qf/quakeforge
0
0
Fork 0
mirror of https://git.code.sf.net/p/quake/quakeforge synced 2024-11-29 23:52:22 +00:00
Code Issues Projects Releases Packages Wiki Activity
quakeforge/tools/qfcc/source/expr_binary.c

1255 lines
32 KiB
C
Raw Normal View History

Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
/*
exprtype.c
Expression type manipulation
Copyright (C) 2013 Bill Currie <bill@taniwha.org>
Author: Bill Currie <bill@taniwha.org>
Date: 2013/06/27
This program is free software; you can redistribute it and/or
modify it under the terms of the GNU General Public License
as published by the Free Software Foundation; either version 2
of the License, or (at your option) any later version.
This program is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.
See the GNU General Public License for more details.
You should have received a copy of the GNU General Public License
along with this program; if not, write to:
Free Software Foundation, Inc.
59 Temple Place - Suite 330
Boston, MA 02111-1307, USA
*/
#ifdef HAVE_CONFIG_H
# include "config.h"
#endif
[qfcc] Start work on implementing geometric algebra This gets only some very basics working: * Algebra (multi-vector) types: eg @algebra(float(3,0,1)). * Algebra scopes (using either the above or @algebra(TYPE_NAME) where the above was used in a typedef. * Basis blades (eg, e12) done via procedural symbols that evaluate to suitable constants based on the basis group for the blade. * Addition and subtraction of multi-vectors (only partially tested). * Assignment of sub-algebra multi-vectors to full-algebra multi-vectors (missing elements zeroed). There's still much work to be done, but I thought it time to get something into git.
2023-08-21 08:37:56 +00:00
#include "tools/qfcc/include/algebra.h"
[build] Move to non-recursive make There's still some cleanup to do, but everything seems to be working nicely: `make -j` works, `make distcheck` passes. There is probably plenty of bitrot in the package directories (RPM, debian), though. The vc project files have been removed since those versions are way out of date and quakeforge is pretty much dependent on gcc now anyway. Most of the old Makefile.am files are now Makemodule.am. This should allow for new Makefile.am files that allow local building (to be added on an as-needed bases). The current remaining Makefile.am files are for standalone sub-projects.a The installable bins are currently built in the top-level build directory. This may change if the clutter gets to be too much. While this does make a noticeable difference in build times, the main reason for the switch was to take care of the growing dependency issues: now it's possible to build tools for code generation (eg, using qfcc and ruamoko programs for code-gen).
2020-06-21 14:15:17 +00:00
#include "tools/qfcc/include/diagnostic.h"
#include "tools/qfcc/include/expr.h"
#include "tools/qfcc/include/options.h"
#include "tools/qfcc/include/type.h"
#include "tools/qfcc/source/qc-parse.h"
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
typedef struct {
int op;
[qfcc] Make a field name more clear Just "type" isn't very informative, but "result_type" is at least a fair bit better.
2022-01-27 14:29:32 +00:00
type_t *result_type;
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
type_t *a_cast;
type_t *b_cast;
Make binary_expr fold constants This is where constant folding should have happened all along. While unary_expr should fold constants too, it seems to already try to do so and it's a bit much of a mess to clean up right now.
2019-06-09 15:36:13 +00:00
expr_t *(*process)(int op, expr_t *e1, expr_t *e2);
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
} expr_type_t;
Make binary_expr fold constants This is where constant folding should have happened all along. While unary_expr should fold constants too, it seems to already try to do so and it's a bit much of a mess to clean up right now.
2019-06-09 15:36:13 +00:00
static expr_t *pointer_arithmetic (int op, expr_t *e1, expr_t *e2);
[qfcc] Correctly implement pointer arithmetic
2020-03-17 03:45:55 +00:00
static expr_t *pointer_compare (int op, expr_t *e1, expr_t *e2);
[qfcc] Cast functions to ints for comparison Yet more (non-)wasted instructions.
2022-01-29 09:59:38 +00:00
static expr_t *func_compare (int op, expr_t *e1, expr_t *e2);
Support vector/quaternion division by float Implemented via v*(1/f) or q*(1/f) to give CSE a chance to optimize the division if necessary as otherwise the engine would have to divide every time.
2019-06-09 23:44:36 +00:00
static expr_t *inverse_multiply (int op, expr_t *e1, expr_t *e2);
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_t *double_compare (int op, expr_t *e1, expr_t *e2);
[qfcc] Add support for horizontal vector ops And reimplement vector comparison for Ruamoko.
2022-01-30 01:56:15 +00:00
static expr_t *vector_compare (int op, expr_t *e1, expr_t *e2);
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
static expr_t *vector_dot (int op, expr_t *e1, expr_t *e2);
[qfcc] Implement dot and cross product for Ruamoko With explicit operators, even. While they're a tad verbose, they're at least unambiguous and most importantly have the right precedence (or at least adjustable precedence if I got it wrong, but vector ops having high precedence than scalar or component seems reasonable to me).
2022-01-30 05:14:15 +00:00
static expr_t *vector_multiply (int op, expr_t *e1, expr_t *e2);
[qfcc] Implement vector scaling for Ruamoko With this, qfcc-tests builds (can't run yet due to unsigned not having any tests and thus the rest the Ruamoko code in QF not building yet).
2022-01-30 05:48:49 +00:00
static expr_t *vector_scale (int op, expr_t *e1, expr_t *e2);
[qfcc] Alias entity to int for comparison The ruamoko ISA has no entity comparison operators because an entity is just an int in disguise.
2022-02-06 11:12:22 +00:00
static expr_t *entity_compare (int op, expr_t *e1, expr_t *e2);
Make binary_expr fold constants This is where constant folding should have happened all along. While unary_expr should fold constants too, it seems to already try to do so and it's a bit much of a mess to clean up right now.
2019-06-09 15:36:13 +00:00
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
static expr_type_t string_string[] = {
{'+', &type_string},
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
{EQ, &type_int},
{NE, &type_int},
{LE, &type_int},
{GE, &type_int},
{LT, &type_int},
{GT, &type_int},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
static expr_type_t float_float[] = {
{'+', &type_float},
{'-', &type_float},
{'*', &type_float},
{'/', &type_float},
{'&', &type_float},
{'|', &type_float},
{'^', &type_float},
{'%', &type_float},
Implement %% (true modulo) support in qfcc However, it's not quite working yet
2020-02-16 02:57:58 +00:00
{MOD, &type_float},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{SHL, &type_float},
{SHR, &type_float},
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
{AND, &type_int},
{OR, &type_int},
{EQ, &type_int},
{NE, &type_int},
{LE, &type_int},
{GE, &type_int},
{LT, &type_int},
{GT, &type_int},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
static expr_type_t float_vector[] = {
[qfcc] Implement vector scaling for Ruamoko With this, qfcc-tests builds (can't run yet due to unsigned not having any tests and thus the rest the Ruamoko code in QF not building yet).
2022-01-30 05:48:49 +00:00
{'*', .process = vector_scale },
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
static expr_type_t float_quat[] = {
{'*', &type_quaternion},
{0, 0}
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
static expr_type_t float_int[] = {
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{'+', &type_float, 0, &type_float},
{'-', &type_float, 0, &type_float},
{'*', &type_float, 0, &type_float},
{'/', &type_float, 0, &type_float},
{'&', &type_float, 0, &type_float},
{'|', &type_float, 0, &type_float},
{'^', &type_float, 0, &type_float},
{'%', &type_float, 0, &type_float},
Implement %% (true modulo) support in qfcc However, it's not quite working yet
2020-02-16 02:57:58 +00:00
{MOD, &type_float, 0, &type_float},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{SHL, &type_float, 0, &type_float},
{SHR, &type_float, 0, &type_float},
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
{EQ, &type_int, 0, &type_float},
{NE, &type_int, 0, &type_float},
{LE, &type_int, 0, &type_float},
{GE, &type_int, 0, &type_float},
{LT, &type_int, 0, &type_float},
{GT, &type_int, 0, &type_float},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
#define float_uint float_int
#define float_short float_int
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_type_t float_double[] = {
{'+', &type_double, &type_double, 0},
{'-', &type_double, &type_double, 0},
{'*', &type_double, &type_double, 0},
{'/', &type_double, &type_double, 0},
{'%', &type_double, &type_double, 0},
Implement %% (true modulo) support in qfcc However, it's not quite working yet
2020-02-16 02:57:58 +00:00
{MOD, &type_double, &type_double, 0},
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
{EQ, .process = double_compare},
{NE, .process = double_compare},
{LE, .process = double_compare},
{GE, .process = double_compare},
{LT, .process = double_compare},
{GT, .process = double_compare},
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
{0, 0}
};
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
static expr_type_t vector_float[] = {
[qfcc] Implement vector scaling for Ruamoko With this, qfcc-tests builds (can't run yet due to unsigned not having any tests and thus the rest the Ruamoko code in QF not building yet).
2022-01-30 05:48:49 +00:00
{'*', .process = vector_scale },
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
{'/', .process = inverse_multiply},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
static expr_type_t vector_vector[] = {
{'+', &type_vector},
{'-', &type_vector},
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
{DOT, .process = vector_dot},
[qfcc] Implement dot and cross product for Ruamoko With explicit operators, even. While they're a tad verbose, they're at least unambiguous and most importantly have the right precedence (or at least adjustable precedence if I got it wrong, but vector ops having high precedence than scalar or component seems reasonable to me).
