Initial version of JerryScript debugger #1557

polaroi8d · 2017-02-01T10:05:13Z

The debugger supports setting breakpoints, execution control (step, next, continue)
and getting backtrace. The communication is websocket based, so a browser can
communicate with JerryScript without any intermediate application.

JerryScript-DCO-1.0-Signed-off-by: Zoltan Herczeg [email protected]
JerryScript-DCO-1.0-Signed-off-by: Levente Orban [email protected]

zherczeg · 2017-02-01T10:17:10Z

jerry-core/jerry-api.h

@@ -335,6 +338,7 @@ jerry_value_t jerry_exec_snapshot (const void *snapshot_p, size_t snapshot_size,
 size_t jerry_parse_and_save_literals (const jerry_char_t *source_p, size_t source_size, bool is_strict,
                                      uint8_t *buffer_p, size_t buffer_size, bool is_c_format);

+


No need this newline.

zherczeg · 2017-02-01T10:18:12Z

jerry-core/jerry.c

+#endif /* JERRY_DEBUGGER */
+
+  return jerry_parse (source_p, source_size, is_strict);
+} /* jerry_parse_named_resource */


We need an api help / description for this function.

Added, thanks.

I'm trying to understand how this works in a scenario where there are multiple .js source files.
It looks like there is no information kept after this function call that can be used to associate the bytecode back to the resource name. Perhaps I am missing something? Or is it just a TODO for the future?

This information is sent to the debugger who keeps and use this. The concept is described in the documentation.

zherczeg · 2017-02-01T10:21:03Z

jerry-core/parser/js/js-parser-statm.c

+      {
+        if (ident_line_counter != context_p->last_breakpoint_line)
+        {
+          JERRY_DEBUG_MSG ("Insert var breakpoint: %d (%d)\n", ident_line_counter, context_p->last_breakpoint_line);


Somehow these messages are appear on the console. Why?

Personally I would remove these two prints.

LaszloLango · 2017-02-01T10:26:49Z

jerry-core/debugger/jerry-debugger.c

+/**
+ * Debugger socket communication port.
+ */
+#define PORT 5001


These defines should use a prefix. What about JDEBUGGER_?

We use JERRY_DEBUGGER prefix.

dbatyai · 2017-02-01T10:29:41Z

jerry-core/debugger/jerry-debugger.c

+  WEBSOCKET_TEXT_FRAME = 1, /**< text frame */
+  WEBSOCKET_BINARY_FRAME = 2, /**< binary frame */
+  WEBSOCKET_CLOSE_CONNECTION = 8, /**< close connection */
+  WEBSOCKET_PING = 8, /**< ping (keep alive) frame */


Do PING and CLOSE_CONNECTION have the same value intentionally?

Nope. I will fix it.

dbatyai · 2017-02-01T10:47:42Z

jerry-core/debugger/jerry-debugger.c

+/**
+ * Process WebSocket handshake.
+ *
+ * @return true - is no error is occured


Should be: "if no error occurred"

dbatyai · 2017-02-01T11:18:54Z

jerry-core/debugger/jerry-debugger.c

+/*
+ * Send the message to the client side
+ *
+ * @return true - if the data was send successfully to the client side


Should be: "if the data was sent successfully ..."

dbatyai · 2017-02-01T11:21:07Z

jerry-core/debugger/jerry-debugger.h

+
+/**
+ * Limited resources available for the engine, so it is important to
+ * check the maximum buffer size. It need to be between 64 and 256.


It needs to be

LaszloLango · 2017-02-01T13:28:58Z

jerry-core/debugger/jerry-sha1.c

+  sha1_context_p->state[4] = 0xC3D2E1F0;
+} /* jerry_sha1_init */
+
+#define P(a, b, c, d, e, x) \


Missing description.

LaszloLango · 2017-02-01T13:29:41Z

jerry-core/debugger/jerry-sha1.c

+  sha1_context_p->total[0] = 0;
+  sha1_context_p->total[1] = 0;
+
+  sha1_context_p->state[0] = 0x67452301;


Magic literals. Please give some comment at least.

This come from the standard, we don't need a description for it, I think.

