As a result of earlier developments, there remained no other
instatiations of the `ALLOC`, `DEALLOC`, and `DECLARE_ROUTINES_FOR`
templates. So, it's simpler to write the alloc/dealloc routines
directly for `ecma_number_t` as well.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
- Code contains syntactic duplications (e.g., build options copy-
pasted repeatedly).
- Build options for test suites contain semantic duplications
(ES5.1 profile builds happen multiple times, once by not
specifying a profile and once by specifying es5.1 profile
explicitly).
- External build options are not taken into account when detecting
duplicated builds.
This patch provides improvement for these issues.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
None of the code paths have been tested for long and especially the
Freya code paths have been long abandoned. This patch:
- Merges Valgrind-Freya into Valgrind code path (there should be no
need to choose between them, Valgrind should work just fine).
- Removes leftover code (`VALGRIND_FREYA_CHECK_MEMPOOL_REQUEST` and
`valgrind_freya_mempool_request`).
- Adds `JMEM_` prefix to Valgrind-related macros (to correctly
leave the `VALGRIND_` prefix to Valgrind).
- Moves the definition of the Valgrind-related macros to a common
header to avoid duplication.
Note: Adding a CI job to perform Valgrind Memchecks is left for
follow-up as it turns out to be excessively slow (>50 mins for a
`--jerry-tests --jerry-test-suite` run, and even a simple
`--jerry-tests` may get terminated prematurely because of timeout
issues).
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
If the result of a build is not directly executable on the host
system or needs an execution runtime for any other reason (e.g.,
cross-compiled ARM binaries on Intel, emscripten-generated JS
files, binaries built with Valgrind support) then the current test
runners cannot work with them. This patch makes test runners
execute binaries via a runtime given in the `$RUNTIME` environment
variable (if it is specified).
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
`JERRY_UNREACHABLE`s often signal code structure that could be
improved: they can usually either be rewritten to `JERRY_ASSERT`s
or eliminated by restructuring loops, `if`s or `#if`s. Roughly,
the only valid occurences are in default cases of `switch`es. And
even they can often be merged into non-default cases.
Moreover, it is dangerous to write meaningful code after
`JERRY_UNREACHABLE` because it pretends as if there was a way to
recover from an impossible situation.
This patch rewrites/eliminates `JERRY_UNREACHABLE`s where possible
and removes misleading code from after them.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
A more recent cppcheck has revealed some extra issues not detected
by the old one running on CI.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The cmakelists contained various compiler-specific parts scattered
across the file. This patch moves them to as few conditional blocks
as possible.
Additional changes:
- `-Werror` does not depend on whether jerry-libc is enabled.
- Some stylistic fixes.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Custom commands with outputs used in multiple targets are prone to
being executed multiple times in parallel builds (and the repeated
overwrites of the outputs may lead to various hard-to-debug
errors). The fix is to add a custom target that depends on the
custom command and make the existing targets using the outputs
depend on the new custom target.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The patch changes the cmakelists of the doctests to extract the
test sources from the markdown docs into the binary tree.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Container-based VMs promise significantly faster boot times.
More or less related changes:
- Added a JOBNAME to all jobs.
- Added `install-noapt` target to several Makefile.travis files to
avoid `sudo apt-get install ...` steps. Those installations are
now handled by the apt addon of Travis. The `install` targets are
kept anyway to keep the makefiles self-contained.
- Removed a legacy workaround from the Coverity Scan job as it
isn't necessary anymore to fiddle with the cerificates of
scan.coverity.com.
- Fixed the Mbed and the Zephyr jobs.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Some code paths explicitly used double instead of ecma_number_t
even though the values assigned from or compared to were of type
ecma_value_t.
Related to #2251
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
In general, public headers should not have compiler-specific
constructs but both the core and the port headers have attributes,
which are non-standard. It's better to factor out such constructs
to a common place (a new header) and hide them behind macros, which
can then be defined on a per-compiler basis.