2022-01-30 05:14:15 +00:00
{CROSS, &type_vector},
[qfcc] Add an explicit hadamard operator While the option to make '*' mean dot product for vectors is important, it breaks vector scaling in ruamoko progs as the resultant vector op becomes a dot product instead of the indented hadamard product (ie, component-wise).
2022-11-15 15:06:21 +00:00
{HADAMARD, &type_vector},
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
{'*', .process = vector_multiply},
{EQ, .process = vector_compare},
{NE, .process = vector_compare},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
#define vector_int vector_float
#define vector_uint vector_float
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
#define vector_short vector_float
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_type_t vector_double[] = {
[qfcc] Auto-demote double for vector scaling This was missed in the switch to an explicit scale instruction.
2022-02-06 11:10:47 +00:00
{'*', &type_vector, 0, &type_float, vector_scale},
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
{'/', .process = inverse_multiply},
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
{0, 0}
};
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
static expr_type_t entity_entity[] = {
[qfcc] Alias entity to int for comparison The ruamoko ISA has no entity comparison operators because an entity is just an int in disguise.
2022-02-06 11:12:22 +00:00
{EQ, &type_int, 0, 0, entity_compare},
{NE, &type_int, 0, 0, entity_compare},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
static expr_type_t field_field[] = {
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
{EQ, &type_int},
{NE, &type_int},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
static expr_type_t func_func[] = {
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
{EQ, .process = func_compare},
{NE, .process = func_compare},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
static expr_type_t pointer_pointer[] = {
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
{'-', .process = pointer_arithmetic},
{EQ, .process = pointer_compare},
{NE, .process = pointer_compare},
{LE, .process = pointer_compare},
{GE, .process = pointer_compare},
{LT, .process = pointer_compare},
{GT, .process = pointer_compare},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
static expr_type_t pointer_int[] = {
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
{'+', .process = pointer_arithmetic},
{'-', .process = pointer_arithmetic},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
#define pointer_uint pointer_int
#define pointer_short pointer_int
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
static expr_type_t quat_float[] = {
{'*', &type_quaternion},
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
{'/', .process = inverse_multiply},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
static expr_type_t quat_vector[] = {
{'*', &type_vector},
{0, 0}
};
static expr_type_t quat_quat[] = {
{'+', &type_quaternion},
{'-', &type_quaternion},
{'*', &type_quaternion},
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
{EQ, &type_int},
{NE, &type_int},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
static expr_type_t quat_int[] = {
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{'*', &type_quaternion, 0, &type_float},
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
{'/', .process = inverse_multiply},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
#define quat_uint quat_int
#define quat_short quat_int
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_type_t quat_double[] = {
{'*', &type_quaternion},
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
{'/', .process = inverse_multiply},
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
{0, 0}
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
static expr_type_t int_float[] = {
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{'+', &type_float, &type_float, 0},
{'-', &type_float, &type_float, 0},
{'*', &type_float, &type_float, 0},
{'/', &type_float, &type_float, 0},
{'&', &type_float, &type_float, 0},
{'|', &type_float, &type_float, 0},
{'^', &type_float, &type_float, 0},
{'%', &type_float, &type_float, 0},
Implement %% (true modulo) support in qfcc However, it's not quite working yet
2020-02-16 02:57:58 +00:00
{MOD, &type_float, &type_float, 0},
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
{SHL, &type_int, 0, &type_int}, //FIXME?
{SHR, &type_int, 0, &type_int}, //FIXME?
{EQ, &type_int, &type_float, 0},
{NE, &type_int, &type_float, 0},
{LE, &type_int, &type_float, 0},
{GE, &type_int, &type_float, 0},
{LT, &type_int, &type_float, 0},
{GT, &type_int, &type_float, 0},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
static expr_type_t int_vector[] = {
[qfcc] Handle vector scaling by ints I missed a change when implementing support for the scale instructions.
2022-02-04 12:57:41 +00:00
{'*', &type_vector, &type_float, 0, vector_scale},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
static expr_type_t int_pointer[] = {
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
{'+', .process = pointer_arithmetic},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
static expr_type_t int_quat[] = {
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{'*', &type_quaternion, &type_float, 0},
{0, 0}
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
static expr_type_t int_int[] = {
{'+', &type_int},
{'-', &type_int},
{'*', &type_int},
{'/', &type_int},
{'&', &type_int},
{'|', &type_int},
{'^', &type_int},
{'%', &type_int},
{MOD, &type_int},
{SHL, &type_int},
{SHR, &type_int},
{AND, &type_int},
{OR, &type_int},
{EQ, &type_int},
{NE, &type_int},
{LE, &type_int},
{GE, &type_int},
{LT, &type_int},
{GT, &type_int},
{0, 0}
};
static expr_type_t int_uint[] = {
[qfcc] Add support for unsigned, long, etc long is ignored for double, and v6p progs are stuck with 32 bits for longs (don't feel like extending v6p any further), but the basics are there for Ruamoko. short is ignored for ints because the minimum size is 32, and signed is just noise for ints anyway (and no chars, so...). unsigned, however, is finally implemented properly (or at least seems to be working correctly: tests pass after getting things compiling again, and lt.u is used where it should be :)
2022-01-19 09:08:58 +00:00
{'+', &type_int, 0, &type_int},
{'-', &type_int, 0, &type_int},
{'*', &type_int, 0, &type_int},
{'/', &type_int, 0, &type_int},
{'&', &type_int, 0, &type_int},
{'|', &type_int, 0, &type_int},
{'^', &type_int, 0, &type_int},
{'%', &type_int, 0, &type_int},
{MOD, &type_int, 0, &type_int},
{SHL, &type_int, 0, &type_int},
{SHR, &type_int, 0, &type_int},
{EQ, &type_int, 0, &type_int},
{NE, &type_int, 0, &type_int},
{LE, &type_int, 0, &type_int},
{GE, &type_int, 0, &type_int},
{LT, &type_int, 0, &type_int},
{GT, &type_int, 0, &type_int},
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
{0, 0}
};
static expr_type_t int_short[] = {
{'+', &type_int, 0, &type_int},
{'-', &type_int, 0, &type_int},
{'*', &type_int, 0, &type_int},
{'/', &type_int, 0, &type_int},
{'&', &type_int, 0, &type_int},
{'|', &type_int, 0, &type_int},
{'^', &type_int, 0, &type_int},
{'%', &type_int, 0, &type_int},
{MOD, &type_int, 0, &type_int},
{SHL, &type_int, 0, &type_int},
{SHR, &type_int, 0, &type_int},
{EQ, &type_int, 0, &type_int},
{NE, &type_int, 0, &type_int},
{LE, &type_int, 0, &type_int},
{GE, &type_int, 0, &type_int},
{LT, &type_int, 0, &type_int},
{GT, &type_int, 0, &type_int},
{0, 0}
};
static expr_type_t int_double[] = {
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
{'+', &type_double, &type_double, 0},
{'-', &type_double, &type_double, 0},
{'*', &type_double, &type_double, 0},
{'/', &type_double, &type_double, 0},
{'%', &type_double, &type_double, 0},
Implement %% (true modulo) support in qfcc However, it's not quite working yet
2020-02-16 02:57:58 +00:00
{MOD, &type_double, &type_double, 0},
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
{EQ, &type_int, &type_double, 0},
{NE, &type_int, &type_double, 0},
{LE, &type_int, &type_double, 0},
{GE, &type_int, &type_double, 0},
{LT, &type_int, &type_double, 0},
{GT, &type_int, &type_double, 0},
{0, 0}
};
#define uint_float int_float
#define uint_vector int_vector
#define uint_pointer int_pointer
#define uint_quat int_quat
static expr_type_t uint_int[] = {
[qfcc] Add support for unsigned, long, etc long is ignored for double, and v6p progs are stuck with 32 bits for longs (don't feel like extending v6p any further), but the basics are there for Ruamoko. short is ignored for ints because the minimum size is 32, and signed is just noise for ints anyway (and no chars, so...). unsigned, however, is finally implemented properly (or at least seems to be working correctly: tests pass after getting things compiling again, and lt.u is used where it should be :)
2022-01-19 09:08:58 +00:00
{'+', &type_int, &type_int, &type_int },
{'-', &type_int, &type_int, &type_int },
{'*', &type_int, &type_int, &type_int },
{'/', &type_int, &type_int, &type_int },
{'&', &type_int, &type_int, &type_int },
{'|', &type_int, &type_int, &type_int },
{'^', &type_int, &type_int, &type_int },
{'%', &type_int, &type_int, &type_int },
{MOD, &type_int, &type_int, &type_int },
{SHL, &type_uint, &type_int, &type_int },
{SHR, &type_uint, 0, &type_int },
{EQ, &type_int, &type_int, &type_int },
{NE, &type_int, &type_int, &type_int },
{LE, &type_int, &type_int, &type_int },
{GE, &type_int, &type_int, &type_int },
{LT, &type_int, &type_int, &type_int },
{GT, &type_int, &type_int, &type_int },
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
{0, 0}
};
static expr_type_t uint_uint[] = {
{'+', &type_uint},
{'-', &type_uint},
{'*', &type_uint},
{'/', &type_uint},
{'&', &type_uint},
{'|', &type_uint},
{'^', &type_uint},
{'%', &type_uint},
{MOD, &type_uint},
{SHL, &type_uint},
{SHR, &type_uint},
