blob: 976f91cb74eb98fb3118043ba5aac701af512054 [file] [log] [blame]
$Log: CHANGES,v $
Revision 1.12 2002/08/29 17:57:08 greg
Donny made the following changes:
More comment on FDSEnumConstByName
A new function : FDSEnumConstByValue (changes in fds.c and fds.h)
Now, printing an enum will print the string
New test files: testaif.c and testaif.h
Revision 1.11 2002/07/22 22:32:03 greg
Greg made the following changes:
Fixed some minor bugs.
Added applypatch and fixed createpatch scripts.
Updated patches to support GDB 5.2
Updated documentation.
Updated licence to GNU.
Revision 1.10 2002/07/04 17:28:33 greg
Greg/Donny made the following changes:
Fixed support for unsigned integers.
Revision 1.9 2002/07/02 20:42:47 greg
Greg made the following changes:
Added support for MacOS X.
Updated copyright notices in source files.
Revision 1.8 2002/05/18 06:11:04 greg
Donny made the following changes:
1. New test for AIFDiff() in testaif.c
2. A modification for the "non-null with null" comparison in calc.c
and cmp.c
3. Some additions in aif-valprint.c, a partial support for
TYPE_CODE_FUNC, and also a change in TYPE_CODE_VOID.
4. A change in AIF_FUNCTION and also a new function in fds.c
that support the changes in aif-valprint.c
CVe: ----------------------------------------------------------------------
Revision 1.8 2002/05/18 06:11:04 greg
Donny made the following changes:
1. New test for AIFDiff() in testaif.c
2. A modification for the "non-null with null" comparison in calc.c
and cmp.c
3. Some additions in aif-valprint.c, a partial support for
TYPE_CODE_FUNC, and also a change in TYPE_CODE_VOID.
4. A change in AIF_FUNCTION and also a new function in fds.c
that support the changes in aif-valprint.c
CVe: ----------------------------------------------------------------------
Thu May 16 15:26:36 EST 2002 greg
Donny made some changes to struct/class code in conv.c.
Mon Apr 29 11:16:26 EST 2002 greg
Donny added new functionality to support circular lists.
I fixed a bug in _aif_array_ref().
Mon Apr 15 09:40:18 EST 2002 greg
Donny made the following changes:
fixed bugs in _aif_string_to_str - wrong arguments
fixed a bug in _aif_enum_to_str - must be advanced until the end
fixed bug in _aif_enum_to_str - wrong way of doing it
fixed a bug in AIF_STRUCT in _aif_unary_op -
_aif_unary_op changes the fds
add AIF_ENUM in _fds_advance(), _fds_count_bytes(),
_aif_binary_op(), _aif_unary_op()
new tests in testaif.c
Wed Apr 10 15:37:08 EST 2002 greg
Donny made the following changes:
fixed memset bug in BoolToAIF().
merged AIF*ToStr() and AIFPrint*() routines.
Thu Mar 21 12:23:47 EST 2002 greg
Donny made the following changes:
updated INTERNAL
changed FDS_ID, so now:
{|aa=is4;;;} for anonymous struct
{id|aa=is4;;;} for non-anonymous struct
changed AIFPrint() so that it doesn't print the FDS_ID
clean up the _aif_diff() & _aif_binary_op()
note: there is still 'case AIF_BOOLEAN' in
_aif_binary_op() for AND & OR operations
depth argument in AIFPrint(), AIFCompare() & AIFDiff()
( depth == 0 ---> all )
changed testaif.c to reflect the changes
Mon Mar 4 13:49:45 EST 2002 greg
Donny fixed a number of bugs in fds.c:
call to START_STRUCT() in _data_len_public() incorrect
FDSAddFieldToClass() needs to check for named struct, also
pointers not handled properly
wrong argument used in AIFAddFieldToStruct()
strncmp() in FDSStructFieldByName()
FDSTypeCompare() returns wrong value
wrong argument to _field_attribute() in FDSStructFieldByName()
Donny fixed arguments to testAscii() in testaif.c
Wed Feb 27 18:48:42 EST 2002 greg
Added Donny's changes to support new data types:
boolean
enumerated types
unions
classes
Fixed a bug in TypeToFDS() - missing va_start().
Removed log messages as they were getting too messy.
Tue Aug 22 14:11:19 EST 2000 rfinkel
cleaned up code
added file internal.doc to explain the data structures
added file internal.h for prototypes of internal routines
removed unused variables
added "const" to many prototypes to enforce readonly status
fixed bugs
_aif_int_to_str invoked _aif_div_int with a bad parameters.
_aif_compare of two arrays called _aif_compare with bad params.
collapsed _field_type and _field_name into _field_attribute; changed
from static to dynamic allocation of type and name return
fields because types can be embedded within types.
added routines to the API for pointers
AIF *PointerToAIF(const AIF *);
AIF *AIFNull(); /* return null value */
void AIFPrintPointer(FILE *, AIF *);
enhanced routines in cmp.c to handle pointers
added routines to the API for structs
AIF *EmptyStructToAIF(const char *);
int AIFAddFieldToStruct(AIF *, char*, AIF *);
void AIFPrintStruct(FILE *, AIF *);
enhanced routines in fds.c for nested structs
enhanced routines in cmp.c to handle structs
added routines to the API for names (for circular structures)
AIF *NameAIF(const AIF *, const char *);
AIF *ReferenceAIF(const char *);
AIF *CanonicalAIF(const AIF *);
enhanced routines in cmp.c to handle names
Fri Feb 9 09:51:44 EST 2001 rfinkel
Reorganized the meaning of the fds and its associated data to conform with
the output of gdb with the new "set print aif on" option.
Added code for a string type.
added routines to API
AIF *FromAscii(const char *, const char *);
modified the comparison, arithmetic, and dump/print routines to advance
file and data pointers so that data length is no longer passed in. In
this way we can handle data of indeterminate length, such as linked
lists.
precision
internal representation of integers is now long long int to increase
range.
_aif_to_double() can now handle f16 format; changed all callers to
understand that, although with possible loss of precision.
algorithm
_aif_add_int and the like now convert to native types, do arithmetic,
then convert back to AIF; they used to simulate arithmetic by
bit manipulation on the AIF_DATA. If necessary, we can simulate
longer precision by splitting into short-precision chunks.
similarly, type conversions now go through native types.
cleaned up code
added "const" to more prototypes to enforce readonly status.
regularized the order of parameters and their formal names.
added some formal names to prototypes as external documention.
After I have modified/checked a routine, I put a comment
at the closing brace naming the routine.