This patch moves the existing definitions of function attributes and
likely/unlikely builtins to the new header. At the same time, it
unifies the names of these attribute defines and where they are
used. Moreover, it touches on jerry-main and removes the uses of
`__attribute__((unused))` entirely and replaces them with the
elsewhere used `(void) ...` pattern.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
- Various constructs could be expressed with simpler and/or more
readable code.
- The jerry_port_log implementation for the debugger case was prone
to buffer overflow error.
- Some documentation was still missing (even from
jerryscript-port.h).
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The patch fixes both the doc comments in the port header and the
corresponding markdown documentation.
The patch also removes an internal configuration macro guard
exposed by the public port header.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
- Status messages in CMakeLists.txt files got misaligned, fixing.
- An extra space sneaked in before an `#ifdef`, fixing.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
On argument handling:
- Merged all signoff check options into one, option value chooses
between strict/tolerant check variants.
- Added sensible metavars to options with arguments (consistent
with metavars of build.py).
- Reordered options to match the order the checks are executed in.
- Added `--all` alias to `--precommit` check option.
- Beautified code (always placed `help` on separate line, removed
unnecessary arguments of `add_argument` that just repeated their
defaults).
On test runner logic:
There was too many code duplication on how checks were executed, it
got refactored. No change in semantics.
The only change was in Travis CI-related invocations,
`--check-signed-off-travis` had to be rewritten to
`--check-signed-off=travis`.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Until now, the engine's feature set was configurable either via
`CONFIG_*` macro guards defined individually on compiler command
line or via profiles (which are text files listing macro guards,
picked up by the cmake build system and turning them into compiler
command line options). And the features under profile control are
all enabled by default (i.e., all macros are `CONFIG_DISABLE_*`).
This causes a maintenance issue when new features are added to the
engine, because the disabling macros have to be added to all
profiles that don't include the new features. This can even cause
"compatibility break" for applications that embed JerryScript but
don't use the cmake or the python build system, because then
profiles are unavailable and all feature disabling guards have to
be explicitly passed to the compiler. (I.e., if such an application
wants to use the ES5.1 feature set, it must define all the ES2015
disable macros; if the engine is developed further and a new ES2015
feature gets implemented, then the new feature will sneak into the
application's binary unless its own build system is changed to add
the new feature guard.) Even the in-repo example Curie BSP target
seems to have suffered from this maintenance problem.
This patch introduces two new grouping macro guards that enable the
disabling of all ES5.1 builtins and all ES2015 features. As the
grouping logic is in config.h, the maintenance of non-cmake-based
build systems becomes easier (and there is no change for the python
and cmake-based build systems).
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The 'external' toolchain file does nothing but transitively sets
some cmake system variables from values received on the command
line, and forcibly sets the C compiler. However, the same cmake
system variables can be directly set via the command line, together
with the C compiler, and specifying a toolchain is not a must.
Thus, this patch drops the superfluous 'external' toolchain file
and updates cmake-based targets to invoke cmake in a simpler form.
Related changes in this commit:
- While updating the cmake invocations, all the command line
arguments have been reviewed and simplified (removed those, which
did not change the defaults).
- Removed unnecessary forced C compiler settings from some
toolchain files (and/or changed them to setting the "compiler
works" flag to true, thus keeping cmake's compiler identification
logic but disabling some of its overzealous compiler sanity
checks).
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Hitherto, code under the `targets` directory was not tested and so
its maintenance was sometimes speculative. This commit adds build
testing for several targets to prevent them from bit rotting.
Targets covered by this commit are: ESP8266, Mbed, Mbed OS 5,
NuttX, RIOT, Tizen RT, and Zephyr.
Some issues were revealed and fixed:
- ESP8266: added missing include for `uint32_t` typedef.
- Tizen RT: replaced missing `str_to_uint` with `strtol`.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Some benchmark suites contain test cases that have nonreproducible
behaviour. This is mostly caused by relying on "now" when dealing
with dates or timestamps, instead of using a fixed moment. (A
notorious example is the crypto-aes.js test case of the sunspider
bechmark suite, where the heap memory consumption can vary between
34K-41K heap because of using `(new Date()).getTime()`.)
