The gl_in array has a special path due to context-specific
gl_MaxPatchVertices, making the code out of order for tagging built-ins.
This commit moves the tagging to the correct location.
This also fixes issue #80.
This fixes a bug where a token that could be a keyword in one version
is not a keyword in another version, but treated like a non-member after
a "." dereference.
4 components are needed when used a texture, but not an image, which multiplies
layers and faces into the same coordinate. This fixes it from using 4 everywhere,
to only using 4 for textures and 3 for images.
Structured control-flow rules allow leaving the middle of a construct through
a return, but not through a jump to a block that does a return.
Addresses issue #58.
This generally simplifies access chain generation, with far fewer type conversions.
It is particularly important to future SPIR-V changes where there is less aggregate
type uniqueness due to carrying different layout information with the type.
If this breaks your AST consumer, best is to modify it to test
against the enum values instead of doing string comparisons on
built-in function names. This is the reason the change was made.
If you need the old behavior, you should be able to get it back by changing
PureOperatorBuiltins to be false instead of true. This path will work for
a while, but is marked deprecated.
Also, the old behavior is tagged as release 2.4.
This is to avoid all need to do text comparison of built-in function names
when consuming the AST. All built-in functions get enumerants.
Will want to turn on soon. See PureOperatorBuiltins. See issue #8.
Previously if a non-void function implictly returned, a dummy variable
was created as return value. Now instead it returns the result of the
OpUndef instruction. This better conveys the presence of undefined
behavior to SPIR-V consuming tools (and humans).
It also saves one ID per occurrence...
There will be subsequent commits to refine semantics, esp. version-specific semantics,
as well as I/O functionality and restrictions.
Note: I'm getting white-space differences in the preprocessor test results,
which I'm not checking in. I think they need to be tagged as binary or something.
This patch introduces a new extension, GL_GOOGLE_include_directive,
to enable support #include directives. It depends on the extension
GL_GOOGLE_cpp_style_line_directive.
When an include directive is recognized by the preprocessor, it
executes a callback on the filepath argument to obtain the file
contents. That way the compilation client can deal with the file
system, include paths, etc.
Currently only accepts quoted filepaths -- no angle brackets yet.
According to the GLSL spec, the second parameter to #line should be
an integer source string number and __FILE__ will be substituted
with the integer source string number currently processed. This
patch extends the syntax of #line and __FILE__. Now #line accepts
as the second parameter a filename string quoted by double quotation
marks. And if such a #line is set, __FILE__ will be substituted with
the currently set filename string. The implementation is done via
introducing a new extension GL_GOOGLE_cpp_style_line_directive using
the extension framework.
The purpose is to support cpp-style #line directives, which is
required by #include.
Fixes issue #25. (char 255 aliased to -1 and missing tests for end of input).
1) All layers of input scanning now share a single EndOfInput value.
This avoids translation of it across layers of encapsulation.
2) Some places looking for end of line were not stopping on EndOfInput.
3) Use of "char" for the input made char values > 127 be negative numbers.
This allowed for aliasing of 255 to -1, etc. This is fixed by using
unsigned char.
