Summary: 447 instances, 305 unique

Text	Count
// TODO: Remove this constructor. This isn't faster than the first one.	2
// FIXME: Need to move to more structured error handling.	2
# FIXME: Move this format to just JSON for ease of loading.	1
// FIXME: This is an inefficent map, the string is duplicated.	10
// FIXME: Is it correct to do this here, or earlier?	1
// TODO: Add the other special variables, ie. depsFile, rspFile, rspFileContent,	1
// FIXME: This is currently commingled with the build file loading, even though	6
// TODO: Add a handler for SIGUSER1-type signals so the user can	1
// FIXME: Implement our formatter to here or Support/Format.h!	1
// TODO: Report error.	1
// FIXME:This should be made into class design,to avoid dupplication.	2
// FIXME: This is frequently a redundant lookup, given that the caller might	1
// FIXME: Figure out if we want a more lazy approach where we make the	2
// FIXME: we need to define `libllbuild_EXPORTS` to ensure that the	1
# FIXME: Should we pass the context pointer separately from the delegate	1
//TODO: Add SFINAE to test that the Container's iterators match the range's	2
// TODO: Report error.	6
// FIXME: We should eliminate this, it isn't well formed when loading	1
// FIXME: This is very performance critical and should be highly	1
// FIXME: We currently have to shoehorn the timestamp into a fake file	1
// FIXME: Think about this representation more. The problem with storing this	2
// FIXME: Use an rvalue reference for the outputs and inputs here to avoid	2
# FIXME: Get utime and stime available here, and copy.	1
// FIXME: This needs to be a per-execution seed. This is just a placeholder	2
// FIXME: One downside of doing the lookup here is that the client cannot	1
// FIXME: handle environment	2
// TODO: Handle newlines properly. We need to remove leading whitespace.	1
// FIXME: This isn't correct, we need utilities for doing shell quoting.	2
// FIXME: This is unfortunate, we should be able to destroy these naturally.	2
// FIXME: This isn't correct, we need utilities for doing shell quoting.	2
// FIXME: Design clean data types for clients to return unowned/unowned	1
// FIXME: This is somewhat inefficient, as we will end up repeating any	1
// FIXME: this chunk of code is duplicated, but it avoids a fstat when	1
// FIXME: Conditionalize these flags since SwiftPM 5.3 and earlier will crash for platforms they don't know about.	1
// FIXME: This API is now gross, there shouldn't be one generic status changed	2
// FIXME: Use a more appropriate hashing infrastructure.	6
// FIXME: This is a bit dubious, and wouldn't work anymore if we moved the	1
// FIXME: these assignments (& the equivalent const T&/const Optional& ctors)	2
// FIXME: should we save this for Clang?	2
// FIXME: If the outputs are empty, it might mean that this instance is for a custom task, which does not	1
// FIXME: The terminology is really confusing here, we have ExternalCommand which is divorced from the actual internal command implementation of the same name.	1
// FIXME: this poorly matches the generated SubtargetFeatureKV table.  There	1
// FIXME parse the optional <template-args> here!	1
// FIXME: this is invalid for WindowsCE.	1
// FIXME: Don't do default initialization.	1
// FIXME:  cmd.exe uses different syntax for I/O redirection to null.	1
// FIXME: Keeping this in a side table is very inefficient, we always have to	1
// FIXME: We ignore errors here, for now, this will be resolved when we move	1
// FIXME: It might be substantially better to evaluate all of these in the	2
// FIXME: Mange recursive lookup? Ninja crashes on it.	1
// FIXME: Need to support configuration behaviors.	1
// TODO - there is scope for optimization here, but at the moment this code	1
// FIXME: Factor out into a parser helper class.	2
// TODO Move this into some kind of libtestSupport?	3
// FIXME: Investigate changing the database error handling model to	1
// FIXME: Eventually, we should evaluate whether it would be more efficient	1
// FIXME:	2
// FIXME: Unfortunately, to make sure this is as efficient as possible we	2
// FIXME: This could use constexpr + static_assert. This way we	2