This commit renames the jerry-minimal command line tool to
jerry-test (to better reflect its purpose) and adds extra code,
which intercepts some calls to libc (`gettimeofday`, `rand`) and
pins their results to some fixed values. This makes the tool
useless in a general case but ensures stable results when
benchmarking -- for which it is mostly used.
As a side effect, the commit also changes jerry-libc by making all
libc functions weak symbols to allow their override from
application code.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Over time, `.travis.yml` grew somewhat organically and became quite
complex. With some rewrites, it can set up the Travis CI stages
better, and become simpler, more logical, and more maintainable.
The refactoring keeps all existing functionality.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Also includes the addition and styling of some doc comments (but
those are whitespace too).
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The issue was reported against v1.0 and isn't reproducible anymore.
Still, adding the then-faulty input to the regression test suite to
prevent it occuring again.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
* Replace JERRY_JS_PARSER feature guard with JERRY_DISABLE_JS_PARSER
All feature guards of jerry-core are deciding about the inclusion
or exclusion of a feature based on the state (defined or undefined)
of a macro -- except for the JS parser guard, which requires
`JERRY_JS_PARSER` to be defined with a value of either 0 or 1. This
has some issues:
- The engine cannot be built with a "clean" compiler invocation,
i.e., without any `-D` command line macro definitions. This is
painful for targets that must use a different build system from
the project's own python/cmake-based one.
- Some build systems in targets and even some code in jerry-code
are already confused about the different semantics of
`JERRY_JS_PARSER`, and simply define it without a value and make
decisions based on the macro being simply defined or not.
This patch renames the guard to `JERRY_DISABLE_JS_PARSER` and makes
use of it in jerry-core based on its state, not based on its value.
As obvious from the guard name, the default for the JS parser is
that it is included in the build.
The patch also touches those targets in the repository that
explicitly defined the original macro (correctly or incorrectly).
* Make cppcheck verbose
Cppcheck can be quite slow sometimes, especially on Travis CI,
which has a "10 mins without output means failure" rule. As the
code base of the project grows, we start to undeterministically
fall over that limit. Thus, this PR makes cppcheck verbose to
ensure that it keeps Travis CI continuously fed with output.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
- On OSX, `find` is more picky about its arguments than on Linux
and requires the start directory(-ies) to be explicitly mentioned
before the expressions (a.k.a. "primaries"). However, until now,
the `pylint` runner script called `find` without a start dir,
which caused an error on OSX.
- Until now, `find` looked for python scripts in all subdirectories
of the project, causing `pylint` to check scripts under `targets/`
and `tests/test262` (if `test262` was checked out). However, the
files under `targets/` are less strictly controlled while
`test262` is completely beyond the control of the project. This
caused a lot of superfluous lint warnings.
This patch makes the `pylint` runner script call `find` in the
cross-platform way, and also limits the scope of the lint checks to
the `tools/` and `jerry-debugger/` directories.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Markdown files in the docs/ directory can now be annotated to turn
fenced C code blocks into unit tests. The recognized syntax is:
[doctest]: # (name="test.c", test="run")
```c
// unit test code
```
The commit also fixes the issues revealed during the initial
annotation.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Convert debug server port from a compile-time constant to a context
variable. This enables each engine instance (either running in the
same process or in different processes) to listen for connections
at different ports, i.e., multiple engines can be debugged at the
same time on the same machine.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
As a side effect, refactor the variable types in
`print_unhandled_exception` to reduce the overuse of sized integer
types (generic `unsigned int` is better than `uint32_t` if there is
no actual requirement on integer width).
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Removed various superfluous declarations, and rewrote some parts to
be nicer or more consistent.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
* Move the job queue from the ports into jerry-core
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
* Remove port notification and keep `jerry_run_all_enqueued_jobs` API only
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
We should keep the `jerry_` prefix for public API which is either
declared in jerry-core/include or implemented in jerry-core/api.
