that 'make target-name' can be called to generate the documentation target.

config.py A configuration file that must live in the 'doc' subdirectory of the top-level directory of your project. This file should contain a single python dictionary opts that sets the Cinch command-line interface options for your docuement.

top-level-search-directory The relative path to the head of the directory tree within which to search for markdown documentation files.

output The name of the output file that should be produced by pandoc.

Doxygen Documentation

CMake Option: ENABLE_DOXYGEN (default OFF) CMake Option: ENABLE_DOXYGEN_WARN (default OFF)

Cinch supports interface documentation using Doxygen. The doxygen configuration file should be called 'doxygen.conf.in' and should reside in the 'doc' subdirectory. For documentation on using Doxygen, please take a look at the Doxygen Homepage.

If ENABLE_DOXYGEN_WARN is set to ON, normal Doxygen diagnostics and warnings will not be suppressed.

Unit Tests

CMake Option: ENABLE_UNIT_TESTS (default OFF)

Cinch has support for unit testing using a combination of CTest (the native CMake testing facility) and GoogleTest (for C++ support). If unit tests are enabled, Cinch will create a 'test' target. Unit tests may be added in any subdirectory of the project simply be creating the test source code and adding a target using the 'cinch_add_unit(target [source list])' function.

Cinch will check for a local GoogleTest installation on the system during the Cmake configuration step. If GoogleTest is not found, it will be built by Cinch (GoogleTest source code is included with Cinch).

Cinch Logging Utilities (CLOG)

Options

CMake Option: CLOG_ENABLE_STDLOG (default OFF)

CMake Option: CLOG_STRIP_LEVEL (default "0")

CMake Option: CLOG_TAG_BITS (default "16")

CMake Option: CLOG_COLOR_OUTPUT (default OFF)

Basic Description

Cinch has support for trace, info, warn, error, and fatal log reporting (similar to Google Log). There are two interface styles for logging information using CLOG: Insertion style, e.g.,

clog(info) << "This is some information" << std::endl;

and a method interface, e.g.,

clog_info("This is some information");

Both interface styles are available for all severity levels (discussed below).

NOTE: CLOG is automatically available for Cinch unit tests.

CLOG Interface Macros

  • clog_init(groups)
    Initialize the CLOG runtime, enabling the tag groups specified in groups, e.g.,
clog_init("group1,group2,group3");
  • clog(severity)
    Insertion style output with severity level severity, e.g.,
clog(error) << "This is an error level severity message" << std::endl;
  • clog_severity(message)
    Method style output with severity level severity and output message message. Note that message can be any valid C++ stream, e.g.,
clog_info("The number is " << number);
  • clog_every_n(severity, message, n)
    Output every nth iteration using severity and message. This method is not defined for fatal severity level or assertions.

  • clog_assert(test, message)
    Assert that test is true. If test is false, this call will execute clog_fatal(message).

  • clog_add_buffer(name, ostream, colorized)
    Add the buffer defined by the ostream argument in rdbuf(). The second parameter name is the string name to associate with the buffer, and can be used in subsequent calls to the CLOG buffer interface. The last parameter indicates whether or not the buffer supports color output.

  • clog_enable_buffer(name)
    Enable the buffer identified by name.

  • clog_disable_buffer(name)
    Disable the buffer identified by name.

Controlling CLOG Output: Output Streams

CLOG can write output to multiple output streams at once. Users can control which CLOG log files and output are created by adding and enabling/disabling various output streams. By default, CLOG directs output to std::clog (this is the default C++ log iostream and is not part of CLOG) when the CLOG_ENABLE_STDLOG environment variable is defined. Other output streams must be added by the user application. As an example, if the user application wanted CLOG output to go to a file named output.log, one could do the following:

#include <ofstream>

#include "cinchlog.h"

int main(int argc, char ** argv) {

  // Initialize CLOG with output for all tag groups (discussed below)
  clog_init("all");

  // Open an output stream for "output.log"
  std::ofstream output("output.log");

  // Add the stream to CLOG:
  // param 1 ("output") The string name of the buffer.
  // param 2 (output)   The stream (CLOG will call stream.rdbuf() on this).
  // param 3 (false)    A boolean denoting whether or not the buffer
  //                    supports colorization.
  //
  // Note that output is automatically enabled for buffers when they
  // are added. Buffers can be disable with clog_disable_buffer(string name),
  // and re-enabled with clog_enable_buffer(string name).
  clog_add_buffer("output", output, false);

  // Write some information to the output file (and to std::clog if enabled)
  clog(info) << "This will go to output.log" << std::endl;

  return 0;
} // main

Controlling CLOG Output: Severity Levels

CLOG output can be controlled at compile time by specifying a particular severity level. Any logging messages with a lower severity level than the one specified by CLOG_STRIP_LEVEL will be disabled. Note that this implies that CLOG will produce no output for CLOG_STRIP_LEVEL >= 5.

