The API doc for jerry_value_is_{true,false} incorrectly stated
that it is a replacment for the `jerry_value_is_boolean` method.
The `jerry_value_is_true` is the direct replacement of the
`jerry_get_boolean_value` method and the "false" version is a new
method.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
During the execution of the TypedArray filter method it is possible
to have a different sized output TypedArray than the input one.
When copying the data to the output array the values must be
correctly converted to the output TypedArray's value range.
Fixes: #4793
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
When executin the TypedArray's slice method it is possible to have
the a different element sized output TypedArray. In such case
the data must be converted to the desired element type/size.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
In the Array.slice method when the engine uses fast arrays the "end" value
was not updated if the input array's length changed. This can occur when the start/end
index normalization executes a method and the length is changed forcefully.
This leads to a buffer-overflow as the element copy reads too much data from the input
array.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
As a few methods were removed and replaced it is a good idea to describe
which method(s) can reproduce the same functionality. In addition add a
bit of extra comments for various methods/structs.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
IRC notification GitHub action returns an error at the moment
and makes the master red (even if all builds passed).
Disable this action till we fix the notification.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
In some parts of Proxy internal methods the compiler can do a bit of
tail call optimalization which would be ok, but the current stack limit
check framework in Jerry uses the stack pointer to calculate the stack depth.
This way of stack depth calculation is affected by the tail call optimalization
(as the stack does not increase).
By disabling the tail call optimalization at given points the stack limit calculation
will work as expected. This causes a bit of stack overhead, but the Proxy in it self
is a fairly big chunk of code and this stack limit would only be relevant if the Proxy
already does recusion which already very edge case.
The stack limit (--stack-limit=..) should be enabled to correctly report such stack
depth errors.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
The usage of a single byte sized array as a flexible array member can lead to
compiler/analyzer errors or warnings. Not specifying the size in the array is
correct as per C99 standard.
Flexible array members are defined in the C Standard, 6.7.2.1 (ISO/IEC 9899:1999).
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
In the lexer_check_property_modifier the assert incorrectly does
an assignment instead of a simple equals check.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
Add additional information for jerry_create_realm method and rework the
example to be a compilable/runnable doctest.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
A few new methods did not have the version info to describe
in which version were they introduced.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
Add previously missed new defines to the documentation and update the
defines to be in lexicographical order.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
Replace the [[NEXT_VERSION]] entries with the exact version number
in the API reference documentation.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
After the introduction of the Proxy builtin object there was
a possibility to traverse the prototype chain with an invalid object.
The prototype was freed before it's data/properties were queried resulting
in accessing invalid information.
By forcing the allocator to always do a gc (`--mem-stres-test=on` build option)
it was possible to trigger the issue without complicated tests.
New internal method:
* `ecma_op_object_get_prototype_of` which always returns the prototype
of an object and the return value must be freed (if it is valid).
Updated prototype chain traversing in:
* `jerry_object_get_property_names`
* `ecma_builtin_object_prototype_lookup_getter_setter`
* `ecma_op_function_has_instance`
* `ecma_op_function_get_super_constructor`
* `ecma_op_object_is_prototype_of`
* `ecma_op_object_enumerate`
Removed method `ecma_proxy_object_prototype_to_cp`
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
When the mem-stress-test is enabled for the objects foreach unittest
the object counting is off by one. This was due to incorrect assumption
on when the gc is triggered.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
In Proxy.[[Get]] if the target.[[GetOwnPropertyDescriptor]] fails the
trap result should be freed.
Fixes: #4466
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
When accessing an error's type there is no need to check the Proxy
prototype (as there is none).
Fixes: #4440
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
When shifting a negative BigInt value if the
shift count is very "big" the result should be -1.
eg.:
```
(-2n >> (2n*32n)) === -1n
```
Note: the `-2n >> (2n*31n)` already returned `-1n` prior this change.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
Proxy flags (IsCallable, IsConstructor) can't be stored on the
property list compressed pointer. As adding a Proxy to a WeakSet
would add a property to the Proxy object causing failures down the line.
The prototype internal "slot" can be used to store there flags as
it is not used in case of Proxies (as per standard).
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
When an error occurs accessing a property during JSON stringify call
the wrapper object is not freed at the correct place.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
Extended the `jerry_type_t` enum with `JERRY_TYPE_BIGINT` and added it to
the `jerry_value_get_type`.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
The assert in the reduce and filter methods did not checked if the given value
is a BigInt. This missing check caused the assert to fail.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
The handling of Proxy.[[Get]] was not fully correctly in the case when the
Proxy was revoked during the execution of the handler.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
The old variant of the syntax error printing method could incorrectly
over-read the source file contents.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
The error object for an incorrect Regexp literal was derefed then the error message
was accessed. This could lead to a state where the message was not available but
it was still accessed.
This could occur in case of heavy memory load or with mem stress mode.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
The previous `ecma_is_constructor` implementation did not checked if the
target function was an arrow or generator function. This resulted in
an incorrect execution for these function types.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
The `IsCallable(target)` and `IsConstructor(target)` info
can't be stored in the target/handler values.
