Summary: 229 instances, 196 unique Text Count // TODO: Implement a fast path for actual arrays. 3 // Remove all existing values .. TODO:: Double check whether this is correct, i.e. double check 1 // TODO: this comes about when we break in a function before its environment 1 // TODO: refine this type annotation to "array" ?. 1 // TODO: Make this as default behavior 1 // TODO: support multiple JS sources from command line. 1 == UNIT)) { // TODO :: Note sure whether the compact notation makes sense for other 1 // TODO: handle other binary operations. 1 // TODO:: Method take String list as argument to avoid conversions between array and list? 1 // TODO:: There must be a better way to do this. 3 // TODO: find a way to track the bytecode modules explicitly. 1 // TODO (T31512967): Prune unused defs. 1 // TODO could be HermesValue 2 // TODO: figure out how to store BoundFunction. 1 // TODO: if we check for OOM here in the future, we must undo the slot 1 // TODO: Verify the instruction is valid, need switch/case on each 1 else return key; // // TODO:: Make it tighter 1 // TODO: setCaseLevel method is not available. Essentially, the "CASE" sensitivity won't 1 // TODO T25663446: This is potentially optimizable to iterate over the inline 1 // TODO: this probably indicates an error. There is no "undefined" in JSON. 1 // TODO :: Avoid fetching the available locales array in the preceeding code when best-fit 1 "property is not configurable"); // TODO: better message. 1 // TODO: make sampling rate configurable. 1 // TODO: support them. For now we have an empty entry for the global scope. 1 # TODO: remember to update or remove them once we update the parser. 1 // TODO: need to come up with a better condition. 1 // TODO - do we care about this? Arrays are pretty safe to concurrently mutate 1 /// TODO: this has to be implemented properly. 1 // TODO: we do not yet have a way of getting the file name separate from 1 // TODO: share BASIC_BLOCK_STAT_VERSION with VM. 1 // FIXME: escaping, etc? 1 // TODO: use "thisArg" when builtins get fixed to support it. 1 // TODO:: I can't find the right rules to mark the type of the an year part as "yearName". 2 // TODO: keep thread names that are still in use. 1 // TODO: This could be implemented via a directional copy which either 1 "property is not writable"); // TODO: better message. 1 * TODO: 1 // TODO (T84179835): Disable this once it is no longer useful for debugging. 1 // TODO T65916424: Move this out of the hermes::vm namespace. 1 // TODO: good candidate for memchr on length-1 needles 1 // TODO: Need to make this accept format strings 1 // TODO: Test Object equality once Runtime::interpretFunction() is written. 1 return LOCALE; // TODO: Ad-hoc default 2 "property is not configurable"); // TODO: better message. 3 // TODO: use sourceMapTranslator for CJS module. 1 // TODO: this is an extremely naive implementation, it can be optimized in multiple ways: 1 // TODO: switch to WritableMemoryBuffer after updating LLVM. Or use mmap() 1 // TODO T65916424: Consider how we can move this somewhere more modular. 1 // TODO: enhance this when supporting more frame types. 1 // TODO: a better message. 1 /// TODO: do something when there is an uncaught exception, e.g. print 1 // TODO: Support class declarations once we support class IRGen. 1 // TODO: more sophisticated path concat handling. 1 // TODO: get the real process name. 1 // Allow the top-level value to be set. TODO: guard against multiple 1 // FIXME: Avoid duplicating this logic. 1 // FIXME: T42569352 TDZ for parameters used in initializer expressions. 1 # TODO: Flow conditionally outputs "inexact" depending on whether the parent 1 // TODO: we can't make HermesInternal.concat a static builtin method now 1 // TODO: Format string according to locale. 1 // TODO: allow a vector of callbacks. 1 // FIXME: obviously this is just an ugly hardcoded hack. 1 /// TODO: it would be nice to have this return a unique_ptr with a custom 1 // TODO: Remove dependency on SamplingProfilerPosix from ChromeTraceSerializer. 