[qfcc] Add support for unsigned, long, etc long is ignored for double, and v6p progs are stuck with 32 bits for longs (don't feel like extending v6p any further), but the basics are there for Ruamoko. short is ignored for ints because the minimum size is 32, and signed is just noise for ints anyway (and no chars, so...). unsigned, however, is finally implemented properly (or at least seems to be working correctly: tests pass after getting things compiling again, and lt.u is used where it should be :)
2022-01-19 09:08:58 +00:00
{EQ, &type_int, &type_int, &type_int},
{NE, &type_int, &type_int, &type_int},
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
{LE, &type_int},
{GE, &type_int},
{LT, &type_int},
{GT, &type_int},
{0, 0}
};
#define uint_short uint_int
#define uint_double int_double
#define short_float int_float
#define short_vector int_vector
#define short_pointer int_pointer
#define short_quat int_quat
static expr_type_t short_int[] = {
{'+', &type_int, &type_int, 0},
{'-', &type_int, &type_int, 0},
{'*', &type_int, &type_int, 0},
{'/', &type_int, &type_int, 0},
{'&', &type_int, &type_int, 0},
{'|', &type_int, &type_int, 0},
{'^', &type_int, &type_int, 0},
{'%', &type_int, &type_int, 0},
{MOD, &type_int, &type_int, 0},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{SHL, &type_short},
{SHR, &type_short},
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
{EQ, &type_int, &type_int, 0},
{NE, &type_int, &type_int, 0},
{LE, &type_int, &type_int, 0},
{GE, &type_int, &type_int, 0},
{LT, &type_int, &type_int, 0},
{GT, &type_int, &type_int, 0},
{0, 0}
};
static expr_type_t short_uint[] = {
{'+', &type_uint, &type_uint, 0},
{'-', &type_uint, &type_uint, 0},
{'*', &type_uint, &type_uint, 0},
{'/', &type_uint, &type_uint, 0},
{'&', &type_uint, &type_uint, 0},
{'|', &type_uint, &type_uint, 0},
{'^', &type_uint, &type_uint, 0},
{'%', &type_uint, &type_uint, 0},
{MOD, &type_uint, &type_uint, 0},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{SHL, &type_short},
{SHR, &type_short},
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
{EQ, &type_int, &type_uint, 0},
{NE, &type_int, &type_uint, 0},
{LE, &type_int, &type_uint, 0},
{GE, &type_int, &type_uint, 0},
{LT, &type_int, &type_uint, 0},
{GT, &type_int, &type_uint, 0},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
static expr_type_t short_short[] = {
{'+', &type_short},
{'-', &type_short},
{'*', &type_short},
{'/', &type_short},
{'&', &type_short},
{'|', &type_short},
{'^', &type_short},
{'%', &type_short},
Implement %% (true modulo) support in qfcc However, it's not quite working yet
2020-02-16 02:57:58 +00:00
{MOD, &type_short},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{SHL, &type_short},
{SHR, &type_short},
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
{EQ, &type_int},
{NE, &type_int},
{LE, &type_int},
{GE, &type_int},
{LT, &type_int},
{GT, &type_int},
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{0, 0}
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
#define short_double int_double
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_type_t double_float[] = {
{'+', &type_double, 0, &type_double},
{'-', &type_double, 0, &type_double},
{'*', &type_double, 0, &type_double},
{'/', &type_double, 0, &type_double},
{'%', &type_double, 0, &type_double},
Implement %% (true modulo) support in qfcc However, it's not quite working yet
2020-02-16 02:57:58 +00:00
{MOD, &type_double, 0, &type_double},
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
{EQ, .process = double_compare},
{NE, .process = double_compare},
{LE, .process = double_compare},
{GE, .process = double_compare},
{LT, .process = double_compare},
{GT, .process = double_compare},
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
{0, 0}
};
static expr_type_t double_vector[] = {
[qfcc] Auto-demote double for vector scaling This was missed in the switch to an explicit scale instruction.
2022-02-06 11:10:47 +00:00
{'*', &type_vector, &type_float, 0, vector_scale},
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
{0, 0}
};
static expr_type_t double_quat[] = {
{'*', &type_quaternion},
{0, 0}
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
static expr_type_t double_int[] = {
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
{'+', &type_double, 0, &type_double},
{'-', &type_double, 0, &type_double},
{'*', &type_double, 0, &type_double},
{'/', &type_double, 0, &type_double},
{'%', &type_double, 0, &type_double},
Implement %% (true modulo) support in qfcc However, it's not quite working yet
2020-02-16 02:57:58 +00:00
{MOD, &type_double, 0, &type_double},
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
{EQ, .process = double_compare},
{NE, .process = double_compare},
{LE, .process = double_compare},
{GE, .process = double_compare},
{LT, .process = double_compare},
{GT, .process = double_compare},
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
{0, 0}
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
#define double_uint double_int
#define double_short double_int
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_type_t double_double[] = {
{'+', &type_double},
{'-', &type_double},
{'*', &type_double},
{'/', &type_double},
{'%', &type_double},
Implement %% (true modulo) support in qfcc However, it's not quite working yet
2020-02-16 02:57:58 +00:00
{MOD, &type_double},
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
{EQ, &type_int},
{NE, &type_int},
{LE, &type_int},
{GE, &type_int},
{LT, &type_int},
{GT, &type_int},
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
{0, 0}
};
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
[qfcc] Add basic support for (u)long expressions It's woefully incomplete, but sufficient to test initializing non-scalars from ivec constants. Fixes #36
2022-11-16 11:48:58 +00:00
static expr_type_t long_long[] = {
{'+', &type_long},
{'-', &type_long},
{'*', &type_long},
{'/', &type_long},
{'&', &type_long},
{'|', &type_long},
{'^', &type_long},
{'%', &type_long},
{MOD, &type_long},
{SHL, &type_long},
{SHR, &type_long},
{EQ, &type_int},
{NE, &type_int},
{LE, &type_int},
{GE, &type_int},
{LT, &type_int},
{GT, &type_int},
{0, 0}
};
static expr_type_t ulong_ulong[] = {
{'+', &type_long},
{'-', &type_long},
{'*', &type_long},
{'/', &type_long},
{'&', &type_long},
{'|', &type_long},
{'^', &type_long},
{'%', &type_long},
{MOD, &type_long},
{SHL, &type_long},
{SHR, &type_long},
{EQ, &type_int},
{NE, &type_int},
{LE, &type_int},
{GE, &type_int},
{LT, &type_int},
{GT, &type_int},
{0, 0}
};
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_type_t *string_x[ev_type_count] = {
[qfcc] Clean up the type tables for binary expressions More index labels and remove the null entries since they're not needed with the labels.
2022-01-19 01:52:07 +00:00
[ev_string] = string_string,
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
};
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_type_t *float_x[ev_type_count] = {
[qfcc] Clean up the type tables for binary expressions More index labels and remove the null entries since they're not needed with the labels.
2022-01-19 01:52:07 +00:00
[ev_float] = float_float,
[ev_vector] = float_vector,
[ev_quaternion] = float_quat,
[ev_int] = float_int,
[ev_uint] = float_uint,
[ev_short] = float_short,
[ev_double] = float_double,
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
};
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_type_t *vector_x[ev_type_count] = {
[qfcc] Clean up the type tables for binary expressions More index labels and remove the null entries since they're not needed with the labels.
2022-01-19 01:52:07 +00:00
[ev_float] = vector_float,
[ev_vector] = vector_vector,
[ev_int] = vector_int,
[ev_uint] = vector_uint,
[ev_short] = vector_short,
[ev_double] = vector_double,
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
};
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_type_t *entity_x[ev_type_count] = {
[qfcc] Clean up the type tables for binary expressions More index labels and remove the null entries since they're not needed with the labels.
2022-01-19 01:52:07 +00:00
[ev_entity] = entity_entity,
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
};
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_type_t *field_x[ev_type_count] = {
[qfcc] Clean up the type tables for binary expressions More index labels and remove the null entries since they're not needed with the labels.
2022-01-19 01:52:07 +00:00
[ev_field] = field_field,
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
};
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_type_t *func_x[ev_type_count] = {
[qfcc] Clean up the type tables for binary expressions More index labels and remove the null entries since they're not needed with the labels.
2022-01-19 01:52:07 +00:00
[ev_func] = func_func,
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
};
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_type_t *pointer_x[ev_type_count] = {
[qfcc] Clean up the type tables for binary expressions More index labels and remove the null entries since they're not needed with the labels.
2022-01-19 01:52:07 +00:00
[ev_ptr] = pointer_pointer,
[ev_int] = pointer_int,
[ev_uint] = pointer_uint,
[ev_short] = pointer_short,
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
};
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_type_t *quat_x[ev_type_count] = {
[qfcc] Clean up the type tables for binary expressions More index labels and remove the null entries since they're not needed with the labels.
2022-01-19 01:52:07 +00:00
[ev_float] = quat_float,
[ev_vector] = quat_vector,
[ev_quaternion] = quat_quat,
[ev_int] = quat_int,
[ev_uint] = quat_uint,
[ev_short] = quat_short,
[ev_double] = quat_double,
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
static expr_type_t *int_x[ev_type_count] = {
[qfcc] Clean up the type tables for binary expressions More index labels and remove the null entries since they're not needed with the labels.