The different severity levels have the following behavior:

  • trace
    Enabled only for severity level 0 (less than 1)
    Trace output is suitable for fine-grained logging information.

  • info
    Enabled for severity levels less than 2
    Info output is suitable for normal logging information.

  • warn
    Enabled for severity levels less than 3
    Warn output is useful for issuing warnings. When CLOG_COLOR_OUTPUT is enabled, warn messages will be displayed in yellow.

  • error
    Enabled for severity levels less than 4
    Error output is useful for issuing non-fatal errors. When CLOG_COLOR_OUTPUT is enabled, error messages will be displayed in red.

  • fatal
    Enabled for severity levels less than 5
    Fatal error output is useful for issuing fatal errors. Fatal errors print a message, dump the current stack trace, and call std::exit(1). When CLOG_COLOR_OUTPUT is enabled, fatal messages will be displayed in red.

Controlling CLOG Output: Tag Groups

Runtime control of CLOG output is possible by adding scoping sections in the source code. These are referred to as tag groups because the scoped section is labeled with a tag. The number of possible tag groups is controlled by CLOG_TAG_BITS (default 16). Tag groups can be enabled or disabled at runtime by specifying the list of tag groups to the clog_init function. Generally, these are controlled by a command-line flag that is interpreted by the user's application. Here is an example code using GFlags to control output:

#include <gflags/gflags.h>

// Create a command-line flag "--groups" with default value "all"
DEFINE_string(groups, "all", "Specify the active tag groups");

#include "cinchlog.h"

int main(int argc, char ** argv) {

  // Parse the command-line arguments
  gflags::ParseCommandLineFlags(&argc, &argv, true);

  // If the user has specified tag groups with --groups=group1, ...
  // these groups will be enabled. Recall that the default is "all".
  clog_init(FLAGS_groups);

  {
  // Create a new tag scope. Log messages within this scope will
  // only be output if tag group "tag1" or tag group "all" is enabled.
  clog_tag_scope(tag1);

  clog(info) << "Enabled for tag group tag1" << std::endl;

  clog(warn) << "This is a warning in group tag1" << std::endl;
  } // scope

  {
  // Create a new tag scope. Log messages within this scope will
  // only be output if tag group "tag2" or tag group "all" is enabled.
  clog_tag_scope(tag2);

  clog(info) << "Enabled for tag group tag2" << std::endl;

  clog(error) << "This is an error in group tag2" << std::endl;
  } // scope

  clog(info) << "This output is not scoped" << std::endl;

  return 0;
} // main

Example code runs:

% ./example --groups=tag1
% [I1225 11:59:59 example.cc:22] Enabled for tag group tag1
% [W1225 11:59:59 example.cc:24] This is a warning in group tag1
% [I1225 11:59:59 example.cc:37] This output is not scoped

% ./example --groups=tag2
% [I1225 11:59:59 example.cc:32] Enabled for tag group tag1
% [E1225 11:59:59 example.cc:34] This is an error in group tag2
% [I1225 11:59:59 example.cc:37] This output is not scoped

% ./example
% [I1225 11:59:59 example.cc:22] Enabled for tag group tag1
% [W1225 11:59:59 example.cc:24] This is a warning in group tag1
% [I1225 11:59:59 example.cc:32] Enabled for tag group tag1
% [E1225 11:59:59 example.cc:34] This is an error in group tag2
% [I1225 11:59:59 example.cc:37] This output is not scoped

Advanced Topics: Predicated Output

The normal CLOG interface is implemented through a set of macros. Advanced users, who need greater control over CLOG, can create their own interfaces (macro or otherwise) to directly access the low-level CLOG interface. Log messages in CLOG derive from the cinch::log_message_t type, which provides a constructor, virtual destructor, and a virtual stream method:

template<typename P>
struct log_message_t
{

  // Constructor:
  // param 1 (file)      The originating file of the message (__FILE__)
  // param 2 (line)      The originating line of the mesasge (__LINE__)
  // param 3 (predicate) A predicate function that can be used to
  //                     control output.
  log_message_t(
    const char * file,
    int line,
    P && predicate
  )
  {
    // See cinchlog.h for implementation.
  } // log_message_t

