The run_cmake_with_options() command already sets the test
command to ${CMAKE_COMMAND}. The options passed to that
command shouldn't also add another ${CMAKE_COMMAND}.
This removes a warning seen in the output of tests that invoke
__ep_test_with_build_with_server().
In ac6a4d4884 (ExternalProject: Improve robustness of update step,
2020-10-17), the method used to check whether we already have a
commit or not was changed from using git rev-list to git rev-parse.
The new logic assumed rev-parse would output nothing if given a commit
hash it didn't know about, but it simply prints the hash again without
raising an error in this scenario. Amend that logic by adding ^{commit} to
the ref to ensure we do get an error if that ref is not currently known.
Fixes: #22166
The previous implementation was scanning the documentation in
the file at runtime to determine the set of supported keywords for
each public function. This was fragile, made it difficult to restructure
the documentation and was sometimes observable in runtime
performance measurements. Change to a more conventional
approach where supported keywords are explicitly listed in the
code.
The internal _ExternalProject_SELF variable is no longer needed.
CMake now provides CMAKE_CURRENT_FUNCTION_LIST_DIR which
can be used for the same purpose and avoids having to set a
variable when the module is read. This also removes the
requirement that the module must be included by the current or a
parent scope. It is now enough that the module has been included
once somewhere before calling any of its functions.
The above changes combined mean that the module can now use
include_guard() and avoid having to re-parse the very long file every
time.
Refactor the update logic to make it easier to follow. The following
fixes/improvements are some consequences of this change:
* Absorb a confusing git checkout failure message when the failure
is allowed and we act on that failure appropriately.
* Fix an unnecessary fetch in some scenarios when checking out a
git hash we already have locally.
* Stash and restore any local changes even when not rebasing.
* Avoid unsafe rebasing where we are not on a branch that is
already tracking the requested branch.
* When fetching, use --tags --force to ensure we get all the tags
and commits leading up to them regardless of whether the tags
are on branches or not. Also update our local tags if they move
on the remote.
Fixes: #20677
The ExternalProject module cannot be implemented in the Xcode "new build
system" without using CMP0114's NEW behavior. When configuring for that
build system, warn if the policy is not set to NEW and use NEW behavior
anyway.
`ExternalProject_Add_StepTargets` and `INDEPENDENT_STEP_TARGETS` have
some limitations and lack some sanity checks. They can cause confusing
build systems to be generated. The basic problems are:
* The notion of step independence is attached to the step target
rather than the step itself.
* The custom commands implementing the steps are duplicated in the
step targets and the primary targets. This can cause races.
It is also incompatible with the Xcode "new build system".
Fix this by introducing policy CMP0114 to change the way step target
dependencies are handled. Define independence from external
dependencies as a property of each individual step regardless of whether
there is a target for it. Add dependencies among the primary target and
the step targets such that each custom command only appears in one
target. When some steps are disconnected from the primary target, add
step targets for the steps commonly depended upon so that there is a
place to hold their custom commands uniquely.
Fixes: #18663
In order to shorten the download failure of ExternalProject download
steps, a download retry is only done when a recoverable network
error is encountered.
This option only has an effect if at least one of the other LOG_<step>
options is enabled. If an error occurs for a step which has logging to
file enabled, that step's output will be printed to the console. For
cases where a large amount of output is recorded, just the end of that
output may be printed to the console.
Since in some situations, ExternalProject module may be included in
a sub-directory, functions will be available in the global scope but
local variables like "_ep_keywords_<keyword>" will not be defined, this
commit checks and reports an error indicating that the ExternalProject
module must be explicitly included before using any of the ExternalProject_*
functions that require the module's inclusion within the current scope
or above.
Co-authored-by: Pablo Hernandez <pablo.hernandez@kitware.com>
Co-authored-by: Craig Scott <craig.scott@crascit.com>
The known keywords for each function are obtained by scraping the
documentation for lines matching a particular regular expression. In
commit 8842a027 (ExternalProject: Improve documentation, 2017-07-09),
the docs were overhauled and the COMMAND docs subsequently matched the
regular expression when they shouldn't have. This made COMMAND appear as
a true keyword, which thwarted the special handling logic elsewhere for
the intended use of COMMAND arguments.
This commit contains a workaround for issue #17229 to force a dependency
of the patch step on the update step to ensure a predictable step order.
Fixes: #17198
Added new USES_TERMINAL option to the ExternalProject_Add_Step
function. This option passes USES_TERMINAL to the underlying
add_custom_command call so that the Ninja console pool is used.
Also, corresponding new USES_TERMINAL_<step> options were added
to the ExternalProject_Add function.
Justification: if using Ninja with a CMake superbuild, it's often
desirable to limit the superbuild to ONE sub-Ninja process at a
time to avoid oversubscribing the CPU. Using the console pool also
makes it easy to monitor the progress of the sub-Ninja process.
Independent USES_TERMINAL_<step> arguments are passed to
ExternalProject_Add instead of one USES_TERMINAL argument that
controls everything. Users may wish to run some steps in parallel
but not others (e.g. parallelize configure but not build).