Moreover, we should also keep the convention to use short
identifiers for componentization within jerry-core, and use these
identifiers as directory names, file name prefixes, and identifier
prefixes.
Therefore, the tools/gen-unicode.py-generated arrays in
jerry-core/lit are renamed to use `lit_` prefix instead of `jerry_`.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Eases command line option and sub-command definition, usage summary
and detailed help message printing, and argv processing.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Added `handler` module to `jerry-ext` to contain implementation of
commonly used external function handlers: `assert`, `gc`, and
`print`.
Also adapted jerry-main to use jerry-ext/handler
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Since JerryScript has been a C99 project for a while now, we can
use variable-length arrays. This allows the removal of the
artificial limit on command line arguments in jerry-main.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Moved all public API headers under the `jerry-core/include`
directory. This makes installing all the public headers easier.
Also, should we have new public headers in the future, their
installation will be automatic, there will be no need to update the
build files. Moreover, this aligns better with the structure of
other libraries in the project (in those cases, public headers
always reside in `<library>/include`).
Moved all public API implementations under the `jerry-core/api`
directory. This cleans up the root directory of `jerry-core`,
moving all implementation code under "modules", i.e.,
subdirectories. This also makes the future splitting of the big and
monolithic `jerry.c` along features easier, if needed. (Debugger
and snapshot-related functions are already in separate sources.)
Notes:
* `jerryscript.h` is split up to separate header files along
feature boundaries. These new headers are included by
`jerryscript.h`, so this is not a breaking change but header
modularization only.
* `jerry-snapshot.h` is still under `jerry-core/api`, keeping it as
a non-public header.
* This commit also adapts all targets to the include path change.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The built-in `print` is removed from jerry-core, but an external
`print` implementation is added to jerry-main. From now on, all
embedders of the engine have to implement their own `print` if they
need such a functionality.
For printing results in REPL mode of jerry-main, the external
`print` handler is called directly instead of looking up the `print`
function registered into the global object. (The two are the same,
but the indirection is not needed anymore.)
Because jerry-core does not contain `print` anymore,
`jerry_port_console` is removed from the port API. The default port
is updated, i.e., the implementation of `jerry_port_console` is
removed. Additionally, all references to `jerry_port_console` in
jerry-main are replaced by `printf`.
Speculatively, `jerry_port_console` is also removed from all
non-default targets. Most targets implemented it for the sake of the
engine only; in those targets the removal was trivial. Where the
function was called from the embedder application as well, the
calls were replaced with equivalents (e.g., `printf`, `printk`).
NOTE 1: This is a breaking change!
NOTE 2: This patch still leaves several targets without a JS `print`
implementation.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The config header was full of legacy, no-longer-in-use defines. It
was also included from multiple files where it was not necessary.
This patch removes the unused defines, and cleans up headers so
that only `ecma-globals.h` includes `config.h`.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
PR #1505 added support for TI compiler. It explicitly added a
message to notify the user that static linking is forced. PR #1755
added a more generic approach to signal such forced settings and
adapted the TI-specific static linking notification to this
approach. However, it turned out that TI forcibly changed another
setting, too: it disabled release binary stripping, but without
notification. This patch fixes this by moving the setting override
to a consistent place and adding a notification.
PR #1505 also added some source code changes, most importantly a
complex struct initialization for a variable in
`ecma-objects-general.c`. However, that initialization was coded
as a macro to trick the style checker. This patch gets rid of that
macro and uses proper C99 struct initializer with designators.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The internal header was a legacy, declared two functions only. One
of them was not used anymore at all, the other was used at a single
call site. Moreover, their documentation was outdated. This patch
removes the header and the related function implementations -- as
well as the temptation to use such internal functions from outside
the library.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
This commit changes the concept of JerryScript port implementations
from a simple directory of C source files (which get injected among
the sources of `jerry-core`) into a proper static library (which
may be linked to an application together with `jerry-core`). As a
consequence, this commit introduces a new library to the
JerryScript component architecture: the sources of the default port
implementation form `jerry-port-default`.
