ninja-1.7.2-7.1

- Add macros.ninja as an external source file and install it to
  the %{_rpmconfigdir}/macros.d/; this installs macros for RPM
  that are in turn used by the RPM macros for meson.

- Update to version 1.7.2:
  * %e and %r work again in dumb terminals (#1146, #1148)
  * `-t browse` now HTML-escapes rule names (#1157)
  * an edge may now have 0 explicit outputs if it has implicit
    outputs (#1159)
  * make misc/measure.py python 3-compatible
  * misc/write_fake_manifest.py now optionally writes source files,
    use it to quickly produce large, buildable projects for testing
    (#1109, #1198)
  * on newer FreeBSDs, use ppoll()
- Refresh ninja-disable-maxprocs-test.patch

- Update to version 1.7.1
  * Change default NINJA_STATUS to [%f/%t] github issue #1143

- update to version 1.7.0
  * ninja now supports implicit outputs.
  * Ninja's output has slightly changed, to make it easier to
    automated tools to look at ninja's output.
  * Ninja now explicitly handles SIGHUP and cleans up after itself.
  * Ninja now uses posix_spawn instead of fork/exec to launch
    subprocesses (on POSIX).
- adjust links to new wesite

- Make it build on SLE11 (use python2 there).

- Update dependencies
- Use optflags when building
- Make building more verbose

- Update to version 1.6.0
  * rules are now scoped to the subninja they're in.
  * dupe edges with multiple outputs now uses first edge
    instead of last (#867)
  * pools actually work now
  * generators can now run more often than twice if needed
  * unexpected output names now cause a rebuild instead of
    a hard error
  * POSIX-specific changes
  + child processes get detached from terminal
  + check for SIGINT after ppoll/pselect -> faster Ctrl-C
  + add an explicit SIGTERM signal handler
  * Ran ninja under afl-fuzz and fixed all crashes and bugs it found
  + fix crashes on cyclic graphs with multiple outputs
  + failing stat() now aborts build
  + cyclic rule bindings no longer crash
  + dependency cycles with multiple outputs no longer get
    ninja into a stuck state
- Update ninja-disable-maxprocs-test.patch to match upstream
  changes
- use Python 3 for configure

- Update to version 1.5.3
  * Reaping finished subprocesses now has precedence over
    starting new jobs
  * ninja no longer depends on googletest
- Install bash/zsh completion
- Install vim syntax file
- Implement %check
- Add ninja-disable-maxprocs-test.patch to disable max procs test,
  it needs a ulimit raise.

- Only depend on python-base to trim down BuildRequires
- Do not build the html manual
- Use upstream pristine tarball

- Update to version 1.5.1
  * Recompacting removes more entries from the build and deps logs.
  * Add a manifest parsing perftest and generator that generates
    random build manifests.
  * $in, $out are escaped more robustly.
  * `depfile=$out.d` now works if $out gets escaped.
  * ninja doesn't error out when it tries to create a directory and
    the directory already exists.
  * Ninja now accepts both / and \ on windows when making directories.
  * \r\n line endings are now accepted in depfiles.
  * depfile file names can now contain '{' and '}'.
  * -t compdb no longer crashes on edges with emtpy inputs.
  * 'keeprsp' is now included in the spelling suggestestions for -d
  * Many improvements to ninja-mode.el.
  * ninja_syntax.py is now faster.
  * bash completion and zsh completion provides completions more often.
  * ninja.vim now highlights $$ correctly.

- Update to version 1.4.0
  * Bug fixes
- Fix version field. Its 1.x not 3.x

- Update to git hash 0f53fd3
  * Version 3.0 now
  * Don't poll pipes using POLLRDHUP
  * Recompact the deps log when it gets too big

- Update to git hash 7bf53c6
  * Build ninja itself in deps mode
  * Bugfixes
  * Documentation updates

- Update to git hash ecc876e
  * Add compdb command to generate compile_commands.json for clang.
  * Add elapsed time (%e) to NINJA_STATUS.

- Update to git hash b26d217
  * Documentation updates
  * Fix wrong usage of std::string::find

- First release