Rather than save the entire CookedSource object in the module scope,
all we need is the vector of characters that CharBlock has references
into. This avoids the problem that CookedSource object has a dangling
reference to an AllSources object (which shouldn't be needed by the
CharBlocks in the symbol table, but is there).
This also decouples Scope from parser classes other than CharBlock.
Original-commit: flang-compiler/f18@fdbf5c841e
Reviewed-on: https://github.com/flang-compiler/f18/pull/149
Tree-same-pre-rewrite: false
parser::Parsing does most of the work needed to prescan and parse a file.
The only problem is that for module files we need to take ownership of
the cooked character stream so that it lives as long as the module scope
does.
To achieve this, change Parsing::cooked_ to be a unique_ptr and add a
member function to allow the caller to take ownership of it.
Original-commit: flang-compiler/f18@9088891ab2
Reviewed-on: https://github.com/flang-compiler/f18/pull/149
Tree-same-pre-rewrite: false
The `-module` option specifies where to write module files. It is also
the first place search for modules files. The current options name and
behavior match PGI compilers.
To support this, reorganize `ModFileWriter` to be more like `ModFileReader`.
The gives the client a way to set the output directory and fetch the
error messages. A bunch of static member functions in `ModFileWriter`
were changed to non-member functions so they didn't need to be exposed
in `mod-files.h`.
Original-commit: flang-compiler/f18@2717846e81
Reviewed-on: https://github.com/flang-compiler/f18/pull/147
When a use-stmt is encountered for a module that isn't in the global
scope, search for and read the appropriate `.mod` file. To perform the
search, pass the search directories in to ResolveNames.
For modules that were read from `.mod` files, we have to keep the cooked
source from being deleted so that the names so that references to names
stay valid. So we store the cooked source in the Scope of the module as
a `unique_ptr`.
Add `Symbol::Flag::ModFile` to distinguish module symbols that were read
from a `.mod` file rather than from the current compilation. Use it to
prevent writing those back out.
Fix test_errors.sh to run the compiler in the temp subdirectory --
otherwise tests could be affected by `.mod` files left from previous
tests.
Original-commit: flang-compiler/f18@207065999c
Reviewed-on: https://github.com/flang-compiler/f18/pull/145
Omppar
Original-commit: flang-compiler/f18@d341464e7f
Reviewed-on: https://github.com/flang-compiler/f18/pull/137
Due to a conflicting rebase during the linearizing of flang-compiler/f18, this commit squashes a number of other commits:
flang-compiler/f18@cfbc8bd16b Add OpenMP Block directives and some declarative directive(TBD:declare reduction and declare target)
flang-compiler/f18@d198352f84 Removed extra code from resolveing conflict of the merge.
flang-compiler/f18@85911a0e80 fix some grammar and unparsing.
flang-compiler/f18@43bcfb7cd3 OpenMP declartive directive can be after function declaration line. OpenMP Loop construct will be treated as if statement instead of block because currently parser cannot construct the loop if loop end with Label [Continue] Use DefinedOperator and ProcedureDesignator instead of create a new Parser and postpone the validity of operator in semantic.
flang-compiler/f18@5d83b50ad3 Merge branch 'master' into omppar
flang-compiler/f18@8c666436a5 remove "HEAD" code and resolve conflict.
When a procedure is included by name in a generic (either with a
procedure statement in a generic interface or with a generic statement)
we can't immediately resolve it to a subprogram symbol. That is because
the subprogram may be defined later in the specification part. Instead,
collect the names (and whether they should be module procedures) and
resolve them to symbols at the end of the specification part. Also
detect duplicate names then.
This is needed for module file reading as these forward references can
occur there.
Write generics to module file using generic statements. This allows us to
include the access-spec.
Original-commit: flang-compiler/f18@d11d002084
Reviewed-on: https://github.com/flang-compiler/f18/pull/132
Write symbols for external subprogram interfaces as interface-stmts.
Those go in the decls part of the module file, as opposed to contained
subprograms which go in the contains part. See modfile06.f90.
Write symbols with GenericDetails to module files. The specific
procedures of a generic interface are always written as procedure-stmts.
If they also have specific interfaces those are written in a separate
interface-stmt. See modfile07.f90.
Fix a bug where `real, external :: f` was not written like
`real f; external f`. We have to notice the EXTERNAL attribute on the
type-declaration-stmt and convert the entity to a procedure entity.
See modfile08.f90.
Fix a bug where a use-associated symbol is referenced in a
procedure-designator. We were not resolving that correctly.
Change ModFileWriter::PutEntity to include the kind of Details when
it reports an internal error due to a kind it can't handle.
Make DetailsToString public to support that.
Change test_errors.sh to fail if the f18 command exits due to a signal.
We were missing bugs where the correct errors were written out but then
module file writing crashed (due to failure to handle generics mentioned
above). Non-zero exit status is okay because we are expecting
compilation errors.