Changes in more detail:
- The sources in `targets/default` are moved to `jerry-port/default`
and are turned into a proper static library.
- Actually, the default port implementation has two library
variants, one that implements the bare minimum only
(`jerry-port-default-minimal`) and one that has some extra
functionalities specific to this implementation (the "full"
`jerry-port-default`).
- The new libraries have an interface header in
`jerry-port/default/include`, which extends the common
`jerryscript-port.h` API with functions specific to these
libraries.
- All non-standard port functions have now the
`jerry_port_default_` prefix (this affects `jobqueue_init` and
`jobqueue_run`).
- The jobqueue implementation functions became config macro
independent: it is now the responsibility of the linker to pick
the needed objects from the library, and omit those (e.g.,
jobqueue-related code) that are not referenced.
- Build of the libraries can be controlled with the new
`JERRY_PORT_DEFAULT` cmake option.
- The cmake option `PORT_DIR` is dropped, and `PORT_DIR/*.c` is not
appended to `jerry-core` sources.
- Instead, the `jerry` tool of `jerry-main` links to
`jerry-port-default`, while `jerry-minimal` links to
`jerry-port-default-minimal`.
- `tests/unit-core` tests are also linked to
`jerry-port-default-minimal`.
- Tools adapted.
- `build.py` has `--jerry-port-default` instead of `--port-dir`.
- `check-*.sh` have paths updated (`jerry-port/default` instead
of `targets/default`).
- Miscellaneous.
- Dropped `#ifndef`s from `jerryscript-port.h`. It is a public
header of the `jerry-core` library, which means that it must
not contain configuration-dependent parts (once the library is
built with some config macros and the archive and the headers
are installed, there is no way for the header to tell what
those config macrose were).
- Added documentation comments to the JobQueue Port API (in
`jerryscript-port.h`) and to several default port
implementation functions (in `jerry-port/default`).
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The `conversions` and `ranges` files had separate generators, which
shared some functionality through a 3rd python module, and were
quite similar to each other -- but also had some differences.
Moreover, as they were separate scripts, nothing forced to keep the
outputs in sync (to re-generate them at the same time from the same
inputs).
To fix these maintainability issues, this patch
- merges the two python files and part of the 3rd utility module
into a single script while keeping the existing functionality,
- names the new file `gen-unicode.py` to align better with other
generator script in `tools`, and
- adds some extra documentation how to retrieve the input files
(as they are not part of the repository).
The refactoring affected the utility module as well, so this patch
also
- renames `c_source_helper.py` to `gen_c_source.py` (again, for
naming consistency),
- reorganizes some of its functions not to export module-local
helper functions, and
- adapts `js2c.py`, as it also uses this utility module.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
In #1761, not all unit test paths have been updated in check tools.
This triggers no errors as `check-vera` emits a warning only while
`check-cppcheck` signals nothing. Still, this means that some of
the code in the repository that has been quality-assured, is now
avoiding checks. This patch fixes this.
Moreover, `check-cppcheck` has had incorrect paths for a while: it
looked for source files in the root of the repository, but those
files have been in `jerry-main` for almost a year now. This patch
fixes this, too.
Fallout of the above: `main-unix-minimal.c` had to be improved to
make `check-cppcheck` pass.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
In #1738, `JERRY_FEATURE_DEBUGGER` was added to the sources, but
the docs were not updated.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The unit tests should follow the component structure, so this patch
moves all `jerry-core` unit tests under `tests/unit-core` and the
`jerry-libm` unit tests under `tests/unit-libm`.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The patch aims at easing the startup of the HTML-based JerryScript
debugger client by adding a python script that starts the default
browser with the HTML page implementing the debugger client.
Additionally, the pyton script encodes the host:port address
information of the debug server in the file url and the HTML client
is extended to retrieve this information. Thus, both the console
and the browser-based debugger client can be started similarly.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
All public headers should follow the pattern `jerryscript[-*].h`.
