I've just released gflags 2.1.2.
This release completes the namespace change fixes. In particular, it restores binary ABI compatibility with release version 2.0. The deprecated "google" namespace is by default still kept as primary namespace while symbols are imported into the new "gflags" namespace. This can be overridden using the CMake variable GFLAGS_NAMESPACE.
Other fixes of the build configuration are related to the (patched) CMake modules FindThreads.cmake and CheckTypeSize.cmake. These have been removed and instead the C language is enabled again even though gflags is written in C++ only.
This release also marks the complete move of the gflags project from Google Code to GitHub. Email addresses of original issue reporters got lost in the process. Given the age of most issue reports, this should be negligable.
Please report any further issues using the GitHub issue tracker.
I've just released gflags 2.1.1.
This release fixes a few bugs in the configuration of gflags_declare.h and adds a separate GFLAGS_INCLUDE_DIR CMake variable to the build configuration. Setting GFLAGS_NAMESPACE to "google" no longer changes also the include path of the public header files. This allows the use of the library with other Google projects such as glog which still use the deprecated "google" namespace for the gflags library, but include it as "gflags/gflags.h".
I've just released gflags 2.1.
The major changes are the use of CMake for the build configuration instead of the autotools and packaging support through CPack. The default namespace of all C++ symbols is now "gflags" instead of "google". This can be configured via the GFLAGS_NAMESPACE variable.
This release compiles with all major compilers without warnings and passed the unit tests on Ubuntu 12.04, Windows 7 (Visual Studio 2008 and 2010, Cygwin, MinGW), and Mac OS X (Xcode 5.1).
The SVN repository on Google Code is now frozen and replaced by a Git repository such that it can be used as Git submodule by projects. The main hosting of this project remains at Google Code. Thanks to the distributed character of Git, I can push (and pull) changes from both GitHub and Google Code in order to keep the two public repositories in sync. When fixing an issue for a pull request through either of these hosting platforms, please reference the issue number as described here. For the further development, I am following the Git branching model with feature branch names prefixed by "feature/" and bugfix branch names prefixed by "bugfix/", respectively.
Binary and source packages are available on GitHub.
The migration of the build system to CMake is almost complete. What remains to be done is rewriting the tests in Python such they can be executed on non-Unix platforms and splitting them up into separate CTest tests. Though merging these changes into the master branch yet remains to be done, it is recommended to already start using the cmake-migration branch.
More than a year has past since I (Andreas) took over the maintenance for
gflags. Only few minor changes have been made since then, much to my regret.
To get more involved and stimulate participation in the further
development of the library, I moved the project source code today to
GitHub.
I believe that the strengths of Git will allow for better community collaboration
as well as ease the integration of changes made by others. I encourage everyone
who would like to contribute to send me pull requests.
Git's lightweight feature branches will also provide the right tool for more
radical changes which should only be merged back into the master branch
after these are complete and implement the desired behavior.
The SVN repository remains accessible at Google Code and I will keep the master branch of the Git repository hosted at GitHub and the trunk of the Subversion repository synchronized. Initially, I was going to simply switch the Google Code project to Git, but in this case the SVN repository would be frozen and force everyone who would like the latest development changes to use Git as well. Therefore I decided to host the public Git repository at GitHub instead.
Please continue to report any issues with gflags on Google Code. The GitHub project will only be used to host the Git repository.
One major change of the project structure I have in mind for the next weeks is the migration from autotools to CMake. Check out the (unstable!) cmake-migration branch on GitHub for details.
I've just released gflags 2.0.
The google-gflags project has been renamed to gflags. I
(csilvers) am stepping down as maintainer, to be replaced by Andreas
Schuh. Welcome to the team, Andreas! I've seen the energy you have
around gflags and the ideas you have for the project going forward,
and look forward to having you on the team.
I bumped the major version number up to 2 to reflect the new community
ownership of the project. All the changes
are related to the renaming. There are no functional changes from
gflags 1.7. In particular, I've kept the code in the namespace
google, though in a future version it should be renamed to gflags.
I've also kept the /usr/local/include/google/ subdirectory as
synonym of /usr/local/include/gflags/, though the former name has
been obsolete for some time now.