Change test_modfile.sh to allow for the expected module file contents to
be indented so the tests are easier to read.
Original-commit: flang-compiler/f18@82a7931e51
Reviewed-on: https://github.com/flang-compiler/f18/pull/132
Tree-same-pre-rewrite: false
Allow for bash anywhere on path.
Allow for utilites to be in /bin as well as /usr/bin.
Don't count on sed understanding '\t'.
The options to mktemp aren't portable. Instead just create a per-test
temp directory in the working directory (i.e. <build-dir>/test/semantics).
Setting the KEEP environment variable causes them all to be kept to help
in investigating failures.
Fixesflang-compiler/f18#128.
Original-commit: flang-compiler/f18@3736ceeca8
Reviewed-on: https://github.com/flang-compiler/f18/pull/129
The source files contain the expected contents of generated .mod files.
`test_modfile.sh` compiles the source file and verifies that the correct
.mod files are generated.
Original-commit: flang-compiler/f18@a23f53c1a7
Reviewed-on: https://github.com/flang-compiler/f18/pull/126
Tree-same-pre-rewrite: false
Module file writing is implemented in mod-file.cc. They need to be
written after all semantic checking. Until then, for testing, write
them out whenever names are resolved.
There is a header comment in the .mod files but it is mostly a
placeholder until we can read them in and do something with it.
Rename `Symbol::details<D>` to `Symbol::get<D>`. This asserts that the
details of the symbol match D and returns that type. But we need a way
to access the details as a variant as well (not just one of its types).
`details()` is the best name for that, especially as we already have
`set_details()`. Renaming the old `details` to `get` also better matches
`has` which is used to check which variant is present.
Original-commit: flang-compiler/f18@8d14be1a16
Reviewed-on: https://github.com/flang-compiler/f18/pull/126
Tree-same-pre-rewrite: false
Recognize `IntentStmt` and use `HandleAttributeStmt()` to implement it
as is done with other attribute statements. Add `Attr::INTENT_INOUT` as
a separate attribute for `INTENT(INOUT)`.
Collect attributes from the prefix and suffix of `FunctionStmt` and
`SubroutineStmt` (including `BIND(C)`) and set them on the subprogram
symbol.
Create a test for this using `test_symbol.sh`. It compiles with
`-funparse-with-symbols` and compares the output with the symbols in
comments in the input.
Change `test_errors.sh` to be similar to `test_symbol.sh`: check usage
and allow `F18` environment variable to override the path to the
compiler.
Original-commit: flang-compiler/f18@384828a22f
Reviewed-on: https://github.com/flang-compiler/f18/pull/120
Tree-same-pre-rewrite: false
A ProcedureDesignator is a derived type constructor call if the name
resolves to a derived type. We need to recognize it as such and not
report an error. It may be in another derived type so we have to call
FindSymbol() first to find the derived type.
Most of the diffs are due to `symbol` changing from a reference to a
pointer.
In this example we were reporting an error on the first `t1()` and
resolving the second as an external function call.
```
module m
type :: t1
end type
type(t1) :: x = t1()
type t2
type(t1) :: y = t1()
end type
end module
```
Original-commit: flang-compiler/f18@717b22ca2b
Reviewed-on: https://github.com/flang-compiler/f18/pull/116
When `-fdebug-dump-symbols` is supplied with `-funparse`, include symbol
information in comments in the Fortran output. This will be used for
testing to verify that correct symbols are defined and references in
the right place.
In `UnparseWithSymbols()`, walk the parse tree and collect symbol
definitions and references, organized by statement. When a symbol is
defined across several statement it is associated with the first.
The definition of implicitly defined symbols is associated with the
first reference.
To write out the symbol information, a new optional argument is added to
`Unparse()`: it is a function that is called immediately before each
statement is unparsed. We pass in a function that prints out the symbol
information collected for that statement.
Add `Symbol::GetType()` to make it easier to write the symbol types
and add `Symbol::SetType()` for uniformity.
Original-commit: flang-compiler/f18@2e827de4ad
Reviewed-on: https://github.com/flang-compiler/f18/pull/112
Tree-same-pre-rewrite: false
This gets the code base back to compiling cleanly with clang after
pull request 109.
There were two overloadings of `Post(const parser::DeclarationTypeSpec::Type &)`.
The one in DeclarationVisitor needed to call the one in DeclTypeSpecVisitor.
This was fixed by introducing a new function, SetDerivedDeclTypeSpec, to do
the equivalent thing.
Original-commit: flang-compiler/f18@81e447bf4e
Reviewed-on: https://github.com/flang-compiler/f18/pull/110
This consists of:
- a new kind of symbols to represent them with DerivedTypeDetails
- creating symbols for derived types when they are declared
- creating a new kind of scope for the type to hold component symbols
- resolving entity declarations of objects of derived type
- resolving references to objects of derived type and to components
- handling derived types with same name as generic
Type parameters are not yet implemented.