The `jerry-api.h` to `jerryscript.h` renaming has already happened
a while ago, now it's time for the port API header to follow.
This patch
* renames the public header file,
* updates all includes to use the new file name (in `jerry-main`,
in all the targets, and in the docs), and
* keeps `jerry-port.h` as a deprecated forwarding header to leave
some time for external users to follow up with this change.
As a related change, the header of the default port implementation
is also changed to `jerryscript-port-default.h`.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Currently, `lit-magic-strings.inc.h` is manually maintained. This
has several drawbacks:
- keeping the list of magic strings sorted first by length then
alphabetically is error prone,
- it is easy to leave unused magic strings in the list by accident
(e.g., `LIT_MAGIC_STRING_JERRY_UL` is defined as a magic string
but not used anywhere in the code) and,
- it is very hard to add `CONFIG_DISABLE_*_BUILTIN` guards to the
list, even though there are several magic strings, which are used
in some of the configurations only (e.g, "setPrototypeOf" is used
in ES2015 only).
To ease the maintenance of magic strings, this commit moves the
definition of magic strings to a config file
(`lit-magic-strings.ini`), and adds `tools/gen-magic-strings.py` to
generate the `.inc.h` file from this config file and from the use
cases of the strings in the code.
- The magic strings in the config file can appear in any order, the
generator will ensure that they are correctly sorted.
- The generator skips those definitions that are not used anywhere
(and emits a warning to signal that such definitions can be
removed).
- The generator applies the same guards to the definitions in the
`.inc.h` file as found in the code around the use of the strings
to optimize for size.
The commit also changes some builtin-related `.inc.h` files by
adding guards that don't affect functionality but improve the
results of the generator.
To ensure that the invocation of the generator does not get
forgotten, the commit also adds `tools/check-magic-strings.sh` and
binds it into the testing infrastructure.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
* `to_string` is unpythonic, let's use `__str__` instead.
* `\"%s\"` for strings is also unpythonic, let's use `%r`.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
First issue was that `diff -u0` was not supported on OSX, thus got
replaced by `-U0`. Then, it turned out that test outcome was
determined based on diff's stdout, which was empty even though the
exit code was non-0 - this got changed, too.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Currently, the Darwin platform and the TI compiler force some of
the settings to predefined values. This patch ensures that these
overrides are highlighted during build.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
* OSX Finder leaves `.DS_Store` files around; make them ignored.
* The mbedk64f target has been renamed to mbed a long time ago;
follow that change in .gitignore, too.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The file `main-unix.c` contained a lot of recurring patterns in the
command line option handling parts. This patch removes these code
clones with the introduction of two helper functions. The goal of
the patch is to have a smaller and maintainable `jerry` tool (mostly
source-wise).
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Disable release binary stripping on OSX: The linker of OSX does not
support `-s` for link-time stripping. (The option does not cause an
error but clutters build output with warnings.)
No OSX-dependent LTO flags: LTO is disabled on OSX.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Some functions were incorrectly marked as const but were pure only
(or not even pure). Some functions were marked as pure but
qualified as const. Some functions were not attributed at all but
qualified either as pure or const. Some functions had attributes
at definition but not at declaration. This commit fixes these
inconsistencies.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Until now, it was not possible to query the engine whether its
build configuration contained debugging support. This commit adds
the `JERRY_FEATURE_DEBUGGER` label to `jerry_feature_t` and extends
`jerry_is_feature_enabled` to support the new feature flag.
Additionally, the command line tool `jerry` was enhanced to report
a warning when invoked with `--start-debug-server` but linked to a
non-debuggable engine.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The arrays contain the addresses of functions annotated with
constructor or destructor attributes. The support is optional,
requires FEATURE_INIT_FINI cmake option to be set. As of now, the
option is _not_ available in tools/build.py directly, only via
`--cmake-param="-DFEATURE_INIT_FINI=ON"`.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The `run-tests.py` test execution harness validates the correctness
of JerryScript by building various configurations of jerry and
running various test suites and subsets against it. However, until
now, it always built jerry for each (build configuration, test suite)
pair. This commit improves `run-tests.py` by skipping duplicate
builds, sharing the same-built binaries across multiple test suite
runs.