2022-01-19 01:52:07 +00:00
[ev_float] = int_float,
[ev_vector] = int_vector,
[ev_ptr] = int_pointer,
[ev_quaternion] = int_quat,
[ev_int] = int_int,
[ev_uint] = int_uint,
[ev_short] = int_short,
[ev_double] = int_double,
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
};
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
static expr_type_t *uint_x[ev_type_count] = {
[qfcc] Clean up the type tables for binary expressions More index labels and remove the null entries since they're not needed with the labels.
2022-01-19 01:52:07 +00:00
[ev_float] = uint_float,
[ev_vector] = uint_vector,
[ev_ptr] = uint_pointer,
[ev_quaternion] = uint_quat,
[ev_int] = uint_int,
[ev_uint] = uint_uint,
[ev_short] = uint_short,
[ev_double] = uint_double,
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
};
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_type_t *short_x[ev_type_count] = {
[qfcc] Clean up the type tables for binary expressions More index labels and remove the null entries since they're not needed with the labels.
2022-01-19 01:52:07 +00:00
[ev_float] = short_float,
[ev_vector] = short_vector,
[ev_ptr] = short_pointer,
[ev_quaternion] = short_quat,
[ev_int] = short_int,
[ev_uint] = short_uint,
[ev_short] = short_short,
[ev_double] = short_double,
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
};
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_type_t *double_x[ev_type_count] = {
[qfcc] Clean up the type tables for binary expressions More index labels and remove the null entries since they're not needed with the labels.
2022-01-19 01:52:07 +00:00
[ev_float] = double_float,
[ev_vector] = double_vector,
[ev_quaternion] = double_quat,
[ev_int] = double_int,
[ev_uint] = double_uint,
[ev_short] = double_short,
[ev_double] = double_double,
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
};
[qfcc] Add basic support for (u)long expressions It's woefully incomplete, but sufficient to test initializing non-scalars from ivec constants. Fixes #36
2022-11-16 11:48:58 +00:00
static expr_type_t *long_x[ev_type_count] = {
[ev_long] = long_long,
};
static expr_type_t *ulong_x[ev_type_count] = {
[ev_ulong] = ulong_ulong,
};
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_type_t **binary_expr_types[ev_type_count] = {
[qfcc] Clean up the type tables for binary expressions More index labels and remove the null entries since they're not needed with the labels.
2022-01-19 01:52:07 +00:00
[ev_string] = string_x,
[ev_float] = float_x,
[ev_vector] = vector_x,
[ev_entity] = entity_x,
[ev_field] = field_x,
[ev_func] = func_x,
[ev_ptr] = pointer_x,
[ev_quaternion] = quat_x,
[ev_int] = int_x,
[ev_uint] = uint_x,
[ev_short] = short_x,
[qfcc] Add basic support for (u)long expressions It's woefully incomplete, but sufficient to test initializing non-scalars from ivec constants. Fixes #36
2022-11-16 11:48:58 +00:00
[ev_double] = double_x,
[ev_long] = long_x,
[ev_ulong] = ulong_x,
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
};
[qfcc] Implement automatic casting between same-width vectors This allows all the tests to build and pass. I'll need to add tests to ensure warnings happen when they should and that all vec operations are correct (ouch, that'll be a lot of work), but vectors and quaternions are working again.
2022-04-29 06:27:12 +00:00
// supported operators for scalar-vector expressions
static int scalar_vec_ops[] = { '*', '/', '%', MOD, 0 };
static expr_t *
convert_scalar (expr_t *scalar, int op, expr_t *vec)
{
int *s_op = scalar_vec_ops;
while (*s_op && *s_op != op) {
s_op++;
}
if (!*s_op) {
return 0;
}
[qfcc] Implement support for the swizzle operator The destination operand must be a full four component vector, but the source can be smaller and small sources do not need to be aligned: the offset of the source operand and the swizzle indices are adjusted. The adjustments are done during final statement emission in order to avoid confusing the data flow analyser (and that's when def offsets are known).
2022-05-01 01:02:26 +00:00
[qfcc] Implement automatic casting between same-width vectors This allows all the tests to build and pass. I'll need to add tests to ensure warnings happen when they should and that all vec operations are correct (ouch, that'll be a lot of work), but vectors and quaternions are working again.
2022-04-29 06:27:12 +00:00
// expand the scalar to a vector of the same width as vec
[qfcc] Implement support for the swizzle operator The destination operand must be a full four component vector, but the source can be smaller and small sources do not need to be aligned: the offset of the source operand and the swizzle indices are adjusted. The adjustments are done during final statement emission in order to avoid confusing the data flow analyser (and that's when def offsets are known).
2022-05-01 01:02:26 +00:00
type_t *vec_type = get_type (vec);
[qfcc] Create vector lists only for constants Defs and symbols benefit from swizzling as that's one instruction vs 2-3 for loading a scalar into a vector component by component. Constants are ok because the result gets converted to a vector constant.
2022-05-01 05:12:43 +00:00
if (is_constant (scalar)) {
[qfcc] Implement support for the swizzle operator The destination operand must be a full four component vector, but the source can be smaller and small sources do not need to be aligned: the offset of the source operand and the swizzle indices are adjusted. The adjustments are done during final statement emission in order to avoid confusing the data flow analyser (and that's when def offsets are known).
2022-05-01 01:02:26 +00:00
for (int i = 1; i < type_width (get_type (vec)); i++) {
expr_t *s = copy_expr (scalar);
s->next = scalar;
scalar = s;
}
return new_vector_list (scalar);
[qfcc] Implement automatic casting between same-width vectors This allows all the tests to build and pass. I'll need to add tests to ensure warnings happen when they should and that all vec operations are correct (ouch, that'll be a lot of work), but vectors and quaternions are working again.
2022-04-29 06:27:12 +00:00
}
[qfcc] Implement support for the swizzle operator The destination operand must be a full four component vector, but the source can be smaller and small sources do not need to be aligned: the offset of the source operand and the swizzle indices are adjusted. The adjustments are done during final statement emission in order to avoid confusing the data flow analyser (and that's when def offsets are known).
2022-05-01 01:02:26 +00:00
[qfcc] Support the extended extend instruction That... feels odd :)
2023-08-25 08:20:28 +00:00
return new_extend_expr (scalar, vec_type, 2, false);//2 = copy
[qfcc] Implement automatic casting between same-width vectors This allows all the tests to build and pass. I'll need to add tests to ensure warnings happen when they should and that all vec operations are correct (ouch, that'll be a lot of work), but vectors and quaternions are working again.
2022-04-29 06:27:12 +00:00
}
Make binary_expr fold constants This is where constant folding should have happened all along. While unary_expr should fold constants too, it seems to already try to do so and it's a bit much of a mess to clean up right now.
2019-06-09 15:36:13 +00:00
static expr_t *
pointer_arithmetic (int op, expr_t *e1, expr_t *e2)
{
[qfcc] Correctly implement pointer arithmetic
2020-03-17 03:45:55 +00:00
type_t *t1 = get_type (e1);
type_t *t2 = get_type (e2);
[qfcc] Fix some not uninitialized warnings The "not" because I'm pretty sure they're false positives due to when the function is called, but clang doesn't know that (wonder why gcc was ok with it).
2022-03-30 15:37:53 +00:00
expr_t *ptr = 0;
expr_t *offset = 0;
[qfcc] Correctly implement pointer arithmetic
2020-03-17 03:45:55 +00:00
expr_t *psize;
[qfcc] Fix some not uninitialized warnings The "not" because I'm pretty sure they're false positives due to when the function is called, but clang doesn't know that (wonder why gcc was ok with it).
2022-03-30 15:37:53 +00:00
type_t *ptype = 0;
Make binary_expr fold constants This is where constant folding should have happened all along. While unary_expr should fold constants too, it seems to already try to do so and it's a bit much of a mess to clean up right now.
2019-06-09 15:36:13 +00:00
[qfcc] Use pr_type_names to generate is_TYPE This means basic types will always have a check function automatically.
2022-01-19 08:48:46 +00:00
if (!is_ptr (t1) && !is_ptr (t2)) {
Make binary_expr fold constants This is where constant folding should have happened all along. While unary_expr should fold constants too, it seems to already try to do so and it's a bit much of a mess to clean up right now.
2019-06-09 15:36:13 +00:00
internal_error (e1, "pointer arithmetic on non-pointers");
}
[qfcc] Use pr_type_names to generate is_TYPE This means basic types will always have a check function automatically.
2022-01-19 08:48:46 +00:00
if (is_ptr (t1) && is_ptr (t2)) {
[qfcc] Correctly implement pointer arithmetic
2020-03-17 03:45:55 +00:00
if (op != '-') {
return error (e2, "invalid pointer operation");
}
if (t1 != t2) {
return error (e2, "cannot use %c on pointers of different types",
op);
}
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
e1 = cast_expr (&type_int, e1);
e2 = cast_expr (&type_int, e2);
psize = new_int_expr (type_size (t1->t.fldptr.type));
[qfcc] Correctly implement pointer arithmetic
2020-03-17 03:45:55 +00:00
return binary_expr ('/', binary_expr ('-', e1, e2), psize);
[qfcc] Use pr_type_names to generate is_TYPE This means basic types will always have a check function automatically.
