Driver uses its path to derive the plugin paths (i.e.
'lib/swift/host/plugins' et al.) Previously it was a constant string
'swiftc' that caused SourceKit failed to find dylib plugins in the
toolchain. Since 'SwiftLangSupport' knows the swift-frontend path,
use it, but replacing the filename with 'swiftc', to derive the plugin
paths.
rdar://107849796
d8d8bdc007 changed how supplementary
outputs were removed, but since tests weren't able to be run there was
inevitably a bug. The `return` should have been a `continue`.
`ArgList::eraseArgIf` was removed in apple/llvm-project
000a90ab63ce0595d442b3e7e8adc9c0506d2525. Re-implement by retrieving the
option specifier ID for all options that have the `SupplementaryOutput`
flag in the given arguments and then erasing those.
If the compiler arguments have errors in them (e.g. because a file with the same name is used twice), we can often still fulfill SourceKit requests because the compiler argument errors are only relevant for later stages of the compilation process.
Instead of bailing out early, do a best effor retrieving the compiler arguments that are valid and ignoring the errors.
Fixes rdar://77618144
If the driver is only set up to get the frontend invocation (e.g. from sourcekitd), don't validate that we can link against ARCLite because
a) we don't care about link-time when we are only interested in the frontend invocation
b) finding arclite might cause us to find clang using xcrun which in turn can cause sourcekitd to hang.
rdar://50659268
We've seen bad file descriptor errors in certain build environments (rdar://73157185) and retrying
opening those files seems to be a walkaround. In this change, we allow the
compiler to retry expanding reponse files in argument lists.
rdar://73892564
getSingleFrontendInvocationFromDriverArguments is set up to never produce file
lists in the output frontend arguments, but since the driver accepts file lists
as input arguments, this should too.
When producing frontend arguments for sourcekitd, force the output mode
to -typecheck so that we do not create any temporary output files in the
driver. Previously, any sourcekitd operation that created a compiler
invocation would create 0-sized .o file inside $TMPDIR that would never
be cleaned up.
The new swift-driver project handles temporaries much better as
VirtualPath, and should not need this approach.
rdar://62366123
The `-force-single-frontend-invocation` flag predates WMO and is now an
alias for `-whole-module-optimization`. We should use the latter and let
the former fade into history.
But not even that much uglier; at all three call sites this will save
an allocation, and for the most important one (SourceKit) we can now
avoid creating a temporary CompilerInvocation just to copy into a
longer-lived one.
With this change, Driver no longer depends on Frontend, which means...
well, slightly faster builds of the compiler itself, but not much
else.
Explicitly disable batch mode in createCompilerInvocation, since it uses
-force-single-frontend-invocation. Previously we were getting spurious
warnings. Also add a test that -disable-batch-mode will allow commands
that use -index-file to avoid the same warning, since that is likely
what they want to do as well.
rdar://39581506
When generating a compiler invocation in driver::createCompilerInvocation()
we end up using filelists if the number of inputs is > 128 (to work around
command line arg limits). We never actually write them out though, and so
fail when parsing the frontend arguments that reference them.
As this function is called frequently by SourceKit and command line limits
aren't a concern here, this patch makes the 128 threshold value configurable
via a new -driver-filelist-threshold option. This is set to its maximum value
in driver::createCompilerInvocation() to ensure filelists aren't used. This
new option makes the existing -driver-use-filelists (that forces filelists to
be used) redundant as it's now equivalent to -driver-filelist-threshold=0.
Resolves rdar://problem/38231888
The "Tool" abstraction wasn't buying us enough to deserve the added
complexity. Now a ToolChain turns Actions into Jobs, and every helper
tool is searched for relative to Swift first. Much simpler.
Swift SVN r31563
...and rename Command to Job (previously the name of the base class).
We never generated job lists directly contained in other job lists, so
let's not even worry about this case. We may some day need to break Job
out into separate subclasses (Clang has Command and FallbackCommand in
addition to JobList), but we should be able to keep the list separate.
No intended functionality change.
Swift SVN r23144
Added a new API, swift::driver::createCompilerInvocation. This takes an array
of driver arguments, constructs a Driver and a Compilation, and then uses the
Compilation's frontend arguments to create a CompilerInvocation.
This works by forcing Driver to create a Compilation which contains a single
compile command. (It achieves this by passing
"-force-single-frontend-invocation" after all other arguments.) This approach
roughly matches Clang's clang::createInvocationFromCommandLine.
As implied by the namespacing, this lives in swiftDriver. As a result,
swiftDriver now depends on swiftFrontend.
In support of this, added a couple of driver diagnostics for exceptional error
cases (where Driver produced something other than a single Command, or if that
Command is not a frontend command).
This fixes <rdar://problem/16125395>.
Swift SVN r20972