The google-gflags Google Code page has been renamed to
gflags, in preparation for the project being renamed to
gflags. In the coming weeks, I'll be stepping down as
maintainer for the gflags project, and as part of that Google is
relinquishing ownership of the project; it will now be entirely
community run. The name change reflects that shift.
I've just released gflags 1.7. This is a minor release; the major
change is that CommandLineFlagInfo now exports the address in memory
where the flag is located. There has also been a bugfix involving
very long --help strings, and some other minor changes.
I've just released gflags 1.6. The major new feature in this release is support for setting version info, so that --version does something useful.
One minor change has required bumping the library number:
ReparseCommandlineFlags now returns void instead of int (the int
return value was always meaningless). Though I doubt anyone ever used
this (meaningless) return value, technically it's a change to the ABI
that requires a version bump. A bit sad.
There's also a procedural change with this release: I've changed the
internal tools used to integrate Google-supplied patches for gflags
into the opensource release. These new tools should result in more
frequent updates with better change descriptions. They will also
result in future ChangeLog entries being much more verbose (for better
or for worse).
See the ChangeLog for a full list of changes for this release.
I've just released gflags 1.5. This release has only minor changes
from 1.4, including some slightly better reporting in --help, and
an new memory-cleanup function that can help when running gflags-using
libraries under valgrind. The major change is to fix up the macros
(DEFINE_bool and the like) to work more reliably inside namespaces.
If you have not had a problem with these macros, and don't need any of the other changes described, there is no need to upgrade. See the ChangeLog for a full list of changes for this release.
I've just released gflags 1.4. This release has only minor changes from 1.3, including some documentation tweaks and some work to make the library smaller. If 1.3 is working well for you, there's no particular reason to upgrade.
I've just released gflags 1.3. gflags now compiles under MSVC, and all tests pass. I really never thought non-unix-y Windows folks would want gflags, but at least some of them do.
The major news, though, is that I've separated out the python package into its own library, python-gflags. If you're interested in the Python version of gflags, that's the place to get it now.
I've just released gflags 1.2. The major change from gflags 1.1 is it now compiles under MinGW (as well as cygwin), and all tests pass. I never thought Windows folks would want unix-style command-line flags, since they're so different from the Windows style, but I guess I was wrong!
The other changes are minor, such as support for --htmlxml in the python version of gflags.
I've just released gflags 1.1. It has only minor changes fdrom gflags 1.0 (see the ChangeLog for details). The major change is that I moved to a new system for creating .deb and .rpm files. This allows me to create x86_64 deb and rpm files.
In the process of moving to this new system, I noticed an inconsistency: the tar.gz and .rpm files created libraries named libgflags.so, but the deb file created libgoogle-gflags.so. I have fixed the deb file to create libraries like the others. I'm no expert in debian packaging, but I believe this has caused the package name to change as well. Please let me know (at [mailto:google-gflags@googlegroups.com google-gflags@googlegroups.com]) if this causes problems for you -- especially if you know of a fix! I would be happy to change the deb packages to add symlinks from the old library name to the new (libgoogle-gflags.so -> libgflags.so), but that is beyond my knowledge of how to make .debs.
If you've tried to install a .rpm or .deb and it doesn't work for you, let me know. I'm excited to finally have 64-bit package files, but there may still be some wrinkles in the new system to iron out.
gflags 1.0rc2 was out for a few weeks without any issues, so gflags
1.0 is now released. This is much like gflags 0.9. The major change
is that the .h files have been moved from /usr/include/google to
/usr/include/gflags. While I have backwards-compatibility
forwarding headeds in place, please rewrite existing code to say
#include <gflags/gflags.h>
instead of
#include <google/gflags.h>
I've kept the default namespace to google. You can still change with
with the appropriate flag to the configure script (./configure --help to see the flags). If you have feedback as to whether the
default namespace should change to gflags, which would be a
non-backwards-compatible change, send mail to
google-gflags@googlegroups.com!
Version 1.0 also has some neat new features, like support for bash commandline-completion of help flags. See the ChangeLog for more details.
If I don't hear any bad news for a few weeks, I'll release 1.0-final.