2022-01-19 08:48:46 +00:00
} else if (is_ptr (t1)) {
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
offset = cast_expr (&type_int, e2);
[qfcc] Rework the functionality of address expressions The goal was to get lea being used for locals in ruamoko progs because lea takes the base registers into account while the constant pointer defs used by v6p cannot. Pointer defs are still used for gobals as they may be out of reach of 16-bit addressing. address_expr() has been simplified in that it no longer takes an offset: the vast majority of the callers never passed one, and the few that did have been reworked to use other mechanisms. In particular, offset_pointer_expr does the manipulations needed to add an offset (unscaled by type size) to a pointer. High-level pointer offsets still apply a scale, though. Alias expressions now do a better job of hanling aliasing of aliases by simply replacing the target type when possible.
2022-01-25 14:39:17 +00:00
ptr = e1;
[qfcc] Correctly implement pointer arithmetic
2020-03-17 03:45:55 +00:00
ptype = t1;
[qfcc] Use pr_type_names to generate is_TYPE This means basic types will always have a check function automatically.
2022-01-19 08:48:46 +00:00
} else if (is_ptr (t2)) {
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
offset = cast_expr (&type_int, e1);
[qfcc] Rework the functionality of address expressions The goal was to get lea being used for locals in ruamoko progs because lea takes the base registers into account while the constant pointer defs used by v6p cannot. Pointer defs are still used for gobals as they may be out of reach of 16-bit addressing. address_expr() has been simplified in that it no longer takes an offset: the vast majority of the callers never passed one, and the few that did have been reworked to use other mechanisms. In particular, offset_pointer_expr does the manipulations needed to add an offset (unscaled by type size) to a pointer. High-level pointer offsets still apply a scale, though. Alias expressions now do a better job of hanling aliasing of aliases by simply replacing the target type when possible.
2022-01-25 14:39:17 +00:00
ptr = e2;
[qfcc] Correctly implement pointer arithmetic
2020-03-17 03:45:55 +00:00
ptype = t2;
}
[qfcc] Rework the functionality of address expressions The goal was to get lea being used for locals in ruamoko progs because lea takes the base registers into account while the constant pointer defs used by v6p cannot. Pointer defs are still used for gobals as they may be out of reach of 16-bit addressing. address_expr() has been simplified in that it no longer takes an offset: the vast majority of the callers never passed one, and the few that did have been reworked to use other mechanisms. In particular, offset_pointer_expr does the manipulations needed to add an offset (unscaled by type size) to a pointer. High-level pointer offsets still apply a scale, though. Alias expressions now do a better job of hanling aliasing of aliases by simply replacing the target type when possible.
2022-01-25 14:39:17 +00:00
// op is known to be + or -
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
psize = new_int_expr (type_size (ptype->t.fldptr.type));
[qfcc] Rework the functionality of address expressions The goal was to get lea being used for locals in ruamoko progs because lea takes the base registers into account while the constant pointer defs used by v6p cannot. Pointer defs are still used for gobals as they may be out of reach of 16-bit addressing. address_expr() has been simplified in that it no longer takes an offset: the vast majority of the callers never passed one, and the few that did have been reworked to use other mechanisms. In particular, offset_pointer_expr does the manipulations needed to add an offset (unscaled by type size) to a pointer. High-level pointer offsets still apply a scale, though. Alias expressions now do a better job of hanling aliasing of aliases by simply replacing the target type when possible.
2022-01-25 14:39:17 +00:00
offset = unary_expr (op, binary_expr ('*', offset, psize));
return offset_pointer_expr (ptr, offset);
Make binary_expr fold constants This is where constant folding should have happened all along. While unary_expr should fold constants too, it seems to already try to do so and it's a bit much of a mess to clean up right now.
2019-06-09 15:36:13 +00:00
}
[qfcc] Correctly implement pointer arithmetic
2020-03-17 03:45:55 +00:00
static expr_t *
pointer_compare (int op, expr_t *e1, expr_t *e2)
{
type_t *t1 = get_type (e1);
type_t *t2 = get_type (e2);
expr_t *e;
if (!type_assignable (t1, t2)) {
return error (e2, "cannot use %s on pointers of different types",
get_op_string (op));
}
[qfcc] Cast pointers to ints for comparison Ruamoko doesn't wast instructions on otherwise compatible low-level types.
2022-01-29 09:23:07 +00:00
if (options.code.progsversion < PROG_VERSION) {
e = new_binary_expr (op, e1, e2);
} else {
e = new_binary_expr (op, cast_expr (&type_int, e1),
cast_expr (&type_int, e2));
}
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
e->e.expr.type = &type_int;
[qfcc] Correctly implement pointer arithmetic
2020-03-17 03:45:55 +00:00
return e;
}
[qfcc] Cast functions to ints for comparison Yet more (non-)wasted instructions.
2022-01-29 09:59:38 +00:00
static expr_t *
func_compare (int op, expr_t *e1, expr_t *e2)
{
expr_t *e;
if (options.code.progsversion < PROG_VERSION) {
e = new_binary_expr (op, e1, e2);
} else {
e = new_binary_expr (op, new_alias_expr (&type_int, e1),
new_alias_expr (&type_int, e2));
}
e->e.expr.type = &type_int;
[qfcc] Use float for v6 comparisons While simple types used floats, functions, vectors and entities used ints. This doesn't go so well when compiling for v6 progs.
2023-02-11 11:11:46 +00:00
if (options.code.progsversion == PROG_ID_VERSION) {
e->e.expr.type = &type_float;
}
[qfcc] Cast functions to ints for comparison Yet more (non-)wasted instructions.
2022-01-29 09:59:38 +00:00
return e;
}
Support vector/quaternion division by float Implemented via v*(1/f) or q*(1/f) to give CSE a chance to optimize the division if necessary as otherwise the engine would have to divide every time.
2019-06-09 23:44:36 +00:00
static expr_t *
inverse_multiply (int op, expr_t *e1, expr_t *e2)
{
// There is no vector/float or quaternion/float instruction and adding
// one would mean the engine would have to do 1/f every time
expr_t *one = new_float_expr (1);
return binary_expr ('*', e1, binary_expr ('/', one, e2));
}
[qfcc] Add support for horizontal vector ops And reimplement vector comparison for Ruamoko.
2022-01-30 01:56:15 +00:00
static expr_t *
vector_compare (int op, expr_t *e1, expr_t *e2)
{
if (options.code.progsversion < PROG_VERSION) {
expr_t *e = new_binary_expr (op, e1, e2);
e->e.expr.type = &type_int;
[qfcc] Use float for v6 comparisons While simple types used floats, functions, vectors and entities used ints. This doesn't go so well when compiling for v6 progs.
2023-02-11 11:11:46 +00:00
if (options.code.progsversion == PROG_ID_VERSION) {
e->e.expr.type = &type_float;
}
[qfcc] Add support for horizontal vector ops And reimplement vector comparison for Ruamoko.
2022-01-30 01:56:15 +00:00
return e;
}
int hop = op == EQ ? '&' : '|';
e1 = new_alias_expr (&type_vec3, e1);
e2 = new_alias_expr (&type_vec3, e2);
expr_t *e = new_binary_expr (op, e1, e2);
e->e.expr.type = &type_ivec3;
return new_horizontal_expr (hop, e, &type_int);
}
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
static expr_t *
vector_dot (int op, expr_t *e1, expr_t *e2)
[qfcc] Implement dot and cross product for Ruamoko With explicit operators, even. While they're a tad verbose, they're at least unambiguous and most importantly have the right precedence (or at least adjustable precedence if I got it wrong, but vector ops having high precedence than scalar or component seems reasonable to me).
2022-01-30 05:14:15 +00:00
{
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
expr_t *e = new_binary_expr (DOT, e1, e2);
if (options.code.progsversion == PROG_VERSION) {
e->e.expr.type = &type_vector;
if (op == '*') {
// vector * vector is dot product in v6 progs (ick)
[qfcc] Make the meaning of vec * vec selectable Currently only via pragma (not command line options), but I needed to test the concept. Converting legacy code is just too error prone. Telling the compiler how to treat the operator makes more sense. When * acts as @dot with Ruamoko progs, the result is automatically aliased as a float as this is the legacy meaning (ie, float result for dot product).
2022-02-06 11:05:35 +00:00
e = new_alias_expr (&type_float, e);
}
[qfcc] Implement dot and cross product for Ruamoko With explicit operators, even. While they're a tad verbose, they're at least unambiguous and most importantly have the right precedence (or at least adjustable precedence if I got it wrong, but vector ops having high precedence than scalar or component seems reasonable to me).
2022-01-30 05:14:15 +00:00
} else {
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
e->e.expr.type = &type_float;
[qfcc] Implement dot and cross product for Ruamoko With explicit operators, even. While they're a tad verbose, they're at least unambiguous and most importantly have the right precedence (or at least adjustable precedence if I got it wrong, but vector ops having high precedence than scalar or component seems reasonable to me).
2022-01-30 05:14:15 +00:00
}
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
return e;
}
[qfcc] Implement dot and cross product for Ruamoko With explicit operators, even. While they're a tad verbose, they're at least unambiguous and most importantly have the right precedence (or at least adjustable precedence if I got it wrong, but vector ops having high precedence than scalar or component seems reasonable to me).
2022-01-30 05:14:15 +00:00
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
static expr_t *
vector_multiply (int op, expr_t *e1, expr_t *e2)
{
if (options.math.vector_mult == DOT) {
// vector * vector is dot product in v6 progs (ick)
return vector_dot (op, e1, e2);
}
// component-wise multiplication
expr_t *e = new_binary_expr ('*', e1, e2);
e->e.expr.type = &type_vector;
[qfcc] Implement dot and cross product for Ruamoko With explicit operators, even. While they're a tad verbose, they're at least unambiguous and most importantly have the right precedence (or at least adjustable precedence if I got it wrong, but vector ops having high precedence than scalar or component seems reasonable to me).
