integration
Motivation: libIDE clients should be simple, and they should not have to
translate token-based SourceRanges to character locations.
This also allows us to remove the dependency of DiagnosticConsumer on the
Lexer. Now the DiagnosticEngine translates the diagnostics to CharSourceRanges
and passes character-based ranges to the DiagnosticConsumer.
Swift SVN r7173
Every valid source location corresponds to a source buffer. There should be no
cases where we create a source location for a random string. Thus,
findBufferContainingLoc() always succeeds.
Swift SVN r7120
Lexer::BufferID is the single point of truth, Lexer::BufferStart and
Lexer::BufferEnd is just a cache -- they always point to the beginning and end
of the buffer, even in a sublexer.
Swift SVN r7079
Now we have a clear separation between a primary lexer, which scans the whole
buffer, and a sublexer, which can be created from a primary lexer to scan a
part of the buffer.
Swift SVN r7077
Decouple splitting an interpolated string to segments, from encoding the string segments.
This allows us to tokenize or re-lex a string literal without having to allocate memory for
encoding the string segments when we don't need them encoded.
Swift SVN r6940
-Parse the expression using a begin/end state sub-Lexer instead of a general StringRef Lexer
-Introduce a Lexer constructor accepting a BufferID and a range inside the buffer, and use it for swift::tokenize.
Swift SVN r6938
...instead of returning tok::unknown, which may confuse the Parser and cause it to crash with llvm::unreachable.
By ignoring them we reduce the complexity in the Parser by not having it worry about tok::unknown.
Swift SVN r6910
Also centralizes the knowledge about whether the hashbang is allowed in the
SourceManager. This fixes a bug in tokenize() because previously it just had
to guess.
Swift SVN r6822
around everywhere
Fixes:
rdar://14585108 Code completion does not work at the beginning of the file
rdar://14592634 Code completion returns zero results at EOF in a function
without a closing brace
Swift SVN r6820
Modules can be in either 'Raw' or 'Canonical' form, with different invariants on each. We don't actually distinguish those invariants yet, but this patch adds the field to SILModule and adds a "sil_stage" declaration to SIL printer/parser syntax.
Swift SVN r6793
This makes long literals like 1_000_000_000 or 0x7FFF_FFFF_FFFF_FFFF much easier to read, and has precedent in Perl, OCaml, Python, Ruby, Rust, ... Fixes <rdar://problem/14247571>.
Swift SVN r6681
completion token
This is required to handle cases like fooObject.#^A^#.bar where code completion
is invoked inside the ".." token. Previously, the token would not be split and
the lexer would produce an incorrect tokenization for this case. Now we
produce ".", tok::code_complete, ".".
Swift SVN r6635
-Introduce PersistentParserState to represent state persistent among multiple parsing passes.
The advantage is that PersistentParserState is independent of a particular Parser or Lexer object.
-Use PersistentParserState to keep information about delayed function body parsing and eliminate parser-specific
state from the AST (ParserTokenRange).
-Introduce DelayedParsingCallbacks to abstract out of the parser the logic about which functions should be delayed
or skipped.
Many thanks to Dmitri for his valuable feedback!
Swift SVN r6580
* Added a mode in swift-ide-test to test code completion. Unlike c-index-test,
the code completion token in tests is a real token -- we don't need to
count lines and columns anymore.
* Added support in lexer to produce a code completion token.
* Added a parser interface to code completion. It is passed down from the
libFrontend to the parser, but its functions are not called yet.
* Added a sketch of the interface of code completion consumer and code
completion results.
Note: all this is not doing anything useful yet.
Swift SVN r6128
This is to make overload resolution for the protected constructor call sites more explicit.
Otherwise, adding a parameter in the protected constructor and a default parameter in the public one
can inadvertently result in the protected constructor call sites switching to picking the public one.
Swift SVN r6007
When we are interpreting escape sequences in the lexer, we copy the string
literal bytes to ASTContext instead of storing a pointer to the source buffer.
But then we used to try to get a source location for that string in the heap,
which is not a valid source buffer. It succeeds during parsing, but breaks
when we try to print a diagnostic using this location.
Added a verifier check for this.
Also added a real source range for StringLiteralExpr, instead of a source range
with begin == end, produced from the beginning location.
Swift SVN r5961
Reimplement 'switch' parsing for our new AST representation, where cases contain patterns and 'where' guards, case blocks can have multiple cases, and 'default' is constrained to being the lone label of the last block if present. No type-checking or parsing of actual pattern productions yet.
Swift SVN r5834
This makes parsing patterns a bit more straightforward and prevents us from inadvertently accepting '_' as an identifier in a place we shouldn't.
Swift SVN r5806
parser state. Backtracking will be used a lot when we implement delayed
parsing for function bodies, and we don't want to leak lexer and parser state
details to AST classes when we store the state for the first and last token for
the function body.
Swift SVN r5759
The analyzer can't reason about the following complicated
control-dependency:
if (!isxdigit(BytesPtr[0]) || !isxdigit(BytesPtr[1]) ||
!isxdigit(BytesPtr[2]) || !isxdigit(BytesPtr[3])
continue; // Ignore invalid escapes.
StringRef(BytesPtr, 4).getAsInteger(16, CharValue);
BytesPtr += 4;
In this case, we cap the number of bytes read at 4, and we have
verified that the bytes will always form an integer. The analyzer
thinks there could be a false positive uninitialized value warning
here since 'getAsInteger()' does its own checking to see if the
parsed bytes actually compose an integer value.
Swift SVN r5566