Perhaps a description from what standard it came? ;-)

The SHA-1 standard is mentioned in the first non-license block comment of the source file. I don't see the point of repeating it here. We could repeat it in every line of the code, since we invented nothing new here. I think the purpose (setting the initial state) is quite obvious.

Right. Which only shows why comments might be necessary even in places where they don't seem to be, initially - because people will mis-read code, ignore context, etc. In this case I take the blame for doing so, and agree that surrounding context gives enough hints about it.

LaszloLango · 2017-02-01T13:31:16Z

jerry-core/debugger/jerry-sha1.c

+
+#define SHIFT(x, n) ((x << n) | ((x & 0xFFFFFFFF) >> (32 - n)))
+
+#define R(t) \


LaszloLango · 2017-02-01T13:31:57Z

jerry-core/debugger/jerry-sha1.c

+  GET_UINT32_BE (W[14], data, 56);
+  GET_UINT32_BE (W[15], data, 60);
+
+#define SHIFT(x, n) ((x << n) | ((x & 0xFFFFFFFF) >> (32 - n)))


SHA1_SHIFT

LaszloLango · 2017-02-01T13:34:18Z

jerry-core/ecma/base/ecma-helpers.c

@@ -19,6 +19,8 @@
 #include "ecma-helpers.h"
 #include "ecma-lcache.h"
 #include "ecma-property-hashmap.h"
+#include "jcontext.h"
+#include "jerry-debugger.h"


#ifdef JERRY_DEBUGGER #include "jcontext.h" #include "jerry-debugger.h" #endif /* JERRY_DEBUGGER */

LaszloLango · 2017-02-01T13:38:09Z

jerry-core/parser/js/js-parser-internal.h

+/**
+ * Extra information for each breakpoint
+ */
+typedef struct parser_breakpoint_info_t


typedef struct {

LaszloLango · 2017-02-01T13:44:07Z

jerry-core/parser/js/js-parser-statm.c

@@ -1559,6 +1599,14 @@ parser_parse_statements (parser_context_t *context_p) /**< context */
  parser_stack_push_uint8 (context_p, PARSER_STATEMENT_START);
  parser_stack_iterator_init (context_p, &context_p->last_statement);

+#ifdef JERRY_DEBUGGER
+  /* Set lexical enviroment for the debugger*/


missing space before */

LaszloLango · 2017-02-01T13:47:11Z

jerry-core/vm/vm-defines.h

@@ -48,6 +48,7 @@ typedef struct
  ecma_value_t *stack_top_p;                          /**< stack top pointer */
  jmem_cpointer_t *literal_start_p;                   /**< literal list start pointer */
  ecma_object_t *lex_env_p;                           /**< current lexical environment */
+  struct vm_frame_ctx_t *prev_context_p;              /**< previous context */


Is this used, when the debugger is disabled?

It is used now. Instead of a local variable it is stored as part of the structure.

LaszloLango · 2017-02-01T13:58:57Z

jerry-main/main-unix.c

+    else if (!strcmp ("--start-debug-server", argv[i]))
+    {
+      flags |= JERRY_INIT_DEBUGGER;
+      jerry_port_default_set_log_level (JERRY_LOG_LEVEL_DEBUG);


https://github.com/jerryscript-project/jerryscript/pull/1557/files#r98863430
This is why.

Please remove log level setting.

jiangzidong · 2017-02-03T01:08:07Z

@polaroi8d It's a very useful feature!
Could you also please add some docs about the config/instructions/examples about "How to debug JS code in JerryScript" after this patch is landed?

zherczeg · 2017-02-03T08:46:27Z

@jiangzidong you need to pass --jerry-libc off and --jerry-debugger on to build.py.

After that run jerry with --start-debug-server and pass a JS file as well. The debugger is a remote debugger, so it will listen for a connection. In the jerry-debugger/ directory there are two debugger clients: a python and a html client. Both are stand-alone. I think the best feature of JerryDebugger is that you only need a browser to do debugging, no need any other application. Run either of these clients and type help. Have fun.

martijnthe · 2017-02-03T13:25:10Z

@polaroi8d this is incredibly awesome!