2022-01-30 05:14:15 +00:00
return e;
}
[qfcc] Support explicit dot product in v6p code The dot product in v6p code returns a float rather than a vector.
2023-08-17 07:19:37 +00:00
static expr_t *
vector_scale (int op, expr_t *e1, expr_t *e2)
[qfcc] Implement vector scaling for Ruamoko With this, qfcc-tests builds (can't run yet due to unsigned not having any tests and thus the rest the Ruamoko code in QF not building yet).
2022-01-30 05:48:49 +00:00
{
// Ensure the expression is always vector * scalar. The operation is
// always commutative, and the Ruamoko ISA supports only vector * scalar
// (though v6 does support scalar * vector, one less if).
if (is_scalar (get_type (e1))) {
expr_t *t = e1;
e1 = e2;
e2 = t;
}
expr_t *e = new_binary_expr (SCALE, e1, e2);
e->e.expr.type = get_type (e1);
return e;
}
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
static expr_t *
double_compare (int op, expr_t *e1, expr_t *e2)
{
type_t *t1 = get_type (e1);
type_t *t2 = get_type (e2);
expr_t *e;
Handle implicit casts for double/float comparisons
2020-02-23 14:10:10 +00:00
if (is_constant (e1) && e1->implicit && is_double (t1) && is_float (t2)) {
t1 = &type_float;
[qfcc] Clean up handling of value expressions I'd created new_value_expr some time ago, but never used it... Also, replace convert_* with cast_expr to the appropriate type (removes a pile of value check and create code).
2022-04-28 08:37:56 +00:00
e1 = cast_expr (t1, e1);
Handle implicit casts for double/float comparisons
2020-02-23 14:10:10 +00:00
}
if (is_float (t1) && is_constant (e2) && e2->implicit && is_double (t2)) {
t2 = &type_float;
[qfcc] Clean up handling of value expressions I'd created new_value_expr some time ago, but never used it... Also, replace convert_* with cast_expr to the appropriate type (removes a pile of value check and create code).
2022-04-28 08:37:56 +00:00
e2 = cast_expr (t2, e2);
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
}
if (is_double (t1)) {
if (is_float (t2)) {
warning (e2, "comparison between double and float");
} else if (!is_constant (e2)) {
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
warning (e2, "comparison between double and int");
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
}
e2 = cast_expr (&type_double, e2);
Handle implicit casts for double/float comparisons
2020-02-23 14:10:10 +00:00
} else if (is_double (t2)) {
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
if (is_float (t1)) {
warning (e1, "comparison between float and double");
} else if (!is_constant (e1)) {
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
warning (e1, "comparison between int and double");
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
}
e1 = cast_expr (&type_double, e1);
}
e = new_binary_expr (op, e1, e2);
[gamecode] Rename ev_integer to ev_int And other related fields so integer is now int (and uinteger is uint). I really don't know why I went with integer in the first place, but this will make using macros easier for dealing with types.
2022-01-18 04:21:06 +00:00
e->e.expr.type = &type_int;
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
return e;
}
[qfcc] Alias entity to int for comparison The ruamoko ISA has no entity comparison operators because an entity is just an int in disguise.
2022-02-06 11:12:22 +00:00
static expr_t *
entity_compare (int op, expr_t *e1, expr_t *e2)
{
if (options.code.progsversion == PROG_VERSION) {
e1 = new_alias_expr (&type_int, e1);
e2 = new_alias_expr (&type_int, e2);
}
expr_t *e = new_binary_expr (op, e1, e2);
e->e.expr.type = &type_int;
[qfcc] Use float for v6 comparisons While simple types used floats, functions, vectors and entities used ints. This doesn't go so well when compiling for v6 progs.
2023-02-11 11:11:46 +00:00
if (options.code.progsversion == PROG_ID_VERSION) {
e->e.expr.type = &type_float;
}
[qfcc] Alias entity to int for comparison The ruamoko ISA has no entity comparison operators because an entity is just an int in disguise.
2022-02-06 11:12:22 +00:00
return e;
}
[qfcc] Promote types before expanding scalars This fixes the internal error generated by the likes of `(sv_gravity * '0 0 1')` where sv_gravity is a float and `'0 0 1'` is an ivec3: the vector is promoted to vec3 first so that expanding sv_gravity is expanded to vec3 instead of ivec3 (which is not permitted for a float: expansion requires the destination base type to be the same as the source).
2022-09-14 05:39:09 +00:00
#define invalid_binary_expr(_op, _e1, _e2) \
[qfcc] Include C function in diagnostic messages It's kind of redundant with the line number, but it's helpful for seeing the context at a glance.
2023-08-20 05:44:59 +00:00
_invalid_binary_expr(_op, _e1, _e2, __FILE__, __LINE__, __FUNCTION__)
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
static expr_t *
[qfcc] Promote types before expanding scalars This fixes the internal error generated by the likes of `(sv_gravity * '0 0 1')` where sv_gravity is a float and `'0 0 1'` is an ivec3: the vector is promoted to vec3 first so that expanding sv_gravity is expanded to vec3 instead of ivec3 (which is not permitted for a float: expansion requires the destination base type to be the same as the source).
2022-09-14 05:39:09 +00:00
_invalid_binary_expr (int op, expr_t *e1, expr_t *e2,
[qfcc] Include C function in diagnostic messages It's kind of redundant with the line number, but it's helpful for seeing the context at a glance.
2023-08-20 05:44:59 +00:00
const char *file, int line, const char *func)
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
{
[qfcc] Indicate type width in type strings Makes for much more informative error messages for type mismatches (confusing when both sides look the same).
2022-04-26 10:57:43 +00:00
type_t *t1, *t2;
t1 = get_type (e1);
t2 = get_type (e2);
[qfcc] Include C function in diagnostic messages It's kind of redundant with the line number, but it's helpful for seeing the context at a glance.
2023-08-20 05:44:59 +00:00
return _error (e1, file, line, func, "invalid binary expression: %s %s %s",
[qfcc] Indicate type width in type strings Makes for much more informative error messages for type mismatches (confusing when both sides look the same).
2022-04-26 10:57:43 +00:00
get_type_string (t1), get_op_string (op),
get_type_string (t2));
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
}
static expr_t *
reimplement_binary_expr (int op, expr_t *e1, expr_t *e2)
{
expr_t *e;
if (options.code.progsversion == PROG_ID_VERSION) {
switch (op) {
case '%':
{
Don't truncat float % float This allows full usage, eg, x % pi, but otherwise maintains compatibility with integer %
2020-02-14 04:54:26 +00:00
expr_t *tmp1, *tmp2;
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
e = new_block_expr ();
tmp1 = new_temp_def_expr (&type_float);
tmp2 = new_temp_def_expr (&type_float);
Don't truncat float % float This allows full usage, eg, x % pi, but otherwise maintains compatibility with integer %
2020-02-14 04:54:26 +00:00
append_expr (e, assign_expr (tmp1, binary_expr ('/', e1, e2)));
append_expr (e, assign_expr (tmp2, binary_expr ('&', tmp1, tmp1)));
e->e.block.result = binary_expr ('-', e1, binary_expr ('*', e2, tmp2));
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
return e;
}
break;
}
}
return 0;
}
static expr_t *
check_precedence (int op, expr_t *e1, expr_t *e2)
{
if (e1->type == ex_uexpr && e1->e.expr.op == '!' && !e1->paren) {
if (options.traditional) {
[qfcc] Back out incorrect precedence check changes The precedence check changes done in 63795e790bef3b202c3c1baa3174c3c0d42821ea seem to have been incorrect (game-source/ctf produced many false positives), so putting that check against '=' back into the code seems like a good idea (no more false positives). That sounds a bit cargo-cult, but I'm really not sure what I was thinking when I did the changes (probably just tired).
2023-02-11 06:29:16 +00:00
if (op != AND && op != OR && op != '=') {
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
notice (e1, "precedence of `!' and `%s' inverted for "
"traditional code", get_op_string (op));
e1->e.expr.e1->paren = 1;
return unary_expr ('!', binary_expr (op, e1->e.expr.e1, e2));
}
} else if (op == '&' || op == '|') {
if (options.warnings.precedence)
warning (e1, "ambiguous logic. Suggest explicit parentheses "
"with expressions involving ! and %s",
get_op_string (op));
}
}
if (options.traditional) {
if (e2->type == ex_expr && !e2->paren) {
if (((op == '&' || op == '|')
&& (is_math_op (e2->e.expr.op) || is_compare (e2->e.expr.op)))
[qfcc] Back out incorrect precedence check changes The precedence check changes done in 63795e790bef3b202c3c1baa3174c3c0d42821ea seem to have been incorrect (game-source/ctf produced many false positives), so putting that check against '=' back into the code seems like a good idea (no more false positives). That sounds a bit cargo-cult, but I'm really not sure what I was thinking when I did the changes (probably just tired).