// FIXME: The intent of having a callback like Rule structure and a decoupled	2
// FIXME: We should revist this, as we probably wouldn't want a crash in the	1
// FIXME: Should we provide the input key here? We have it available	1
// FIXME: Figure out how this is going to be organized.	2
// FIXME - implement these	1
// FIXME: As above, this structure has redundancy in it.	1
assert(0 && "FIXME: invalid target");	1
// FIXME: Make efficient.	1
// FIXME: We want to remove this protocol eventually and use the FileSystem	1
// FIXME: We should eventually eliminate the need for this.	1
// FIXME: We need to continue scanning the rest of the inputs to ensure we	1
// FIXME: Support custom attributes in client tools.	3
// FIXME: SubArch values were simplified to fit into the expectations	1
// FIXME: This test is rather cumbersome for all it is trying to check, maybe	1
// TODO: Tag resolution.	1
// FIXME: Need to include additional information on the status here, e.g., the	2
// FIXME: We should sort out eventually whether the engine itself should	1
// FIXME: We need a deallocation callback in order to ensure this is released.	1
// FIXME: We need to fix the directory list to not get contents for symbolic	2
/// TODO: Remove this layer of indirection.  It's not necessary.	2
Size = Capacity = 0; // FIXME: Setting Capacity to 0 is suspect.	2
// FIXME: Note that this structure is redundant here, as	1
/// We maintain a number of invariants on Twine objects (FIXME: Why):	2
// FIXME: This should be cached.	1
// FIXME: This is not going through the filesystem object. Indeed the fs	2
// FIXME: Raise proper error.	2
// FIXME: This is O(N). We don't expect N to be large in practice, but it	1
// FIXME: Clean this up, we should be able to come up with a better definition	1
// FIXME: We should make this explicit once we have actual support for must	1
// TODO: Properly handle tags. "[!!str ]" should resolve to !!str "", not	1
// FIXME: This algorithm currently does a redundant stat for each directory,	2
// FIXME: From here on we assume that both sets have the same small size.	1
// FIXME: Optimize.	1
// FIXME: Is it right to bring this up-to-date when one of the inputs	1
// TODO: support skipping from the middle of a parsed collection ;/	2
// FIXME: Support dynamic tool commands.	1
// FIXME: Clang uses it, but it's bogus, since neon defaults to vfpv3.	1
// FIXME: This is incredibly wasteful to copy everywhere. Rephrase things so	1
// FIXME: Need to add support for required parameters.	1
// FIXME: This needs to be reworked to go through a C API.	1
// FIXME: In our model, we still wait for everything to terminate, which	1
// FIXME: This is a total hack, and should be cleaned up.	4
// FIXME: Diagnose this sooner.	1
// FIXME: Implement a shared cache for this, to reduce the number of	1
// FIXME: This feels awkward, maybe this isn't the right way to manage	2
// FIXME: Make the type of the result buffer correct instead of	2
# FIXME: Quote arguments better.	1
/// FIXME: remove this, since there is no implicit conversion to NodeTy.	2
// FIXME: might be better to just fake the timestamp on the file	1
// FIXME: Why do I have to name the closure signature here?	1
// FIXME: We shouldn't need to expose this to use for mapping purposes, we should be able to use something more efficient.	1
char Buffer[len] = {'\0'};  // FIXME: working around MSan false positive.	1
// FIXME: This is an unfortunate amount of copying.	1
// FIXME: This really should move out of subprocess and up a layer or two. The	1
//TODO: Consider ADL/non-member begin/end calls.	2
// FIXME: This involves a lot of redundant allocation, currently. We could	2
// FIXME: Find the beginning of the line properly for multibyte characters.	1
// FIXME: Why do we need this, why can't we get a pointer to the protocol	1
// FIXME: Use atomic operations here so that there is no data race.	1
// FIXME: This assertion is intended to catch unintended use of multibyte	1
// TODO: Make this const, if it's safe...	2
// FIXME: open files cannot be deleted.	1
// FIXME: Consider trying to share this with the Ninja implementation.	1
