Previously, the following rules were used for binary operators:
* If one of the values is a long, the result is long.
* If one of the values is unsigned, the result is also unsigned.
* Otherwise the result is an int.
C89 specifies the "usual arithmetic conversions" as:
* The integral promotions are performed on both operands.
* Then the following rules are applied:
* If either operand has type unsigned long int, the other operand is
converted to unsigned long int.
* Otherwise, if one operand has type long int and the other has type
unsigned int, if a long int can represent all values of an unsigned int,
the operand of type unsigned int is converted to long int; if a long int
cannot represent all the values of an unsigned int, both operands are
converted to unsigned long int.
* Otherwise, if either operand has type long int, the other operand is
converted to long int.
* Otherwise, if either operand has type unsigned int, the other operand is
converted to unsigned int.
* Otherwise, both operands have type int.
https://port70.net/~nsz/c/c89/c89-draft.html#3.2.1.5
As one example, these rules give a different result for an operator
with one long operand and one unsigned int operand. Previously,
the result type was unsigned long. With C89 semantics, it is just long,
since long can represent all unsigned ints.
Integral promotions convert types shorter than int to int (or unsigned int).
Both char and unsigned char are promoted to int since int can represent
all unsigned chars.
https://port70.net/~nsz/c/c89/c89-draft.html#3.2.1.1
Rename promoteint to ArithmeticConvert, since this is more accurate.
Fixes#170
Renamed GetReplacementType() to GetStructReplacementType().
Clarified in comments that most *Struct* facilities work for unions as well.
Made it clear in some error messages with regards to structs/unions.
didn't look "deep enough" into nested arrays to determine the constness
correctly.
git-svn-id: svn://svn.cc65.org/cc65/trunk@5622 b7a2c559-68d2-44c3-8de9-860c34a00d81
be converted to pointer-to-function and an array to pointer-to-member resp.
git-svn-id: svn://svn.cc65.org/cc65/trunk@3977 b7a2c559-68d2-44c3-8de9-860c34a00d81
type. Started to add general handling of address size flags in types.
git-svn-id: svn://svn.cc65.org/cc65/trunk@3891 b7a2c559-68d2-44c3-8de9-860c34a00d81
using signed integers as data types. This led to wrong results for mod, div
and compares.
git-svn-id: svn://svn.cc65.org/cc65/trunk@3797 b7a2c559-68d2-44c3-8de9-860c34a00d81
elements to a string of structs, each representing one type element. This
should fix problems on unusual architectures, since it is no longer necessary
to embedd pointers and other data converted to numbers in the string of
unsigned shorts.
Increased the TypeCode length to unsigned long to make room for more type
bits.
Inline more functions in datatype.h.
git-svn-id: svn://svn.cc65.org/cc65/trunk@3709 b7a2c559-68d2-44c3-8de9-860c34a00d81
When assigning structures, copy them using the primary register when the
size is 1, 2, or 4.
When assigning structures, allow structures by value when the size is 1,
2, or 4. These structures are expected to be in the primary register. The
only case when this can happen is by return from function, so this change
makes div() work.
git-svn-id: svn://svn.cc65.org/cc65/trunk@1477 b7a2c559-68d2-44c3-8de9-860c34a00d81
Rewrite of CharMapPragma in pragma.c
Fixed bug in scanner.c: Invalid octal constants containing the numbers 8
and 9 were accepted by the compiler.
Moved SignExtendChar from scanner.c to datatype.c.
git-svn-id: svn://svn.cc65.org/cc65/trunk@1416 b7a2c559-68d2-44c3-8de9-860c34a00d81
internal compiler errors in several places.
An old cc65 extension (variables of type void) was broken and did no longer
work. Fixed that.
git-svn-id: svn://svn.cc65.org/cc65/trunk@1184 b7a2c559-68d2-44c3-8de9-860c34a00d81
unused stores into these registers.
Because of this, the old code using regsave does no longer work. Started
to rewrite it.
git-svn-id: svn://svn.cc65.org/cc65/trunk@966 b7a2c559-68d2-44c3-8de9-860c34a00d81
* __fixargs__ is now an actual function symbol and no longer handled in the
scanner.
* Additional symbol __argsize__ that is a constant in normal functions and
a const local variable for variadic functions. Using this symbol, the
va_arg macro gets a lot simpler and smaller.
* Added special code to handle the fixed parameters of a variadic function.
The code has some overhead, but the va_fix macro is no longer needed (and
the compiler generated code is better than va_fix anyway).
git-svn-id: svn://svn.cc65.org/cc65/trunk@652 b7a2c559-68d2-44c3-8de9-860c34a00d81
optimization for C functions. The last parameter is not pushed onto the
stack by the caller but in the function entry code, so there is no speed
but a size gain.
git-svn-id: svn://svn.cc65.org/cc65/trunk@645 b7a2c559-68d2-44c3-8de9-860c34a00d81