2023-02-11 06:29:16 +00:00
|| (op == '='
&&(e2->e.expr.op == OR || e2->e.expr.op == AND))) {
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
notice (e1, "precedence of `%s' and `%s' inverted for "
"traditional code", get_op_string (op),
get_op_string (e2->e.expr.op));
e1 = binary_expr (op, e1, e2->e.expr.e1);
e1->paren = 1;
return binary_expr (e2->e.expr.op, e1, e2->e.expr.e2);
}
if (((op == EQ || op == NE) && is_compare (e2->e.expr.op))
|| (op == OR && e2->e.expr.op == AND)
|| (op == '|' && e2->e.expr.op == '&')) {
notice (e1, "precedence of `%s' raised to `%s' for "
"traditional code", get_op_string (op),
get_op_string (e2->e.expr.op));
e1 = binary_expr (op, e1, e2->e.expr.e1);
e1->paren = 1;
return binary_expr (e2->e.expr.op, e1, e2->e.expr.e2);
}
} else if (e1->type == ex_expr && !e1->paren) {
if (((op == '&' || op == '|')
&& (is_math_op (e1->e.expr.op) || is_compare (e1->e.expr.op)))
[qfcc] Back out incorrect precedence check changes The precedence check changes done in 63795e790bef3b202c3c1baa3174c3c0d42821ea seem to have been incorrect (game-source/ctf produced many false positives), so putting that check against '=' back into the code seems like a good idea (no more false positives). That sounds a bit cargo-cult, but I'm really not sure what I was thinking when I did the changes (probably just tired).
2023-02-11 06:29:16 +00:00
|| (op == '='
&&(e2->e.expr.op == OR || e2->e.expr.op == AND))) {
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
notice (e1, "precedence of `%s' and `%s' inverted for "
"traditional code", get_op_string (op),
get_op_string (e1->e.expr.op));
e2 = binary_expr (op, e1->e.expr.e2, e2);
e2->paren = 1;
return binary_expr (e1->e.expr.op, e1->e.expr.e1, e2);
}
}
} else {
if (e2->type == ex_expr && !e2->paren) {
if ((op == '&' || op == '|' || op == '^')
&& (is_math_op (e2->e.expr.op)
|| is_compare (e2->e.expr.op))) {
if (options.warnings.precedence)
warning (e2, "suggest parentheses around %s in "
"operand of %c",
is_compare (e2->e.expr.op)
? "comparison"
: get_op_string (e2->e.expr.op),
op);
}
}
if (e1->type == ex_expr && !e1->paren) {
if ((op == '&' || op == '|' || op == '^')
&& (is_math_op (e1->e.expr.op)
|| is_compare (e1->e.expr.op))) {
if (options.warnings.precedence)
warning (e1, "suggest parentheses around %s in "
"operand of %c",
is_compare (e1->e.expr.op)
? "comparison"
: get_op_string (e1->e.expr.op),
op);
}
}
}
return 0;
}
[qfcc] Move alias expr inside call block expression This fixes the trampled return value when the first expression aliases the return result.
2020-03-26 11:16:52 +00:00
static int is_call (expr_t *e)
{
return e->type == ex_block && e->e.block.is_call;
}
[qfcc] Promote types before expanding scalars This fixes the internal error generated by the likes of `(sv_gravity * '0 0 1')` where sv_gravity is a float and `'0 0 1'` is an ivec3: the vector is promoted to vec3 first so that expanding sv_gravity is expanded to vec3 instead of ivec3 (which is not permitted for a float: expansion requires the destination base type to be the same as the source).
2022-09-14 05:39:09 +00:00
static type_t *
promote_type (type_t *dst, type_t *src)
{
if (is_vector (dst) || is_quaternion (dst)) {
return dst;
}
return vector_type (base_type (dst), type_width (src));
}
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
expr_t *
binary_expr (int op, expr_t *e1, expr_t *e2)
{
type_t *t1, *t2;
etype_t et1, et2;
expr_t *e;
expr_type_t *expr_type;
convert_name (e1);
[qfcc] Move return save temp into call block This fixes func-expr after the assignment rewrite. Now all tests pass when not optimizing (something not quite right in assignchain when optimizing).
2020-03-13 16:24:13 +00:00
// FIXME this is target-specific info and should not be in the
// expression tree
[qfcc] Give alias expressions their own type While this was a pain to get working, that pain only went to prove the value of using proper "types" (even if only an enum) for different expression types: just finding all the places to edit was a chore, and easy to make mistakes (forgetting bits here and there). Strangely enough, this exposed a pile of *type* aliasing bugs (next commit).
2022-01-08 03:06:52 +00:00
if (e1->type == ex_alias && is_call (e1->e.alias.expr)) {
[qfcc] Move alias expr inside call block expression This fixes the trampled return value when the first expression aliases the return result.
2020-03-26 11:16:52 +00:00
// move the alias expression inside the block so the following check
// can detect the call and move the temp assignment into the block
[qfcc] Give alias expressions their own type While this was a pain to get working, that pain only went to prove the value of using proper "types" (even if only an enum) for different expression types: just finding all the places to edit was a chore, and easy to make mistakes (forgetting bits here and there). Strangely enough, this exposed a pile of *type* aliasing bugs (next commit).
2022-01-08 03:06:52 +00:00
expr_t *block = e1->e.alias.expr;
e1->e.alias.expr = block->e.block.result;
[qfcc] Move alias expr inside call block expression This fixes the trampled return value when the first expression aliases the return result.
2020-03-26 11:16:52 +00:00
block->e.block.result = e1;
e1 = block;
}
Fix multiple function calls in expressions.
2013-09-27 14:07:38 +00:00
if (e1->type == ex_block && e1->e.block.is_call
&& has_function_call (e2) && e1->e.block.result) {
[qfcc] Move alias expr inside call block expression This fixes the trampled return value when the first expression aliases the return result.
2020-03-26 11:16:52 +00:00
// the temp assignment needs to be insided the block so assignment
// code generation doesn't see it when applying right-associativity
[qfcc] Move return save temp into call block This fixes func-expr after the assignment rewrite. Now all tests pass when not optimizing (something not quite right in assignchain when optimizing).
2020-03-13 16:24:13 +00:00
expr_t *tmp = new_temp_def_expr (get_type (e1->e.block.result));
e = assign_expr (tmp, e1->e.block.result);
append_expr (e1, e);
e1->e.block.result = tmp;
Fix multiple function calls in expressions.
2013-09-27 14:07:38 +00:00
}
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
if (e1->type == ex_error)
return e1;
convert_name (e2);
if (e2->type == ex_error)
return e2;
if (e1->type == ex_bool)
e1 = convert_from_bool (e1, get_type (e2));
if (e2->type == ex_bool)
e2 = convert_from_bool (e2, get_type (e1));
if ((e = check_precedence (op, e1, e2)))
return e;
t1 = get_type (e1);
t2 = get_type (e2);
if (!t1 || !t2)
internal_error (e1, "expr with no type");
[qfcc] Start work on implementing geometric algebra This gets only some very basics working: * Algebra (multi-vector) types: eg @algebra(float(3,0,1)). * Algebra scopes (using either the above or @algebra(TYPE_NAME) where the above was used in a typedef. * Basis blades (eg, e12) done via procedural symbols that evaluate to suitable constants based on the basis group for the blade. * Addition and subtraction of multi-vectors (only partially tested). * Assignment of sub-algebra multi-vectors to full-algebra multi-vectors (missing elements zeroed). There's still much work to be done, but I thought it time to get something into git.
2023-08-21 08:37:56 +00:00
if (is_algebra (t1) || is_algebra (t2)) {
return algebra_binary_expr (op, e1, e2);
}
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
if (op == EQ || op == NE) {
if (e1->type == ex_nil) {
t1 = t2;
convert_nil (e1, t1);
} else if (e2->type == ex_nil) {
t2 = t1;
convert_nil (e2, t2);
}
}
Implicitly cast unadorned floating point constants Floating point constants without f or d adornments will automatically cast, without warnings, to the type appropriate to the rest of the expression.
2020-02-23 13:28:54 +00:00
if (is_constant (e1) && is_double (t1) && e1->implicit && is_float (t2)) {
t1 = &type_float;
[qfcc] Clean up handling of value expressions I'd created new_value_expr some time ago, but never used it... Also, replace convert_* with cast_expr to the appropriate type (removes a pile of value check and create code).
2022-04-28 08:37:56 +00:00
e1 = cast_expr (t1, e1);
Implicitly cast unadorned floating point constants Floating point constants without f or d adornments will automatically cast, without warnings, to the type appropriate to the rest of the expression.
2020-02-23 13:28:54 +00:00
}
if (is_constant (e2) && is_double (t2) && e2->implicit && is_float (t1)) {
t2 = &type_float;
[qfcc] Clean up handling of value expressions I'd created new_value_expr some time ago, but never used it... Also, replace convert_* with cast_expr to the appropriate type (removes a pile of value check and create code).
2022-04-28 08:37:56 +00:00
e2 = cast_expr (t2, e2);
Implicitly cast unadorned floating point constants Floating point constants without f or d adornments will automatically cast, without warnings, to the type appropriate to the rest of the expression.
2020-02-23 13:28:54 +00:00
}
[qfcc] Support pointer arithmetic on arrays That is, `array + offset`. This actually works around the bug highlighted by arraylife.r (because the array is explicitly used), but is not a proper solution, so that test still fails of course. However, with this, it's no longer necessary to use `&array[index]` instead of `array + index`.