// FIXME: Implement checks without weak.	3
/// FIXME: We use EM_CurrentCodePage to write response files for GNU tools in	2
// FIXME: We should flag the task to ensure this is only called once, and	1
// FIXME: Twine.str() is not very efficient.	4
// FIXME: Design a way to communicate the "lane" here, for use in "super	2
// FIXME: Error handling.	2
// FIXME: This is an inefficent map, given that the string is contained	6
// FIXME: This is a good candidate for a 'long running' testing mode. Increase	1
// FIXME: We also need to figure out if a richer concurrency model is needed for	2
// FIXME: It would be nice to optimize this more, and the common case is that	2
// FIXME: We should probably try and move this so that it can be done by	1
// FIXME: We should support update-if-newer for commands with deps.	1
// FIXME: Eventually, it might be nice to harden the format so that we know	1
// FIXME: It would be nice to run this on a thread with a small stack to	1
// FIXME: should probably dump filters here too	1
// FIXME: Error handling.	1
// FIXME: This is due to clang 3.3 (or older version) always checks for the	2
// FIXME: Figure out what the deal is with normalization.	2
// FIXME: Delegate is lost.	1
// FIXME: don't do this if they're efficiently moveable.	2
// FIXME: These should be moved to TargetTuple once it exists	4
PathExts.push_back(".exe"); // FIXME: This must be in %PATHEXT%.	1
// FIXME: We are currently paying the cost for carrying this around on every	2
// FIXME: This is naively mapping all keys with no caching at this point,	1
# FIXME: Need a way to find this.	1
/// FIXME: It would be nice to have a StringSet.	1
// FIXME: This is not very efficient, if there are a lot of diagnostics.	1
// FIXME: This is currently an O(n) operation that could be relatively	1
// FIXME: We should do this work in the background.	3
// FIXME: When we move to C++11, use %t length modifier. It's not in	1
// FIXME: This is inefficient, we will end up doing the stat twice, once	1
// FIXME: We may eventually want to allow the individual job to provide some	2
// FIXME: vector is used because that's what clang uses for subtarget feature	2
// TODO: Is 16-byte alignment enough?  We copy small object files with large	1
// FIXME: Handle multibyte characters in the source line.	2
// FIXME: Consider assigning over existing elements, rather than clearing &	2
// FIXME: This doesn't belong here, move once the status is fully delegated.	1
// FIXME:         ::= L <type> <real-part float> _ <imag-part float> E   # complex floating point literal (C 2000)	1
// FIXME: Do we need to support these?	1
// FIXME: this is invalid for WindowsCE	1
// FIXME: We should be able to use <unwind.h> for any target that has an	1
// FIXME: Once we have custom ValueType objects, we would like to	1
// FIXME: This method is hard to use correctly, we should modify build to	2
// FIXME: Need a decent subprocess interface.	1
// FIXME: Provide client control.	2
// FIXME: Move to a shared location.	1
// FIXME: Shared behavior with ExternalCommand.	1
// FIXME: We should output this, but it requires updating lots of tests.	2
// TODO: the long colored prefix output might mess with this logic:	1
// FIXME: Check that the client created the rule appropriately. We should	1
// FIXME: That isn't necessarily true, we should try to mmap stdin and	1
// FIXME: This needs to honor certain properties of the execution queue	1
<!-- FIXME: This view is not currently supported	1
// FIXME: We should strictly enforce the integrity of this validity routine	1
// FIXME: We should consider making this interface capable of feeding	1
// ^ FIXME: This follows boost::make_zip_iterator's assumption that all	2
// FIXME: Need to move to more structured error handling.	2
/* FIXME: should this just take a const APInt reference?  */	1
// FIXME: Make this more complete, or move to a better dependency format.	1
// FIXME: This should really be done using an additional implicit input, so	1
// FIXME: Find a way to harden this more, or see if we can just get CMake	1
TODO: Implement all operations in semPPCDoubleDouble, and delete these	1