Some (unsolicited ;) comments:

Right now jerry_debugger.c requires having an IP stack in the system. IMHO, it would be better to layer this such that the underlying transport could be swapped out easily, to support devices that do not have an IP stack but want to debug over another type of transport (serial port, Bluetooth, ...) Would it make sense to add an abstraction and let the target provide it through a set of jerry-port.h functions/types?
I'm having a similar reaction to the protocol choice ("simplified version of Websockets"). I'd argue that it's better to keep jerry_debugger.c "transport protocol agnostic" and let the target/port take care of implementing the transport protocol (Websockets may make sense for certain targets, but not others).
How much does jerry-sha1.c add to the code size footprint (for example when targeting ARM CM3)?

zherczeg · 2017-02-03T13:59:50Z

Well, it is certainly possible to make this less dependent on the transportation protocol. Especially the connection create / destroy and message send/receive parts. For packet format I would keep the websocket packet format since it is simple and it would control the complexity of the code (no need an ifdef around every code).

Anyway I would complete the debugger first and based on its final form we could think about how to make it less dependent on network communication. Refactoring is easier in the current form when the code is not full with ifdefs.

galpeter

Just a quick review on html/python parts.

galpeter · 2017-02-03T18:56:06Z

jerry-debugger/jerry-client.html

+
+      assert(i != -1);
+
+      for (; i < newLength; i++)


I think a simple array.splice(i, 1); is enough, no need to re-implement the functionality.

galpeter · 2017-02-03T18:59:13Z

jerry-debugger/jerry-client.html

+
+  socket.onmessage = function(event)
+  {
+    message = new Uint8Array(event.data);


Shouldn't this be var message =...?

galpeter · 2017-02-03T18:59:26Z

jerry-debugger/jerry-client.html

+
+    var length = getFormatSize(format);
+
+    message = new Uint8Array(length);


Shouldn't this be var message =...?

galpeter · 2017-02-03T19:03:10Z

jerry-debugger/jerry-client.html

+    return true;
+  }
+
+  if (args[1] == "b" || args[1] == "break")


Is it possible to convert these if blocks converted into a switch-case?

galpeter · 2017-02-03T19:05:43Z

jerry-debugger/jerry-client.html

+      {
+        var breakpoint = decodeMessage("CI", message, 1);
+
+        breakpoint = functions[breakpoint[0]].offsets[breakpoint[1]];


So... is this a breakpoint? an offset? or something else? It's a bit confusing.

galpeter · 2017-02-03T19:10:32Z

jerry-debugger/jerry-client.py

+# WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+# See the License for the specific language governing permissions and
+# limitations under the License.
+


This script will not run with python 3. We should use print functions (instead of print statement):

from __future__ import print_function

zherczeg · 2017-02-06T09:40:54Z

jerry-core/CMakeLists.txt

@@ -162,6 +170,11 @@ if(FEATURE_MEM_STATS)
  set(DEFINES_JERRY ${DEFINES_JERRY} JMEM_STATS)
 endif()

+# Enable debugger
+if(FEATURE_DEBUGGER)
+  set(DEFINES_JERRY ${DEFINES_JERRY} JERRY_DEBUGGER)


Currently we requires standard libc. Perhaps we should also add that feature.

I added a check for it, if we build with python and use jerry-libc, the builder throw an error.

LaszloLango · 2017-02-06T10:07:12Z

jerry-core/debugger/jerry-debugger.h

+ * check the maximum buffer size. It need to be between 64 and 256.
+ */
+#if JERRY_DEBUGGER_MAX_BUFFER_SIZE < 64 || JERRY_DEBUGGER_MAX_BUFFER_SIZE > 256
+#error "Please define the MAX_BUFFER_SIZE between 64 and 256."


Quote characters are not necessary

zherczeg · 2017-02-06T11:49:04Z

jerry-core/debugger/jerry-sha1.c

+ *
+ *  This file is part of mbed TLS (https://tls.mbed.org)
+ */
+/*


We could add a newline here.

zherczeg · 2017-02-07T04:46:29Z

Error reported by travis:
jerry-core/debugger/jerry-debugger-ws.c:371: style(variableScope): The scope of the variable 'request_buffer_size' can be reduced.

zherczeg · 2017-02-07T10:08:13Z

jerry-core/debugger/jerry-sha1.c

+
+/* 32-bit integer manipulation macros (big endian). */
+
+#define GET_UINT32_BE(n, b, i) \


We need to add a JERRY_SHA1_ prefix. JERRY_SHA1_GET_UINT32_BE

Okey, thanks for the review.

janjongboom · 2017-02-07T12:59:03Z

This is very cool of course. Love to implement this for the mbedos5 target too.