2023-05-14 03:39:11 +00:00
if (is_array (t1) && (is_ptr (t2) || is_integral (t2))) {
t1 = pointer_type (t1->t.array.type);
e1 = cast_expr (t1, e1);
}
if (is_array (t2) && (is_ptr (t1) || is_integral (t1))) {
t2 = pointer_type (t2->t.array.type);
e2 = cast_expr (t2, e2);
}
Implicitly cast unadorned floating point constants Floating point constants without f or d adornments will automatically cast, without warnings, to the type appropriate to the rest of the expression.
2020-02-23 13:28:54 +00:00
Fix order of operations for implicit casts
2020-02-23 14:18:31 +00:00
et1 = low_level_type (t1);
et2 = low_level_type (t2);
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
if (et1 >= ev_type_count || !binary_expr_types[et1])
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
return invalid_binary_expr(op, e1, e2);
Create more double related tests Including catching warnings :) (yay -Werror)
2020-02-14 14:36:36 +00:00
if (et2 >= ev_type_count || !binary_expr_types[et1][et2])
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
return invalid_binary_expr(op, e1, e2);
[qfcc] Implement some basics for the vector types They're now properly part of the type system and can be used for declaring variables, initialized (using {} block initializers), operated on (=, *, + tested) though much work needs to be done on binary expressions, and indexed. So far, only ivec2 has been tested.
2022-02-03 15:25:31 +00:00
if ((t1->width > 1 || t2->width > 1)) {
[qfcc] Implement automatic casting between same-width vectors This allows all the tests to build and pass. I'll need to add tests to ensure warnings happen when they should and that all vec operations are correct (ouch, that'll be a lot of work), but vectors and quaternions are working again.
2022-04-29 06:27:12 +00:00
// vector/quaternion and scalar won't get here as vector and quaternion
// are distict types with type.width == 1, but vector and vec3 WILL get
// here because of vec3 being float{3}
[qfcc] Promote types before expanding scalars This fixes the internal error generated by the likes of `(sv_gravity * '0 0 1')` where sv_gravity is a float and `'0 0 1'` is an ivec3: the vector is promoted to vec3 first so that expanding sv_gravity is expanded to vec3 instead of ivec3 (which is not permitted for a float: expansion requires the destination base type to be the same as the source).
2022-09-14 05:39:09 +00:00
if (t1 != t2) {
type_t *pt1 = t1;
type_t *pt2 = t2;
if (is_float (base_type (t1)) && is_double (base_type (t2))
&& e2->implicit) {
pt2 = promote_type (t1, t2);
} else if (is_double (base_type (t1)) && is_float (base_type (t2))
&& e1->implicit) {
pt1 = promote_type (t2, t1);
} else if (type_promotes (base_type (t1), base_type (t2))) {
pt2 = promote_type (t1, t2);
} else if (type_promotes (base_type (t2), base_type (t1))) {
pt1 = promote_type (t2, t1);
} else if (base_type (t1) == base_type (t2)) {
if (is_vector (t1) || is_quaternion (t1)) {
pt2 = t1;
} else if (is_vector (t2) || is_quaternion (t2)) {
pt1 = t2;
}
} else {
debug (e1, "%d %d\n", e1->implicit, e2->implicit);
return invalid_binary_expr (op, e1, e2);
}
if (pt1 != t1) {
e1 = cast_expr (pt1, e1);
t1 = pt1;
}
if (pt2 != t2) {
e2 = cast_expr (pt2, e2);
t2 = pt2;
}
}
[qfcc] Add an explicit hadamard operator While the option to make '*' mean dot product for vectors is important, it breaks vector scaling in ruamoko progs as the resultant vector op becomes a dot product instead of the indented hadamard product (ie, component-wise).
2022-11-15 15:06:21 +00:00
int scalar_op = 0;
[qfcc] Implement automatic casting between same-width vectors This allows all the tests to build and pass. I'll need to add tests to ensure warnings happen when they should and that all vec operations are correct (ouch, that'll be a lot of work), but vectors and quaternions are working again.
2022-04-29 06:27:12 +00:00
if (type_width (t1) == 1) {
// scalar op vec
if (!(e = convert_scalar (e1, op, e2))) {
return invalid_binary_expr (op, e1, e2);
}
[qfcc] Add an explicit hadamard operator While the option to make '*' mean dot product for vectors is important, it breaks vector scaling in ruamoko progs as the resultant vector op becomes a dot product instead of the indented hadamard product (ie, component-wise).
2022-11-15 15:06:21 +00:00
scalar_op = 1;
[qfcc] Implement automatic casting between same-width vectors This allows all the tests to build and pass. I'll need to add tests to ensure warnings happen when they should and that all vec operations are correct (ouch, that'll be a lot of work), but vectors and quaternions are working again.
2022-04-29 06:27:12 +00:00
e1 = e;
t1 = get_type (e1);
}
if (type_width (t2) == 1) {
// vec op scalar
if (!(e = convert_scalar (e2, op, e1))) {
return invalid_binary_expr (op, e1, e2);
}
[qfcc] Add an explicit hadamard operator While the option to make '*' mean dot product for vectors is important, it breaks vector scaling in ruamoko progs as the resultant vector op becomes a dot product instead of the indented hadamard product (ie, component-wise).
2022-11-15 15:06:21 +00:00
scalar_op = 1;
[qfcc] Implement automatic casting between same-width vectors This allows all the tests to build and pass. I'll need to add tests to ensure warnings happen when they should and that all vec operations are correct (ouch, that'll be a lot of work), but vectors and quaternions are working again.
2022-04-29 06:27:12 +00:00
e2 = e;
t2 = get_type (e2);
}
[qfcc] Add an explicit hadamard operator While the option to make '*' mean dot product for vectors is important, it breaks vector scaling in ruamoko progs as the resultant vector op becomes a dot product instead of the indented hadamard product (ie, component-wise).
2022-11-15 15:06:21 +00:00
if (scalar_op && op == '*') {
op = HADAMARD;
}
[qfcc] Implement automatic casting between same-width vectors This allows all the tests to build and pass. I'll need to add tests to ensure warnings happen when they should and that all vec operations are correct (ouch, that'll be a lot of work), but vectors and quaternions are working again.
2022-04-29 06:27:12 +00:00
if (type_width (t1) != type_width (t2)) {
// vec op vec of different widths
[qfcc] Implement some basics for the vector types They're now properly part of the type system and can be used for declaring variables, initialized (using {} block initializers), operated on (=, *, + tested) though much work needs to be done on binary expressions, and indexed. So far, only ivec2 has been tested.
2022-02-03 15:25:31 +00:00
return invalid_binary_expr (op, e1, e2);
}
[qfcc] Implement automatic casting between same-width vectors This allows all the tests to build and pass. I'll need to add tests to ensure warnings happen when they should and that all vec operations are correct (ouch, that'll be a lot of work), but vectors and quaternions are working again.
2022-04-29 06:27:12 +00:00
t1 = get_type (e1);
t2 = get_type (e2);
et1 = low_level_type (t1);
et2 = low_level_type (t2);
// both widths are the same at this point
if (t1->width > 1) {
e = new_binary_expr (op, e1, e2);
e->e.expr.type = t1;
return e;
}
[qfcc] Implement some basics for the vector types They're now properly part of the type system and can be used for declaring variables, initialized (using {} block initializers), operated on (=, *, + tested) though much work needs to be done on binary expressions, and indexed. So far, only ivec2 has been tested.
2022-02-03 15:25:31 +00:00
}
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
expr_type = binary_expr_types[et1][et2];
while (expr_type->op && expr_type->op != op)
expr_type++;
if (!expr_type->op)
return invalid_binary_expr(op, e1, e2);
if (expr_type->a_cast)
e1 = cast_expr (expr_type->a_cast, e1);
if (expr_type->b_cast)
e2 = cast_expr (expr_type->b_cast, e2);
Make binary_expr fold constants This is where constant folding should have happened all along. While unary_expr should fold constants too, it seems to already try to do so and it's a bit much of a mess to clean up right now.
2019-06-09 15:36:13 +00:00
if (expr_type->process) {
return fold_constants (expr_type->process (op, e1, e2));
}
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
if ((e = reimplement_binary_expr (op, e1, e2)))
Make binary_expr fold constants This is where constant folding should have happened all along. While unary_expr should fold constants too, it seems to already try to do so and it's a bit much of a mess to clean up right now.
2019-06-09 15:36:13 +00:00
return fold_constants (e);
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
e = new_binary_expr (op, e1, e2);
[qfcc] Make a field name more clear Just "type" isn't very informative, but "result_type" is at least a fair bit better.
2022-01-27 14:29:32 +00:00
e->e.expr.type = expr_type->result_type;
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
if (is_compare (op) || is_logic (op)) {
Move short-circuit boolean code to its own file
2020-02-23 16:20:24 +00:00
if (options.code.progsversion == PROG_ID_VERSION) {
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
e->e.expr.type = &type_float;
Move short-circuit boolean code to its own file
2020-02-23 16:20:24 +00:00
}
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
}
Make binary_expr fold constants This is where constant folding should have happened all along. While unary_expr should fold constants too, it seems to already try to do so and it's a bit much of a mess to clean up right now.
2019-06-09 15:36:13 +00:00
return fold_constants (e);
Rewrite binary_expr(). It is now in its own file and uses table lookups to check for valid type and operator combinations, and also the resulting type of the expression. This probably breaks multiple function calls in the one expression.
2013-09-27 08:48:06 +00:00
}
Reference in a new issue Copy permalink