The patch also improves on the style of how `Option`s are specified
and fixes the name of the `jerry_tests-debug-cpointer_32bit`
option.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Profiles used to be in alpha order. However, after the recent
`CONFIG_DISABLE_[ES2015_]*_BUILTIN` renamings, the minimal profile
became unsorted. This commit re-sorts the file.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The default definitions and undefinitions of macros used for built-in
descriptions (SIMPLE_VALUE, NUMBER_VALUE, STRING_VALUE, OBJECT_VALUE,
ROUTINE, ACCESSOR_READ_WRITE, ACCESSOR_READ_ONLY) are heavily cloned
all over the builtin-objects directory. This commit factors them out
into two header files, which are then included in the place of the
clones. This way, maintenance becomes a lot easier: e.g., if a new
macro gets introduced, default definition and undefinition don't have
to be added to all description files (of which there are 56 right now,
and their number will most probably just grow).
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The comments of `ecma_lexical_environment_type_t` document why the
values of the enum are as they are, while referring to
`ecma_object_type_t`. However, they became out of sync lately and
refer to nonexistent enum labels and/or incorrect values. This
commit gets the two enums in sync.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
On OSX, the traditional GNU `timeout` command is available as
`gtimeout`. So, this commit adds some logic to the test runners to
try and find the right command name on the host.
Moreover, the commit also unifies the timeout mechanism across all
test runners by rewriting occurrences of `ulimit`-based timeouting
to use the `[g]timeout` command.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
`Object.prototype.__proto__` has been implemented by most JS
engines to give R/W access to prototype chains, well before it made
it into the standard. JerryScript has decided not to implement it,
exactly because it was not part of ES 5.1. The only fully
ES5.1-compatible way of accessing the prototype chain is
`Object.getPrototypeOf` for reading.
However, ES2015 defines `Object.setPrototypeOf` for rewriting the
prototype chain, and JerryScript has now an ES2015 subset profile.
So, this commit adds its implementation to JerryScript.
Note, this commit does _not_ add `Object.prototype.__proto__`,
since that is in the Annex B of ES2015 specification, which is
optional for non-web-browser hosts.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
For negative out-of-range time values, ecma_date_year_from_time
fell into an infinite loop. (Moreover, for positive out-of-range
time values, it returned an incorrect year.) This patch fixes the
helper to return NaN in these cases (and ensures that its call
sites are prepared for NaN values).
Fixes#1657.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Doxygen reported a few issues, which are fixed by this patch:
* ArrayBuffer-related documentation group names and titles were
incorrect.
* Some Date-related helper functions used incorrect parameter
documentation syntax.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Legacy implementation made incorrect assumptions on how many bits
of useful information is returned by libc's `rand()`. If `RAND_MAX`
had more than 16 useful bits, it assumed that `rand()` could return
32 useful random bits. However, e.g., jerry-libc's `RAND_MAX` is
`0x7fffffff`, which denotes 31 useful bits only. The consequence
was that `Math.random()` covered only the lower half of the
standard-mandated `[0,1)` interval.
This path fixes the error and always uses the exact value of
`RAND_MAX` to compute the random value, which will thus fully cover
`[0,1)`.
Fixes#1414
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
`-g` is enough (most modern compilers use DWARF4 as default, at
least, e.g., GCC has switched from DWARF2 in 4.8), more future
proof (if a compiler moves to a newer version, it wont get forced
to a fixed older version), and less restrictive (some compilers
handle the "standard" `-g` only but don't understand its variants).
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Defining the same macros in multiple source files can cause macro
re-definition warnings or errors in all-in-one builds.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Newer compilers (especially clang) warn (and fail) on
`assert (!"message");` constructs typically used to assert on
unreachable code paths (multiple occurrences in jerry-libc). A more
up-to-date approach is to use `assert (false && "message");`. This
patch applies the pattern to all relevant asserts in jerry-libc.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
This patch improves the command line tool in two ways:
* The tool can be instructed to read and execute a script from
stdin by invoking it as `jerry -`.
