Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Jump to bottom

Assertion failures in mruby-pack on big endian #4190

Closed
jeremyevans opened this issue Dec 17, 2018 · 4 comments
Closed

Assertion failures in mruby-pack on big endian #4190

jeremyevans opened this issue Dec 17, 2018 · 4 comments

Comments

@jeremyevans
Copy link

The good news is mruby 2.0.0 works on OpenBSD/sparc64 (previous versions failed to build). However, running the tests shows the following failures:

Fail: pack float (mrbgems: mruby-pack)
 - Assertion[1] Failed: Expected to be equal
    Expected: "\x00\x00@@"
      Actual: "@@\x00\x00"
 - Assertion[2] Failed: Expected to be equal
    Expected: [3]
      Actual: [2.304855714121459e-41]
 - Assertion[3] Failed: Expected to be equal
    Expected: "@@\x00\x00"
      Actual: "\x00\x00@@"
 - Assertion[4] Failed: Expected to be equal
    Expected: [3]
      Actual: [2.304855714121459e-41]
 - Assertion[5] Failed: Expected to be equal
    Expected: "@@\x00\x00"
      Actual: "\x00\x00@@"
 - Assertion[6] Failed: Expected to be equal
    Expected: [3]
      Actual: [2.304855714121459e-41]
 - Assertion[7] Failed: Expected to be equal
    Expected: "@@\x00\x00"
      Actual: "\x00\x00@@"
 - Assertion[8] Failed: Expected to be equal
    Expected: [3]
      Actual: [2.304855714121459e-41]
Fail: pack double (mrbgems: mruby-pack)
 - Assertion[1] Failed: Expected to be equal
    Expected: "\x00\x00\x00\x00\x00\x00\b@"
      Actual: "@\b\x00\x00\x00\x00\x00\x00"
 - Assertion[2] Failed: Expected to be equal
    Expected: [3]
      Actual: [1.043466644016713e-320]
 - Assertion[3] Failed: Expected to be equal
    Expected: "@\b\x00\x00\x00\x00\x00\x00"
      Actual: "\x00\x00\x00\x00\x00\x00\b@"
 - Assertion[4] Failed: Expected to be equal
    Expected: [3]
      Actual: [1.043466644016713e-320]
 - Assertion[5] Failed: Expected to be equal
    Expected: "@\b\x00\x00\x00\x00\x00\x00"
      Actual: "\x00\x00\x00\x00\x00\x00\b@"
 - Assertion[6] Failed: Expected to be equal
    Expected: [3]
      Actual: [1.043466644016713e-320]
 - Assertion[7] Failed: Expected to be equal
    Expected: "@\b\x00\x00\x00\x00\x00\x00"
      Actual: "\x00\x00\x00\x00\x00\x00\b@"
 - Assertion[8] Failed: Expected to be equal
    Expected: [3]
      Actual: [1.043466644016713e-320]

These failures do not occur on OpenBSD/amd64, and the test output leads me to believe this may be a general mruby-pack issue on big endian systems.
@matz
Copy link
Member

matz commented Dec 20, 2018

Do you have MRB_ENDIAN_BIG defined?

@jeremyevans
Copy link
Author

I'm guessing it isn't explicitly defined. I'm going to close this now, and if there are still assertion failures with it defined, I'll reopen.

@jeremyevans
Copy link
Author

I just wanted to confirm that defining MRB_ENDIAN_BIG did fix the problem. We ended up doing something similar to this so the same code works on both big endian and little endian:

 /* define on big endian machines; used by MRB_NAN_BOXING */
-//#define MRB_ENDIAN_BIG
+#include <endian.h>
+#if (BYTE_ORDER == BIG_ENDIAN)
+#define MRB_ENDIAN_BIG
+#endif

@matz
Copy link
Member

matz commented Dec 20, 2018

That would be a nice idea. I was working on a different solution that relies on the dynamic endian check.
But this one is better.

matz added a commit that referenced this issue Dec 20, 2018
@matz
mruby-pack should not rely on MRB_ENDIAN_BIG macro; fix #4190
2fdb309 
The `MRB_ENDIAN_BIG` macro is originally used for `NaN` boxing.
We cannot assume it is defined on every big endian platform (#4190
is the case). So instead of relying on untrusted `MRB_ENDIAN_BIG`, we
use `BYTE_ORDER` macro with a fallback function to check endian in
runtime.
matz added a commit that referenced this issue Dec 20, 2018
@matz
Define MRB_ENDIAN_BIG automatically; ref #4190
f553791 
You had to define this macro on big endian platforms, but it is very
error-prone. So define the macro automatically if possible.
matz added a commit that referenced this issue Dec 21, 2018
@matz
Fixed a bug on platforms without BYTE_ORDER; ref #4190
445c64d
matz added a commit that referenced this issue Dec 21, 2018
@matz
Simplify MRB_ENDIAN_BIG macro definition; ref #4190
ed2bc5d 
`cpp` does not raise error on undefined macro access in condition.
@mimaki mimaki mentioned this issue Apr 5, 2019
Closed
ksekimoto added a commit to ksekimoto/mruby that referenced this issue Jul 16, 2021
@ksekimoto
mruby-pack should not rely on MRB_ENDIAN_BIG macro; fix mruby#4190
24927a8 
The `MRB_ENDIAN_BIG` macro is originally used for `NaN` boxing.
We cannot assume it is defined on every big endian platform (mruby#4190
is the case). So instead of relying on untrusted `MRB_ENDIAN_BIG`, we
use `BYTE_ORDER` macro with a fallback function to check endian in
runtime.
ksekimoto added a commit to ksekimoto/mruby that referenced this issue Jul 16, 2021
@ksekimoto
Define MRB_ENDIAN_BIG automatically; ref mruby#4190
e5d304c 
You had to define this macro on big endian platforms, but it is very
error-prone. So define the macro automatically if possible.
ksekimoto added a commit to ksekimoto/mruby that referenced this issue Jul 16, 2021
@ksekimoto
Fixed a bug on platforms without BYTE_ORDER; ref mruby#4190
06cc5b9
ksekimoto added a commit to ksekimoto/mruby that referenced this issue Jul 16, 2021
@ksekimoto
Simplify MRB_ENDIAN_BIG macro definition; ref mruby#4190
ed9c520 
`cpp` does not raise error on undefined macro access in condition.
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Assignees
No one assigned
Labels
None yet
Projects
None yet
Milestone
No milestone
Development

No branches or pull requests
2 participants
@jeremyevans @matz