+1 on making this more agnostic of the underlying transport layer. I'd probably want to run this over serial.

Also, any numbers on size requirements when running with the debugger?

zherczeg · 2017-02-07T13:31:36Z

@janjongboom binary sizes in bytes on ARM-32

mini-shell, static, jerry-libc: 136128
default-shell, static, jerry-libc: 140192
default-shell, dynamic, standard-libc: 141440
default-shell, dynamic, standard-libc, debugger: 149712

The debugger requires roughly 8K code

zherczeg · 2017-02-07T13:34:01Z

Currenlty a transport layer needs to implement 4 functions:

bool jerry_debugger_accept_connection (void);
void jerry_debugger_close_connection (void);

bool jerry_debugger_send (size_t data_size);
bool jerry_debugger_receive (void);

I think it is easy to implement this over serial port. Please check jerry-debugger-ws.c and .h

LaszloLango

LGTM

galpeter · 2017-02-07T19:06:59Z

tools/build.py

@@ -88,6 +89,10 @@ def devhelp(help):
        parser.print_help()
        sys.exit(0)

+    if arguments.jerry_debugger == 'ON' and arguments.jerry_libc == 'ON':
+        print('JerryScript Debugger need the standard libc. Please use the \'--jerry-libc=off\' for it.')


No need to escape the option, simply do the following:

print(".... '--jerry-...' ....")

I would prefer to check this somewhere in the CMake source with an error message. I think it would be better, especially if you build with the cmake only and don't want to use the script.

Its true, thanks!

galpeter · 2017-02-07T19:08:43Z

jerry-debugger/jerry-client-ws.py

+        self.offsets = {}
+        self.first_line = -1
+
+        if len(lines) > 0:


simply use: if lines:

galpeter · 2017-02-07T19:14:08Z

jerry-debugger/jerry-client-ws.py

+            % (self.byte_code_cp, self.source, self.name))
+
+        comma_needed = False
+        for breakpoint in self.lines.values():


If I understand correctly you just simply join the breakpoints as strings with a separator together. If so this would be better imho:

breakpoints = ','.join([str(breakpoint) for breakpoint in self.lines.values()])

and then you can add the breakpoints into the format string above

Thanks, little more pythonic way 😺

galpeter · 2017-02-07T19:16:05Z

jerry-debugger/jerry-client-ws.py

+            self.offsets[offset] = breakpoint
+
+    def __repr__(self):
+        result = ('Function(byte_code_cp:0x%x, source:\'%s\', name:\'%s\', { '


No need to escape the '-s just use " at the start/end of the string.

Changed all of the ' to ", thanks.

galpeter · 2017-02-07T19:16:56Z

jerry-debugger/jerry-client-ws.py

+                return
+
+        message = pack(self.debugger.byte_order + 'BBIB' + self.debugger.idx_format,
+                       0x82,


What are these numbers?

..from the RFC 6455 standard.

and what do they mean? Could we have extracted as a global variable like the JERRY_ things?

galpeter · 2017-02-07T19:35:41Z

jerry-debugger/jerry-client-ws.html

+
+  if (args[1] == "help")
+  {
+    appendLog("Debugger commands:\n" +


Do we really need the help? Can't we just have it as a div somewhere on the page?