* The tool can be instructed to suppress its prompt when in repl
mode by invoking it as `jerry --no-prompt`.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Several non-script files have been added with exec permission to
the repository. This patch revokes the erroneous permissions.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
run-tests.py used to echo the build commands it executed, but did
not behave the same way when it executed test commands. To make its
behavior more traceable, this patch adds echo for test commands as
well.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
From the beginning, we have been configuring cppcheck to check its
input as C++ source. However, the transition to C99 happened a
while ago. This patch switches cppcheck into C99 mode.
Some related changes:
* Progress reporting of cppcheck just clutters the output and makes
warnings hard to discover. Thus, this patch puts cppcheck into a
quieter mode where it prints anything only if a non-suppressed
warning is found.
* The default warning format of cppcheck is a bit different from
usual compiler error/warning format. This patch configures
cppcheck to use a more familiar warning template.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
It's stated in the Guidelines that all contributions must be under
the Apache License 2.0. To avoid potential mistakes from manual
reviews, this patch adds the check-license.py script to
automatically check all source files for license headers.
Travis CI is also configured to run the check.
Fallout: it turned out that some files already in the code base
either miss a license header or have some minor typo differences.
The patch fixes up some of these deficiences.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The naming convention of the project for `struct`s is
`typedef struct x_t { } x_t`, but only if it has self-recursive
pointer members, otherwise `typedef struct { } x_t` is enough.
This patch applies this style consistently throughout the code
base.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
* Cmake had LTO OFF, build.py had LTO ON by default. ON became the
common setting.
* Cmake had snapshot save/exec OFF, build.py had them ON by
default. OFF became the common setting.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Although both jerry-libc and jerry-libm have configuration options
that enable/disable their build, in practice, only jerry-libc can be
replaced with the system (compiler-default) libc. If jerry-libm is
disabled, the build of jerry-main fails, as there is no way to
instruct the linker to link the system libm to the binary. (The
build system does have a way to pass flags to the linker, but those
flags are listed before the linked objects. For the references to
get resolved correctly, the libraries to be linked have to be
specified _after_ the objects.)
This patch adds the EXTERNAL_LINK_LIBS configuration option to
CMakeLists, which ensures that the specified libraries get
correctly passed to the linker. (E.g, replacing jerry-libm with
system libm becomes possible with
`JERRY_LIBM=OFF EXTERNAL_LINK_LIBS='-lm'`.)
Additionally, the patch also makes the following related changes:
* Removes the COMPILER_DEFAULT_LIBC configuration option, as it is
(almost) always the opposite of JERRY_LIBC. Moreover, its name is
misleading: its only role is to add `-nostdlib` to the linker
flags.
* Makes use of transitive library dependencies: if a library has
another library as dependency, and it is linked to a binary, its
dependency is linked as well. Thus, jerry-libc, jerry-libm, and
any external libraries are added to jerry-core as dependency, and
then only jerry-core is linked to executables (cmake will take
care of the rest).
* build.py and run-tests.py follow up the changes, along with some
minor syntax changes.
* Moves static linking option to global CMakeLists, as unit test
binaries should be linked the same way as jerry-main.
* Adds EXTERNAL_COMPILER_FLAGS and EXTERNAL_LINKER_FLAGS as last to
the flag list, to allow user override of (nearly) anything.
The patch speculatively follows up the build system changes in the
mbed, riot-stm32f4, and zephyr targets.
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
The new build script does not support ninja anymore, and we haven't
been building bare-metal arm targets for long. This patch removes
now-obsolete dependencies from install scripts.
This can also help to reduce the load on Travis CI (as it keeps
installing dependencies for each build job over and over again).
JerryScript-DCO-1.0-Signed-off-by: Akos Kiss [email protected]
Akos Kiss