  // Destructor.
  virtual
  ~log_message_t()
  {
    // See cinchlog.h for implementation.
  } // ~log_message_t

  // Stream method.
  virtual
  std::ostream &
  stream()
  {
    // See cinchlog.h for implementation.
  } // stream

}; // struct log_message_t

Users wishing to customize CLOG can change the default behavior by overriding the virtual methods of this type, and by providing custom predicates. Much of the basic CLOG functionality is implemented in this manner, e.g., the following code implements the trace level severity output:

#define severity_message_t(severity, P, format)                                \
struct severity ## _log_message_t                                              \
  : public log_message_t<P>                                                    \
{                                                                              \
  severity ## _log_message_t(                                                  \
    const char * file,                                                         \
    int line,                                                                  \
    P && predicate = true_state)                                               \
    : log_message_t<P>(file, line, predicate) {}                               \
                                                                               \
  ~severity ## _log_message_t()                                                \
  {                                                                            \
    /* Clean colors from the stream */                                         \
    clog_t::instance().stream() << COLOR_PLAIN;                                \
  }                                                                            \
                                                                               \
  std::ostream &                                                               \
  stream() override                                                            \
    /* This is replaced by the scoped logic */                                 \
    format                                                                     \
};

//----------------------------------------------------------------------------//
// Define the insertion style severity levels.
//----------------------------------------------------------------------------//

#define message_stamp \
  timestamp() << " " << rstrip<'/'>(file_) << ":" << line_

severity_message_t(trace, decltype(cinch::true_state),
  {
#if CLOG_STRIP_LEVEL < 1
    if(clog_t::instance().tag_enabled() && predicate_()) {
      std::ostream & stream = clog_t::instance().stream();
      stream << OUTPUT_CYAN("[T") << OUTPUT_LTGRAY(message_stamp);
      stream << OUTPUT_CYAN("] ");
      return stream;
    }
    else {
      return clog_t::instance().null_stream();
    } // if
#else
    return clog_t::instance().null_stream();
#endif
  });

Interested users should look at the source code for more examples.

Versioning

CMake Option: VERSION_CREATION (default 'git describe')

Cinch can automatically create version information for projects that use git. This feature uses the 'git describe' function, which creates a version from the most recent annotated tag with a patch level based on the number of commits since that tag and a partial hash key. For example, if the most recent annotated tag is "1.0" and there have been 35 commits since, the Cinch-created version would be similar to: 1.0-35-g2f657a

For actual releases, this approach may not be optimal. In this case, Cinch allows you to override the automatic versioning by specifying a static version to cmake via the VERSION_CREATION option. Simply set this to the desired version and it will be used.

Release

This software has been approved for open source release and has been assigned LA-CC-15-070.

Copyright

Copyright (c) 2016, Los Alamos National Security, LLC All rights reserved.

Copyright 2016. Los Alamos National Security, LLC. This software was produced under U.S. Government contract DE-AC52-06NA25396 for Los Alamos National Laboratory (LANL), which is operated by Los Alamos National Security, LLC for the U.S. Department of Energy. The U.S. Government has rights to use, reproduce, and distribute this software. NEITHER THE GOVERNMENT NOR LOS ALAMOS NATIONAL SECURITY, LLC MAKES ANY WARRANTY, EXPRESS OR IMPLIED, OR ASSUMES ANY LIABILITY FOR THE USE OF THIS SOFTWARE. If software is modified to produce derivative works, such modified software should be clearly marked, so as not to confuse it with the version available from LANL.

Additionally, redistribution and use in source and binary forms, with or without modification, are permitted provided that the following conditions are met:

  1. Redistributions of source code must retain the above copyright notice, this list of conditions and the following disclaimer.

  2. Redistributions in binary form must reproduce the above copyright notice, this list of conditions and the following disclaimer in the documentation and/or other materials provided with the distribution.

  3. Neither the name of Los Alamos National Security, LLC, Los Alamos National Laboratory, LANL, the U.S. Government, nor the names of its contributors may be used to endorse or promote products derived from this software without specific prior written permission.

THIS SOFTWARE IS PROVIDED BY LOS ALAMOS NATIONAL SECURITY, LLC AND CONTRIBUTORS "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL LOS ALAMOS NATIONAL SECURITY, LLC OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.