I think it feels better this way.

galpeter · 2017-02-07T19:37:24Z

jerry-debugger/jerry-client-ws.html

+    var length = getFormatSize(format);
+
+    var message = new Uint8Array(length);
+


Why do we need the empty lines between the var-s?

galpeter · 2017-02-07T19:38:26Z

jerry-debugger/jerry-client-ws.html

+      assert(i != -1);
+
+      array.splice(i, 1);
+


Why do we need sooo many empty lines?

galpeter · 2017-02-07T19:41:05Z

jerry-debugger/jerry-client-ws.html

+
+Getting help: type 'help' in the command line below.<br>
+
+<input id="command" type="text" onkeypress="return debuggerCommand(event)">


If we always return true on keypress then maybe something like this would be a bit better:
debuggerCommand(event); return true;

galpeter · 2017-02-07T19:41:38Z

jerry-debugger/jerry-client-ws.html

+
+<textarea id="log" rows="20" cols="80"></textarea><br>
+
+Getting help: type 'help' in the command line below.<br>


We could even have the help here :)

robertsipka · 2017-02-08T09:13:36Z

tools/build.py

@@ -56,6 +56,7 @@ def devhelp(help):
    parser.add_argument('-j', '--jobs', metavar='N', action='store', type=int, default=multiprocessing.cpu_count() + 1, help='Allowed N build jobs at once (default: %(default)s)')
    parser.add_argument('--jerry-cmdline', metavar='X', choices=['ON', 'OFF'], default='ON', type=str.upper, help='build jerry command line tool (%(choices)s; default: %(default)s)')
    parser.add_argument('--jerry-cmdline-minimal', metavar='X', choices=['ON', 'OFF'], default='OFF', type=str.upper, help='build minimal version of the jerry command line tool (%(choices)s; default: %(default)s)')
+    parser.add_argument('--jerry-debugger', metavar='X', choices=['ON', 'OFF'], default='OFF', type=str.upper, help='enalbe the jerry debugger (%(choices)s; default: %(default)s)')


enable the jerry debugger

Thanks Robert!

robertsipka · 2017-02-08T09:32:18Z

tools/build.py

@@ -88,6 +89,10 @@ def devhelp(help):
        parser.print_help()
        sys.exit(0)

+    if arguments.jerry_debugger == 'ON' and arguments.jerry_libc == 'ON':
+        print('JerryScript Debugger need the standard libc. Please use the \'--jerry-libc=off\' for it.')


I would prefer to check this somewhere in the CMake source with an error message. I think it would be better, especially if you build with the cmake only and don't want to use the script.

I am also an author.

dbatyai · 2017-02-08T10:29:51Z

docs/07.DEBUGGER.md

+this information instead of JerryScript. Hence debug information
+is discarded after it is transmitted to the client.
+
+The following argument makes JerryScript to wait for a client


"... makes JerryScript wait ...", without to

dbatyai · 2017-02-08T10:51:19Z

docs/07.DEBUGGER.md

+
+`--log-level 2`
+
+The HTML client can connect the IP address of the server by


"... connect to the IP address of the server with the ..."

dbatyai · 2017-02-08T10:57:28Z

docs/07.DEBUGGER.md

+
+The HTML client can connect the IP address of the server by
+the `connect IP-address` command. The IP address can be
+localhost if the server and client run on the same machine.


"... if the server and the client are running on the same machine."

dbatyai · 2017-02-08T11:15:58Z

jerry-core/debugger/jerry-debugger.c

+  type *name_p = ((type *) recv_buffer_p)
+
+/**
+ * Free all unreferenced byte code structures whose


which instead of whose

dbatyai · 2017-02-08T11:25:46Z

jerry-debugger/jerry-client-ws.py

+def arguments_parse():
+    parser = argparse.ArgumentParser(description='JerryScript debugger client')
+
+    parser.add_argument('address', action='store', nargs='?', default='localhost:5001', help='specified a custom network setup (default: %(default)s)')


Could we improve this help message?

zherczeg · 2017-02-14T08:16:53Z

tools/runners/run-debugger-test.sh

+fi
+
+echo "${TEST_CASE} passed"
+exit 0


Missing newline

zherczeg · 2017-02-14T08:18:28Z

tools/runners/run-debugger-test.sh

+echo "$START_DEBUG_SERVER"
+eval "$START_DEBUG_SERVER"
+
+RESULT=$( (cat "${TEST_CASE}.cmd" | ${DEBUGGER_CLIENT}) 2>&1)


No space after $( or space after 2>&1

zherczeg · 2017-02-14T08:19:42Z

tools/runners/run-debugger-test.sh

+DEBUGGER_CLIENT=$2
+TEST_CASE=$3
+
+START_DEBUG_SERVER="${JERRY} ${TEST_CASE}.js --start-debug-server &"


wait 1 second

zherczeg · 2017-02-14T08:19:57Z

tests/debugger/do_step.js

+}
+
+f1();
+f2();


