SourceProcessing

log

age author description
7 months ago Jean Utke FreeSans no longer includeddefault tip
7 months ago utke on macbook year
10 months ago utke on macbook test case
10 months ago utke on macbook the errror message as it was didn't make any sense - we are interested only in the cases where we create more 4-byte floats
10 months ago utke on macbook issue the warning only for the first case and add more info
10 months ago utke on macbook branch merge
10 months ago utke on macbook shouldn't have 0 length character array
10 months ago utke on macbook 0 lengh string (should have a temporary with length 1 as zero won't do)
10 months ago utke on macbook for this test case insist on hoisting the strings
10 months ago utke on macbook fix up test script and update the references for test files
11 months ago Jean Utke standalone (outside of OpenADFortTk) + updated
19 months ago utke on macbook regenerated
19 months ago utke on macbook comprehensive change of the transformfile implementation to enable common block type changes in all of mitgcm
19 months ago utke on macbook indentation
20 months ago utke on macbook adapt regression test to the changed interface and update on reference output
20 months ago utke on macbook fix various problems with the generated initialization procedures including: collect declarations for all common block variables, not just the active ones. associate the declarations with the common block declaration to avoid problems with non-unique common block variable names; create the declaration tfrom the type table information to prevent problems originating with declarations applying to mutlple variables (some of which could not be common block ones) and variable declarations spread oiver mutliple statements.;
20 months ago utke on macbook I don't want to read all this
20 months ago utke on macbook pydev complains about unused variables
20 months ago utke on macbook fix up timing and progress messages
20 months ago utke on macbook fix up progress messages
20 months ago utke on macbook CharacterEntryKind should not be global by default - and make consistent the naming for character length id
20 months ago utke on macbook use inspect for the debugging messages
20 months ago utke on macbook after being done with the unit we clean up and then verify- the verification part should be made optional to save time later.
20 months ago utke on macbook add the type table to the debugging info
20 months ago utke on macbook stmt2unit - mark things on module exit. If the "END" statement is generic we need to go through the generic_exit to detect if the end applies to a module - NOTE this may turn out also to be necessary for other generic END cases; fix up debug statements - also NOTE using "self" as a parameter name for a stand-along method is misleading;
20 months ago utke on macbook dump all the removed types in the debug log
20 months ago utke on macbook debug info
23 months ago hlcm functionality added for differentiating between global and local types, removing local types at the end of the unit, and verifying all types at the end
2012-09-20 utke on macbook put this to stdout
2012-09-20 utke on macbook add some info on the profiling bit
2012-09-20 utke on macbook add timing info for individual files
2012-09-12 hlcm add subroutine comments
2012-08-27 hlcm Make sure that function names passed in subroutine calls are updated to subroutine names, if appropriate
2012-08-27 hlcm rollback change of SubCallStmt modifications in canonicalizer
2012-08-24 hlcm fix EndInterfaceStmt parsing to save the name of the interface & change interface processing in function2subroutine to update the name in the EndInterfaceStmt
2012-08-17 utke on macbook force hoisting if the identifier is declared a PARAMETER which implies it has a constant init.
2012-08-16 hlcm move printing of internal comments to flow() call in fortStmts NonComment so that format type is correctly updated
2012-08-16 hlcm remove outdated comments
2012-08-16 hlcm add functionaliry for type-bound procedures and implicit initializations of user defined types
2012-08-10 utke on macbook one more step to allow expressions (that evaluate to constants at compile time)
2012-08-10 utke on macbook this change permits declarations such as character*(6) :: ca which previously wasn't covered
2012-08-10 utke on macbook the single quotes always mess up the fontification in emacs
2012-08-10 utke on macbook repr is more informative
2012-08-02 hlcm branch merge
2012-08-02 hlcm fixed parsing for module use statements to handle 'module-nature' attribute & updated fortStmts unit test
2012-07-11 utke on macbook branch merge
2012-07-11 utke on macbook if we see an untyped external symbol it may be function or a subroutine. Even though we "know" what external functions there are that were subroutinized the symbol table look up does not (and should not) see the names of these other compile units and so we allow for checking when the symbol is declared external.
2012-07-11 utke on macbook always make the entryKind value an instance of the entry kind class; don't trigger the name clash message when the symbol in question is declared external with out type which defaults to subroutine but once we get a type we can accpet that type and know it is an external function with a proper return type; use the _getframe to get the proper function name in error and debug messages
2012-07-11 utke on macbook typo
2012-07-11 utke on macbook message formatting
2012-07-11 utke on macbook cover the case of untyped external symbol that is apparently used as a function call
2012-07-11 utke on macbook better error message
2012-07-11 utke on macbook provide a tester for the special case
2012-07-11 utke on macbook we made this more precise
2012-07-11 utke on macbook if we see the external statement first we assume an external subroutine - later it may be changed to a function if we see a type declaration
2012-07-11 utke on macbook add external as special case for external subroutines (i.e. "void" return type)
2012-07-11 hlcm remove obsolete type attribute in symbol table; type table is used now
2012-07-10 hlcm branch merge
2012-07-10 hlcm if named type doesn't have a symtab entry, return False in ntArrayReference (an appropriate exception will then be thrown by calling method) [else case]
2012-07-10 utke on macbook mimic a void return type for external subroutine names