Fix "unhandled exception occurs when a private variable increments"

[hvwyl]

Fix "unhandled exception occurs when a private variable increments" #5238

Added handling for the condition of private variable incr/decr in js-parser-expr.c and vm.c.

JerryScript-DCO-1.0-Signed-off-by: hvwyl [email protected]

[zherczeg]

Thank you! Could you add a test case?

[hvwyl]

Thank you! Could you add a test case?

ok, i have added the test case for issue #5238

[zherczeg]

LGTM

[zherczeg]

Add a newline at the end.

[hvwyl]

LGTM

a newline at the end of file private_fields.js has been added. and it can be merged directly.

[zherczeg]

Somehow these tests are crashing, e.g. FAIL (-6): tests/jerry/private_fields.js

[hvwyl]

Somehow these tests are crashing, e.g. FAIL (-6): tests/jerry/private_fields.js

I checked the latest commit of branch. Files private_field.js and regression-test-issue-5238.js are ok.

I noticed that the crushing dosen't occur in release version, and it only occurs in debug version.

Then I compiled the debug version in vm.

For input class A { #a = 0; incr() { this.#a++ } }, it raises

ICE: Assertion 'context_p->stack_depth == 0' failed at /home/runner/work/jerryscript/jerryscript/jerry-core/parser/js/js-parser.c(parser_post_processing):635.
Error: JERRY_FATAL_FAILED_ASSERTION

The real reason is that context_p->stack_depth will be decremented by 1 after context_p->last_cbc_opcode = PARSER_PUSH_PROP_LITERAL_TO_PUSH_LITERAL (context_p->last_cbc_opcode); (located in js-parser-expr.c)

But context_p->stack_depth will remain the same after context_p->last_cbc_opcode = PARSER_TO_EXT_OPCODE (CBC_EXT_PUSH_PRIVATE_PROP_LITERAL_REFERENCE); (located in js-parser-expr.c).

The way to fix: Manually adjusting stack usage for private-field condition.

However, for release version, assert will be disabled. So I didn't notice this crushing.

In addition, I'm using a nickname on github. So I updated the comment in order to fit signed-off checking. Should I resubmit the pr?

[zherczeg]

There is no need for a second pr.

[hvwyl]

There is no need for a second pr.

I noticed the Checks in Action #4395, it throw

Signed-off-by message is incorrect. The following line should be at the end of the 71a898cb commit's message: 'JerryScript-DCO-1.0-Signed-off-by: hvwyl [email protected]'. 

although I have updated the comment 2 days ago. I don't know if this can be ignored.

Anyway, Win_x86-64_Tests-MINGW in Action #4395. There's also problem in ecma-builtin-helpers-date.c.

[hvwyl]

I've organized the commits log and fixed all minor issues. Here's the latest GitHub Actions run in my fork:
https://github.com/hvwyl/jerryscript/actions/runs/16346678103
It can be merged this time.

[hvwyl]

@zherczeg Hi, I've fixed all minor issues. Could you please take another look when you have a moment? Thanks!

[hvwyl]

@zherczeg Hello, I have added the sign-off line 'DCO-1.0-Signed-off-by' in my commits; maybe a new GitHub Action is needed. I am currently still a student. Owing to lack of project experience, I have made quite a few detours. I am really very sorry if I have disturbed you. Thanks for your support and understanding.