1 /// TODO: Implement this for heaps other than GenGC 1 /// TODO: Record line:col for inner functions. 1 // TODO: Handle other classes of recoverable errors. 1 // FIXME: Pass in file name here. 1 // TODO: show overflow block stats as well. 1 // TODO (T25527350): Heap Moving 1 // TODO: better error message. 1 // TODO: Test Object toString once Runtime::interpretFunction() is written. 1 # TODO: is this intended in hermes or should we fix it? 1 // TODO: If there is a header but no preheader, we can just create one. This 1 // TODO: capture leaf frame IP. 1 /// FIXME: This currently only returns an empty SourceMap. 1 // TODO: Add more stringent criteria for turning off this flag, for instance, 1 // TODO: if range is ASCII and unicode is not set, we can stop the search 1 // TODO: Handle globs 1 // TODO: look into doing this lazily. 1 // TODO:: AvailableLocales list always seems to be sorted. Is it guaranteed so that we can 1 // TODO: Don't lose err? 1 # TODO: The reason Hermes fails could be different from the expected error, so 1 // TODO: disallow "UnicodeIDContinue". 1 // TODO: a little more aggressive constant folding. 1 // TODO: Handle name index 1 availableLocaleIds.add(locale.toLanguageTag()); // TODO:: Not available on platforms <= 20 6 /// TODO (T26751833) figure out if this is a performance problem, and do 1 // TODO: handle other kinds of instructions. 1 "property is not configurable"); // TODO: better message. 1 // TODO T65916424: Consider how we can move this somewhere more modular. 1 // TODO: Utilize sizeHint. 1 // TODO: use com.facebook.common.locale.Locales.getApplicationLocale() as the current locale, 1 // TODO nice to have: don't do this if we are in square brackets. 1 // TODO: MallocGC doesn't yet support credit/debit external memory, so 1 } // TODO: Remove isRequest branches. 1 // TODO: Improve error reporting here. 1 /// TODO T40778724: this is a limited implementation sufficient for 1 // TODO: bravely teach buildSourceAndCaretLine to use wcwidth(), lifting this 1 // TODO : Make sure that this is not shown through the resolvedOptions 1 // TODO (T46136220): Set the .name property for anonymous functions that 1 // TODO: pass a sourceURL derived from a '//# sourceURL' comment. 1 // TODO: this would probably be much more efficient with bitmasks. 1 // TODO: consider writing this in JS. 1 // TODO: right now the NewTarget can only be the RegExp constructor itself, 1 // TODO: handle BoundFunction. 1 // TODO: Should we do something special for NativeConstructor? 1 // TODO: this changes the property map, so if we want to support OOM 1 // TODO: Test Object toNumber once Runtime::interpretFunction() is written. 1 // TODO: better error message. 4 // TODO: clearly this has to be optimized somehow 1 /// TODO: this is a wrapper around \f directRegExpExec right now, and it is 1 // TODO: optimize this by implementing memcpy-like functionality in ArrayImpl. 1 // TODO: this can be optimized by hoisting the branches out of the 1 // TODO: fix this for all GCs. 1 // TODO (T25527350): Debug Dump 1 // TODO: this can be made much more efficient at the expense of moving some 2 // TODO: Add a fast path for actual arrays. 6 // TODO: better message. 1 // TODO: defined at compile time. 1 // TODO: There is a fair amount of logic duplication between this function 1 // TODO: optimize this now that we know the index of the property slot. 1 // TODO (T25686322): In non-Malloc GCs, handle sanitization doesn't fully move 1 // TODO: better message. 2 // TODO: Share the global AST node and scope between calls to `visit_module` to avoid 1 /// TODO: this is wasteful in terms of code size because it will be 1 // TODO: avoid boxing primitives unless we are calling an accessor. 1 elseif (MINGW) # FIXME: Also cygwin? 1 else return key; // TODO:: Make it tighter 1 // TODO: Make this something other than empty string. 1 // TODO: rationalize what should happen if toString_RJS() itself throws. 1 // TODO :: Build/Look-for a BiMap datastructure. 