Zend: fix overflow handling bug in non-x86 fast_add_function()

The 'result' argument of fast_add_function() may alias with either
of its operands (or both). Take care not to write to 're

Zend: fix overflow handling bug in non-x86 fast_add_function()

The 'result' argument of fast_add_function() may alias with either
of its operands (or both). Take care not to write to 'result' before
reading op1 and op2.

show more ...

Zend: fix overflow handling bug in non-x86 fast_add_function()

The 'result' argument of fast_add_function() may alias with either
of its operands (or both). Take care not to write to 're

Zend: fix overflow handling bug in non-x86 fast_add_function()

The 'result' argument of fast_add_function() may alias with either
of its operands (or both). Take care not to write to 'result' before
reading op1 and op2.

show more ...

permanently deactivate that place, not on runtime only

besides the two "if" checks, this fixes static analyze which is sometimes
broken because of this

Constant expressions refactoring

Added a few more operators

Fixed mem leaks, added tests and ternary operator

Working commit for constant scalar expressions (with constants).
Tests will follow.

Fixed compiler warnings

warning: 'local_dval' may be used uninitialized in this function
[-Wmaybe-uninitialized]

warning: 'dval2' may be used uninitialized in this function

Fixed compiler warnings

warning: 'local_dval' may be used uninitialized in this function
[-Wmaybe-uninitialized]

warning: 'dval2' may be used uninitialized in this function
[-Wmaybe-uninitialized]

show more ...

Constify streams API and a few other calls down the rabbit hole.
(`char *` to `const char *` for parameters and few return values)
In a few places int len moved to size_t len.

Implement internal operator overloading

As pre RFC https://wiki.php.net/rfc/operator_overloading_gmp

Fix #64780 (PHP 5.5 builds are broken with GCC 3)

A recent change (by me) introduced a call to __builtin_offsetof()
into zend_operators.h which is not defined by GCC prior to
version

Fix #64780 (PHP 5.5 builds are broken with GCC 3)

A recent change (by me) introduced a call to __builtin_offsetof()
into zend_operators.h which is not defined by GCC prior to
version 4.

Changed the code to use offsetof() instead: this is defined in
<stddef.h>, so #include this header conditionally (#ifdef GNUC)

show more ...

Fix rounding of zend_dval_to_lval

Now rounding is always towards zero; the rounding can be though as if
occurring on the original double. Only relevant for the 32-bit long
variant.

Fix zend_dval_to_lval outside 64bit integers range

PHP should preserve the least significant bits when casting from double
to long. Zend.m4 contains this:

AC_DEFINE([ZEND_DVAL_T

Fix zend_dval_to_lval outside 64bit integers range

PHP should preserve the least significant bits when casting from double
to long. Zend.m4 contains this:

AC_DEFINE([ZEND_DVAL_TO_LVAL_CAST_OK], 1, [Define if double cast to long preserves least significant bits])

If ZEND_DVAL_TO_LVAL_CAST_OK is not defined, zend_operators.h had an
inline implementation of zend_dval_to_lval() that would do a cast to an
int64_t (when sizeof(long) == 4), then a cast to unsigned long and
finally the cast to long.

While this works well for doubles inside the range of values of the type
used in the first cast (int64_t in the 32-bit version and unsigned long
in the 64-bit version), if outside the range, it is undefined behavior
that WILL give varying and not particularly useful results.

This commit uses fmod() to first put the double in a range that can
safely be cast to unsigned long and then casts this unsigned long to
long. This last cast is implementation defined, but it's very likely
that this gives the expected result (i.e. the internal 2's complement
representation is unchanged) on all platforms that PHP supports. In any
case, the previous implementationa already had this assumption.

This alternative code path is indeed significantly slower than simply
casting the double (almost an order of magnitude), but that should not
matter because casting doubles with a very high absolute value is a
rare event.

show more ...

Improve x86 inline assembler

- added cc annotation to inline asm that clobbers the condition
flags
- remove hardcoded constants (IS_LONG,IS_DOUBLE)
- remove hardcoded offsets (

Improve x86 inline assembler

- added cc annotation to inline asm that clobbers the condition
flags
- remove hardcoded constants (IS_LONG,IS_DOUBLE)
- remove hardcoded offsets (zval->value, zval->type)

show more ...

Fixed bug #64142 (dval to lval different behavior on ppc64)

See discussion on internals
http://marc.info/?t=136042277700003&r=1&w=2

revert

Fixed bug #64142 (dval to lval different behavior on ppc64)

On x86_64:
(long)(double)9223372036854775807+1 = -9223372036854775808
On ppc64
(long)(double)9223372036854775807-1 = 9

Fixed bug #64142 (dval to lval different behavior on ppc64)

On x86_64:
(long)(double)9223372036854775807+1 = -9223372036854775808
On ppc64
(long)(double)9223372036854775807-1 = 9223372036854775807

show more ...

Bug #46408: Fix double formatting for PostgreSQL bound parameters

Happy New Year

Happy New Year

Happy New Year

zend_binary_strncasecmp_l used w/out declaration

fix bug #54547

fix bug #54547

- Year++