Newline missing

zherczeg · 2017-02-14T08:20:16Z

tests/debugger/do_step.cmd

+c
+s
+bt
+c


Newline missing

zherczeg · 2017-02-14T08:20:44Z

tests/debugger/do_dump.js

+}
+
+print ("var apple");
+var apple = 'apple';


no apple :)

tilmannOSG

Thanks for working on this! This is a great step forward for JerryScript in terms of usability.
Overall the patch looks very good, the bulk of my comments are just minor text issues.

tilmannOSG · 2017-02-14T09:18:10Z

docs/02.API-REFERENCE.md

+
+**Summary**
+
+Parse script and construct an EcmaScript function. The lexical


EcmaScript -> ECMAScript

tilmannOSG · 2017-02-14T09:18:42Z

docs/02.API-REFERENCE.md

+(usually a file name) is also passed to this function which is
+used by the debugger to find the source code.
+
+*Note*: Returned value must be freed with [jerry_release_value](#jerry_release_value) when it


Returned -> The returned

tilmannOSG · 2017-02-14T09:19:27Z

docs/02.API-REFERENCE.md

+
+```c
+jerry_value_t
+jerry_parse_named_resource (const jerry_char_t *name_p, /**<  */


Missing description of name_p?

tilmannOSG · 2017-02-14T09:20:01Z

docs/02.API-REFERENCE.md

+
+- `name_p` - name, usually a file name
+- `name_length` - size of the file name, in bytes
+- `source_p` - string, containing source code to parse. It must be a valid utf8 string


utf8 -> UTF-8

tilmannOSG · 2017-02-14T09:27:13Z

docs/02.API-REFERENCE.md

+  - function object value, if script was parsed successfully,
+  - thrown error, otherwise
+
+This function is the same as [jerry_parse](#jerry_parse), except a filename parameter is added.


This function is identical to jerry_parse, except that an additional filename parameter has been added.

tilmannOSG · 2017-02-14T10:36:17Z

jerry-debugger/jerry-client-ws.py

+            logging.debug("Little endian machine")
+        else:
+            self.byte_order = ">"
+            logging.debug("Big endian machine")


Big endian -> Big-endian

tilmannOSG · 2017-02-14T10:37:28Z

jerry-debugger/jerry-client-ws.py

+
+        data = debugger.get_message()
+
+    # Copy the ready list to the global storage


Please terminate comment with a period.

tilmannOSG · 2017-02-14T10:38:04Z

jerry-debugger/jerry-client-ws.py

+
+        data = debugger.get_message()
+
+        if not data:  # Break the while loop if there is no more data


Please terminate comment with a period.

tilmannOSG · 2017-02-14T10:39:41Z

tests/debugger/do_step.cmd

@@ -0,0 +1,8 @@
+break do_step.js:25


is the ^M intentional?

tilmannOSG · 2017-02-14T10:40:05Z

tests/debugger/do_step.cmd

+break do_step.js:25
+step
+backtrace
+b do_step.js:26


is the ^M intentional?

polaroi8d · 2017-02-14T11:47:43Z

@tilmannOSG updated the PR following your comments. Thanks!

tilmannOSG · 2017-02-14T13:27:29Z

@polaroi8d Thank you!
It seems one of the debugger tests failed in the Travis CI build, can you have a look? Once the tests are fixed I think this should be ready to go in :)

The debugger supports setting breakpoints, execution control (step, next, continue) and getting backtrace. The communication is websocket based, so a browser can communicate with JerryScript without any intermediate application. JerryScript-DCO-1.0-Signed-off-by: Zoltan Herczeg [email protected] JerryScript-DCO-1.0-Signed-off-by: Levente Orban [email protected]