If the input for the ProxyCreate was a revocable Proxy the original target's
callable/constructor information must be retained even after the
Proxy was revoked.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
In case of `for (const {...rest} ...) ` there was an incorrectly handling
of the destructuring pattern.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
In the newer ecma262 standard (post ES11) the ProxyCreate was
changed and the revoked Proxy handler/target is not checked.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
If the module system is disabled however the es.next is enabled the following
code triggers an assert:
```js
class Demo {}
```
This was due to an incorrect scanner data processing where it was always
assumed that there is a module related information.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
Module info should be freed at a later stage to correctly have
all module data in every step during the execution.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
If the prototype of a fast array is changed, for example to a Proxy object,
the array should be converted back to a normal array. This is required to
correctly handle any [[Get]]/[[Set]]/etc... calls which should trigger
these method calls on the before mentioned Proxy object.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
In case of Proxies the "ownKeys" call can return an incompatible
value for a for-in statement. In such cases the error should be
propagated to the user.
Fixes: #4159
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
When the `toLocaleString` was called on a TypedArray's value
the resulting object's `toString/valueOf` invocations could
create errors. These error values were not released.
In addition the input element value for the toString operation
was released twice in case of an error.
Fixes: #4148.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
In case of ES.next the Array.prototype.filter should raise an error
if the property on the target array's data property creation fails.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
With a disabled ES2015 module system the build fails as an enum value
is incorrectly guarded and used.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
In case of ARM Nuttx build with IoT.js there are small conversion
and signedness warnings resulting in a build failure.
```
ecma-builtin-regexp-prototype.c: error: conversion from 'int' to '_Bool8' {aka 'unsigned char'}
may change value [-Werror=conversion]
116 | return ecma_make_boolean_value (flags & re_flags[offset]);
| ~~~~~~^~~~~~~~~~~~~~~~~~
js-scanner-util.c: In function 'scanner_pop_literal_pool':
js-scanner-util.c:628:43: error: comparison of integer expressions of different signedness:
'intptr_t' {aka 'int'} and 'unsigned int' [-Werror=sign-compare]
628 | else if (diff >= -UINT8_MAX && diff <= UINT16_MAX)
| ^~
js-scanner-util.c:891:43: error: comparison of integer expressions of different signedness:
'intptr_t' {aka 'int'} and 'unsigned int' [-Werror=sign-compare]
891 | else if (diff >= -UINT8_MAX && diff <= UINT16_MAX)
```
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
In case of `-fPIC` and single file build one of the Regexp's method
raises an error during linking:
```
ecma-regexp-object.c: In function ‘ecma_regexp_run’:
re-bytecode.c:258:1: error: inlining failed in call to ‘always_inline’ ‘re_get_value’:
function body can be overwritten at link time
```
By adding the `inline` keyword for the method the build error can be resolved.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
In case of a TypedArray the input `byteOffset` argument must be a
multiple of the `BYTES_PER_ELEMENT` value.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
The property descriptors must be free'd during the iteration in the
Object.assign even if the property is not enumerable (or there is no value/getter).
Fixes: #4048
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
The toJSON method should be searched on the `this` object
only once. In case of normal objects there were no problems
however when a Proxy object is used as the `this` argument
for the Date.prototype.toJSON method the Proxy's `get` method
was invoked twice.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
From ES 6 the ToPropertyKey operation is used to convert elements to a valid property key.
This "new" method uses the ToPrimitive operation which can call the `@@ToPrimitive`
well-known symbol to convert the given element to a key.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
This change introduces simple stack measuring by coloring the stack
before doing the actual JS work.
The dynamic linking should be avoided as it will use the stack during
the symbol resolving.
At this moment the `JERRY_STACK_MEASURE` compile time define will enable
the measurement for the jerry-test binary.
A usable build configuration:
```
./tools/build.py --jerry-cmdline-test=on --compile-flag=-DJERRY_STACK_MEASURE=1 --compile-flag=-static
```
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
The Date.prototype[@@toPrimitive] only allows the "string", "default" and "number"
hint values. Any other value should throw a TypeError.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
* Added more details into documentation.
* Moved the C unit-test into its own file.
* Added extra test cases.
* Extended the API reference documentation with doctests.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
In case of an empty array is used as a replacer for the JSON.stringify,
the output string should not contain any keys.
Fixes: #3926
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
The DataView's third argument incorrectly processed the "undefined"
value resulting in a RangeError.
Fixes: #3927
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
During the `opfunc_spread_arguments` argument release process
the stack pointer was incremented early resulting in a state where
one of the arguments was not freed causing a memory leak.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
When the getOwnPropertyDescriptor method was invoked the input property
descriptor was not cleared in every case. This could lead to problems when
the property descriptor is not set/modified by the getOwnPropertyDescriptor
call, resulting in a failure at a later state.
Related to this the Proxy getOwnPropertyDescriptor method incorrectly returned
"undefined" value in a single case.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
Property key query for Proxy objects always returned all keys
even if no symbols were requested symbols were present in the
resulting array.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
When lastIndexOf is executed the array length can change and in
case of fast arrays the length should be re-checked.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
The regexp flag should be correctly referenced and released
if an existing regexp like object is used for constructing a new one.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
The release of the regexp pattern string during creating
was incorrect as it was dereferenced a bit to early.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
During ecma_collection_append the underlying collection
was not increased in the required case. This triggered
a buffer overflow when processing the bound function's arguments
during call or during the Proxy ownKeys method.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
During "put" with receiver the fast arrays were not converted to normal
objects to correctly set properties on it.
JerryScript-DCO-1.0-Signed-off-by: Peter Gal [email protected]
Péter Gál