// FIXME: Move this to a private config.h as it's not usable in public headers.	2
# FIXME: This is a manually Python translation of the C++	1
// FIXME: Need to use the filesystem interfaces.	1
// FIXME: This should perhaps just be lumped in with the start call? Or	1
// FIXME: Really want this job to go into a high priority fifo queue	1
// ^ TODO: Implement random access methods.	2
// FIXME: We should probably factor this out somewhere else, so we can enforce	2
// FIXME: Invalid target name, produce an error.	1
// FIXME: There is no need to store the parameters in the build decl anymore	1
// FIXME: BE8 vs. BE32?	2
// FIXME: We want to replace is_integral_or_enum and is_pointer here with	2
// FIXME: Handle multibyte characters.	1
// FIXME: This could be packed one entry tighter.	1
// FIXME: Note that this is an Apple-specific extension, and we will have to	1
// FIXME: It probably should be an error to assign to the same parameter	2
// FIXME: This interface will need to get more complicated, and provide the	4
// FIXME: We don't want to process all of these requests, this amounts to	1
// TODO: Drop this assertion and the transitive type traits anytime after	2
/// FIXME: Currently if RangeT && is a rvalue reference to a temporary, the	2
// FIXME: We could save some reallocation by having a templated SmallVector	1
// FIXME: Eliminate this, and make the build value array contain an array of	1
/// FIXME: It'd be nice for this to be a non-templated constructor taking a	2
// FIXME: To TableGen this we need to make some table generated files available	2
// TODO - print build value	1
// FIXME: I am ambivalent about this living in Basic, I want all non-functional	2
// FIXME: We don't currently expose the caret diagnostic information, or	1
// FIXME: Actually parse the c-ns-shorthand-tag rule.	1
// TODO: Make the format method return the number of bytes written, that	2
// FIXME: Design the logging and status output APIs.	2
// FIXME: (possibly) incomplete list of features that clang mangles that this	1
// FIXME: Investigate retrieving this result lazily. If the DB is	1
// TODO instead insert, update and remove files from the archive	1
// FIXME: We may eventually want to allow the individual job to provide	2
// FIXME: This is currently encoding the key by merely stripping the	1
// FIXME: Can we safely always call this for __GNUC__ everywhere?	1
/// TODO: If clang source code is ever allowed to use constexpr in its own	1
// FIXME: It should be generic to C++11.	1
// FIXME:  This assumes the wildcard is only in the file name and not in the	1
// FIXME: These probably don't belong here, clients generally can just manage	2
// FIXME: This is copied from the Core BuildEngine unittest, we should figure	1
// FIXME: this may not actually make any sense if we can efficiently move	2
// FIXME: should support fs_create_symlink, but for now explicitly defers to built-in symlink	1
// FIXME: Introduce a per-build context object to hold this.	1
// FIXME: This is inefficient, we will end up doing the stat twice, once when	1
// FIXME: This needs to delegate, since we want to have a notion of	1
// FIXME: Reduce this API to a more narrow subset of std::vector	2
// FIXME: Remove once llvm optimizes this to the faster version without hints.	1
// FIXME: Provide this for PE/COFF targets.	2
// FIXME: Perhaps should use this encoder for the key itself? Right now	2
// FIXME: Subtract necessary number from StackTrace entries to turn return addresses	1
// TODO: Check if results were fetched	1
// FIXME: This API most likely doesn't belong.	2
inputs[inputID].second = value; // FIXME: avoid copying value ?	1
# FIXME: This creates a cycle.	1
// FIXME: Optimize for single entry list.	2
// FIXME: Parameterize this size?	2
# FIXME: What is to prevent the rule from being deallocated after this	1
// FIXME: Error reporting.	1
// FIXME: Need to get the client name and schema version from	1
// FIXME: This is trying not to break up replacements, but then to re-sync	1
// FIXME: Find a cleaner strategy for merging the internal schema version	1
// FIXME: Make this the default for llvm-symbolizer.	1