1 // TODO: Predefine each error message. 1 // TODO: we can't deal with changing the scope depth of functions yet. 1 // TODO: remove this catch all function to find out which visitors need to be 1 /// TODO: Cache this value to avoid recomputing every time. 1 /// TODO: This format currently does not preserve perfect information, for 1 // TODO mhorowitz: move this to jsi/testlib.cpp once we have impls for all VMs 1 } // TODO (ronag): This should be moved before if(stream.destroyed) 1 // TODO: should we allow recursive style 1 // TODO: Add a fast path for actual arrays. 1 # TODO: Verify the result is string type. 1 // TODO: assert pCur equals to pSegEnd. 1 // TODO: Account for generator methods in SpecialMethod here. 1 // TODO: check each individual basic block's range instead of function. 1 // TODO: good candidate for Boyer-Moore on large needles/haystacks 1 /// TODO: Add support for other sourcemap metadata fields such as `metadata` and `sourcesContent`. 1 // TODO: This isLazy check is because a lazy object is reported as 1 /// TODO: Delete this function once all callers are replaced with 1 // TODO: Deal with different subclasses (e.g. rope) 1 // TODO (T28293178) Remove this once exceptions are supported. 1 // TODO :: I can't find a way to set the decomposition mode. 1 // TODO: this logic of finding overflow headers is duplicated here and in 1 // TODO: enable this when Runtime::raiseTypeError() is implemented. 1 // TODO: get real cpu id. 1 // TODO: find a way to get host function name out of 1 // TODO: Getting the name in this way works in most cases, but not for things 1 // TODO: Is there a portable way of expressing this in the preprocessor? 1 // TODO: maybe add some runtime checks based on codegenned predicates? 1 // TODO: It can be possible to avoid setting this flag if the variable is shadowed 1 // TODO: This leaves an extra global call frame that doesn't make any 1 // TODO: right now the NewTarget can only be the RegExp constructor itself, 1 // TODO: Add records for the bytecode header and similar. 1 /// TODO: consolidate the stack frame struct with other function/extern 1 else // TODO :: Make it more tight. 2 // FIXME: use the builtin value of @@iterator. Symbol could have been 1 // FIXME: May be inaccurate for some hex values. 1 // TODO: Long-term this could instead be handled by creating an internal 1 // TODO :: Make it more robust. 1 // TODO: Dynamic threshold is not used in incremental mode because 1 // TODO: Get a list-separator String for the host environment's locale. 2 // TODO this branch should also be used for module namespace and 1 // FIXME: use proper language semantics. 1 // TODO: truncate traces? Support Error.stackTraceLimit? 1 // TODO: optimize the case for null check. 2 // TODO: This is currently a wrapper around JSRegExp::create, and always creates 1 // TODO: log "GC in process" meta event. 1 // TODO: add validation for the loop and reduce the size of loopCount_ to 1 // TODO: Parse initializer. 1 // Hack: if we are already debugging, do nothing. TODO: in the event that we 1 // TODO: The IPs stored in the C++ callstack via this holder will generally be 1 // TODO: the code used to perform a stricter check which missed some cases. 1 public static String COLLATION_CASEFIRST = "colcasefirst"; // TODO:: double check this 1 // TODO: Support default class declarations once we support class IRGen. 1 // FIXME: Avoid reimplementing this logic. 1 // FIXME: Some padding is not included. 1 // TODO: Implement a purely CompressedPointer version of this. That will let 1 "cannot add a new property"); // TODO: better message. 2 // TODO: support them. 1 // FIXME: implement this. 6 * TODO: Change babel to use the Hermes AST format so this is no longer needed. 1 // TODO (T28293178) Remove this once exceptions are supported. 1 }; // TODO (ronag): Improve soft detection to include core modules and 1 // TODO: optimize this with caching, etc. 1 /// TODO optimization: avoid constructing the submatch array. Instead simply 1 // TODO: Parse modifiers. 1