LaszloLango added the feature request Requested feature label Feb 1, 2017

zherczeg previously requested changes Feb 1, 2017

View reviewed changes

LaszloLango requested changes Feb 1, 2017

View reviewed changes

dbatyai reviewed Feb 1, 2017

View reviewed changes

LaszloLango requested changes Feb 1, 2017

View reviewed changes

galpeter reviewed Feb 3, 2017

View reviewed changes

zherczeg reviewed Feb 6, 2017

View reviewed changes

LaszloLango reviewed Feb 6, 2017

View reviewed changes

zherczeg reviewed Feb 6, 2017

View reviewed changes

polaroi8d force-pushed the remote-debugger branch 2 times, most recently from 5393ff3 to bb137a8 Compare February 6, 2017 15:29

polaroi8d force-pushed the remote-debugger branch from 800833a to 69f52e2 Compare February 7, 2017 09:57

zherczeg reviewed Feb 7, 2017

View reviewed changes

polaroi8d force-pushed the remote-debugger branch from 69f52e2 to 8fbf7c2 Compare February 7, 2017 12:06

LaszloLango approved these changes Feb 7, 2017

View reviewed changes

galpeter reviewed Feb 7, 2017

View reviewed changes

robertsipka reviewed Feb 8, 2017

View reviewed changes

dbatyai reviewed Feb 8, 2017

View reviewed changes

polaroi8d force-pushed the remote-debugger branch 2 times, most recently from 1ef86fb to 8dc2010 Compare February 8, 2017 17:35

zherczeg mentioned this pull request Feb 10, 2017

Improve error message #1577

Merged

polaroi8d force-pushed the remote-debugger branch from 8dc2010 to 0e3278e Compare February 14, 2017 08:12

zherczeg reviewed Feb 14, 2017

View reviewed changes

polaroi8d force-pushed the remote-debugger branch 2 times, most recently from e437766 to 6209482 Compare February 14, 2017 09:08

tilmannOSG requested changes Feb 14, 2017

View reviewed changes

polaroi8d force-pushed the remote-debugger branch 2 times, most recently from f1075ca to 69b9894 Compare February 14, 2017 11:45

tilmannOSG approved these changes Feb 14, 2017

View reviewed changes

polaroi8d force-pushed the remote-debugger branch from 69b9894 to 3595892 Compare February 14, 2017 13:34

tilmannOSG merged commit 025a99c into jerryscript-project:master Feb 14, 2017

martijnthe mentioned this pull request Mar 26, 2018

Debugger Transport APIs and Zephyr port #2247

Closed

zherczeg mentioned this pull request Mar 27, 2018

Nominating Robert Sipka (robertsipka) to become a reviewer #2254

Closed

		@@ -335,6 +338,7 @@ jerry_value_t jerry_exec_snapshot (const void *snapshot_p, size_t snapshot_size,
		size_t jerry_parse_and_save_literals (const jerry_char_t *source_p, size_t source_size, bool is_strict,
		uint8_t *buffer_p, size_t buffer_size, bool is_c_format);


		#define SHIFT(x, n) ((x << n) \| ((x & 0xFFFFFFFF) >> (32 - n)))

		#define R(t) \


		var length = getFormatSize(format);

		message = new Uint8Array(length);


		/* 32-bit integer manipulation macros (big endian). */

		#define GET_UINT32_BE(n, b, i) \

		var length = getFormatSize(format);

		var message = new Uint8Array(length);


		Getting help: type 'help' in the command line below.<br>

		<input id="command" type="text" onkeypress="return debuggerCommand(event)">


		<textarea id="log" rows="20" cols="80"></textarea><br>

		Getting help: type 'help' in the command line below.<br>


		`--log-level 2`

		The HTML client can connect the IP address of the server by

+              }
+              f1();
+              f2();


		Summary

		Parse script and construct an EcmaScript function. The lexical


		data = debugger.get_message()

		# Copy the ready list to the global storage


		data = debugger.get_message()

		if not data: # Break the while loop if there is no more data

Initial version of JerryScript debugger #1557

Initial version of JerryScript debugger #1557

Uh oh!

Conversation

polaroi8d commented Feb 1, 2017

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Choose a reason for hiding this comment

Uh oh!

jiangzidong commented Feb 3, 2017

Uh oh!

zherczeg commented Feb 3, 2017

Uh oh!

martijnthe commented Feb 3, 2017

Uh oh!

zherczeg commented Feb 3, 2017

Uh oh!