// FIXME: We should report a statistic for how much output we read from the	2
#define _WIN32_IE    0x0800 // MinGW at it again. FIXME: verify if still needed.	1
/// TODO: Consider making this not a child of Node.	2
// FIXME: This build graph triggers degenerate build behavior. Due to the	1
//FIXME: include cpuid.h from clang or copy __get_cpuid_max here	1
// FIXME: Add iterators that abstract over the small and large form, and then	2
// TODO: bool parameters are not readable and a source of bugs.	2
// FIXME: How do we pass the client schema version here, if we haven't	1
// FIXME: We need to track this state, and generate an error if this	1
// FIXME: This should be a separate callback, not shared with getFileInfo (or get FileInfo should take a parameter).	1
// FIXME: CONSISTENCY: One consistency issue in this model currently is that	1
# FIXME: This leaks everything, but we are currently dropping a	1
// FIXME: Should create a CustomCommand class, to avoid all the boilerplate	1
// FIXME: We should probably also move this to be dynamic.	1
/// \todo FIXME: remove when \c constexpr becomes really \c constexpr	2
// FIXME: This is gross, come up with a general purpose solution.	1
// FIXME: We should most likely have some kind of registration process so we	1
// FIXME: This is currently useful as a mechanism for defining builds in which	2
// FIXME: This class is only here to support the transition to llvm::Error. It	1
// FIXME: Need llvm::sys::fs.	1
// FIXME: FileOffset in MapViewOfFile() should be aligned to not dwPageSize,	1
// FIXME: If there are multibyte or multi-column characters in the source, all	1
// FIXME: Separate the delegate from the context pointer.	1
// FIXME: We need to delegate to the node here.	1
// FIXME: Design a way to communicate the "lane" here, for use in "super	2
// FIXME: TableGen this.	5
// FIXME: Move all of these things to a real home.	1
// FIXME: This should also account for dependencies that were added dynamically.	1
// FIXME: Need to include additional information on the status here, e.g., the	2
// FIXME: We should give the client the opportunity to support a previous	1
// FIXME: Lift the file access into the delegate, like we do for Ninja.	1
// TODO: util::emitError doesn't use the console queue?	1
// FIXME: We would like to call this Process, but then it conflicts with Swift's builtin Process. Maybe there is another name?	1
// FIXME: We should support BuildValues with arbitrary payloads.	1
// FIXME: Using a formula based purely on the precision is conservative;	1
// TODO: don't do more precise calculations above than are required.	1
// FIXME: We do not support this natively, yet.	1
// FIXME: At some point, figure out the optimal representation for this	2
// FIXME: We shouldn't need to do this, but should switch llbuild to using	1
// FIXME: base 36	1
// FIXME: when the SupportedPlatforms availability directive is updated and	1
// FIXME: This map is duplicated in parseOperatorName and parseExpr.	1
// FIXME: Really want this job to go into a high priority fifo queue	1
// FIXME: This is historical, We should tidy up this class to reflect that	1
// FIXME: Add a utility for capturing a subprocess infocation.	1
// FIXME: We don't try and set this in response to the restat flag on	1
// FIXME: This is not efficient, we could store into a single allocation.	2
// FIXME: Include cycleDetected callback.	1
// FIXME: We could audit these dependencies at this point to verify that	1
// FIXME: We need to fix the input ID situation, this is not extensible. We	1
// FIXME: For now, we just merge all the inputs. This isn't	1
// FIXME: We should reorganize again getPageSize() on Win32.	1
// FIXME: This interface will need to get more complicated, and provide the	2
// FIXME: Design the logging and status output APIs.	1
// FIXME: This might be more efficient if it returns Result.	2
// FIXME: The need to do this makes it questionable that we use this	1
// FIXME: this probably shouldn't use 'round half up'.	1
# FIXME: We need to get back the dependents view, it was super useful.	1
// FIXME: This should be abstracted into a helper class.	1