Refactor DeclTypeSpec to be a value class wrapping an IntrinsicTypeSpec
or a DerivedTypeSpec (or neither in the TypeStar and ClassStar cases).
Store DerivedTypeSpec objects in a new structure the current scope
MakeDerivedTypeSpec so that DeclTypeSpec can just contain a pointer to
them, as it currently does for intrinsic types.
In GenericDetails, add derivedType field to handle case where generic
and derived type have the same name. The generic is in the scope and the
derived type is referenced from the generic, similar to the case where a
generic and specific have the same name. When one of these names is
mis-recognized, we sometimes have to fix up the 'occurrences' lists
of the symbols.
Assign implicit types as soon as an entity is encountered that requires
one. Otherwise implicit derived types won't work. When we see 'x%y' we
have to know the type of x in order to resolve y. Add an Implicit flag
to mark symbols that were implicitly typed
For symbols that introduce a new scope, include a pointer back to that
scope.
Add CurrNonTypeScope() for the times when we want the current scope but
ignoring derived type scopes. For example, that happens when looking for
types or parameters, or creating implicit symbols.
Original-commit: flang-compiler/f18@9bd16da020
Reviewed-on: https://github.com/flang-compiler/f18/pull/109
Make DerivedTypeSpec a simple wrapper around the name of the type.
Leave out type parameter values until we are ready to resolve them
as well.
Change DeclTypeSpec to be an old-fashioned union with an enum to
indicate what is in it. std::variant doesn't work well here because
we wanted the enum visible in the class' API and there is not a
1-to-1 mapping between enumerators and data stored.
Original-commit: flang-compiler/f18@03bdeef790
Reviewed-on: https://github.com/flang-compiler/f18/pull/109
Tree-same-pre-rewrite: false
With this change, all instances Symbol are stored in class Symbols.
Scope.symbols_, which used to own the symbol memory, now maps names to
Symbol* instead. This causes a bunch of reference-to-pointer changes
because of the change in type of key-value pairs. It also requires a
default constructor for Symbol, which means owner_ can't be a reference.
Symbols manages Symbol instances by allocating a block of them at a time
and returning the next one when needed. They are never freed.
The reason for the change is that there are a few cases where we need
to have a two symbols with the same name, so they can't both live in
the map in Scope. Those are:
1. When there is an erroneous redeclaration of a name we may delete the
first symbol and replace it with a new one. If we have saved a pointer
to the first one it is now dangling. This can be seen by running
`f18 -fdebug-dump-symbols -fparse-only test/semantics/resolve19.f90`
under valgrind. Subroutine s is declared twice: each results in a
scope that contains a pointer back to the symbol for the subroutine.
After the second symbol for s is created the first is gone so the
pointer in the scope is invalid.
2. A generic and one of its specifics can have the same name. We currently
handle that by moving the symbol for the specific into a unique_ptr
in the generic. So in that case the symbol is owned by another symbol
instead of by the scope. It is simpler if we only have to deal with
moving the raw pointer around.
3. A generic and a derived type can have the same name. This case isn't
handled yet, but it can be done like flang-compiler/f18#2 above. It's more complicated
because the derived type and the generic can be declared in either
order.
Original-commit: flang-compiler/f18@55a68cf023
Reviewed-on: https://github.com/flang-compiler/f18/pull/107
We were collecting symbols in a map of SourceName to Symbol*.
This is wrong because sometimes different occurrences of a name
map to different symbols (e.g. in different scopes).
SourceName::begin() is unique for each occurrence so use that
as the map key instead.
The problem can be reproduced by running:
`f18 -fdebug-resolve-names -fparse-only -fdebug-dump-parse-tree`
on the following source. The two symbols 'i' should have different
types and they were both coming out as INTEGER because they both
pointed to the first symbol for 'i'.
```
module m
integer :: i
contains
subroutine s
real :: i
end
end
```
Original-commit: flang-compiler/f18@a165c717ff
Reviewed-on: https://github.com/flang-compiler/f18/pull/107
Tree-same-pre-rewrite: false
The Fortran source files in test/semantics all contain expected
errors in comments. The script test_errors.sh compiles a file with
'f18 -fdebug-resolve-names -fparse-only' and compares the actual
errors produced against the expected ones.
The change to resolve15.f90 is necessary because test_errors.sh can't
handle two expected errors for the same source line.
A useful command to run these is 'ctest -R f90 --output-on-failure'.
-R means only run tests with f90 in the name
--output-on-failure prints the output of test_errors.sh when a test
fails, showing the expected and actual messages that differ.
Original-commit: flang-compiler/f18@df18ee7bc9
Reviewed-on: https://github.com/flang-compiler/f18/pull/105
peter klausler