-
Notifications
You must be signed in to change notification settings - Fork 394
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.
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
Fix function pixReadMemBmp for big endian hosts #752
Conversation
Simplify also the code a little bit and remove a comment with test code. Signed-off-by: Stefan Weil <[email protected]>
I still get a failure for
But those issues can be addressed in another pull request. |
Thank you for simplifying & fixing it. |
Failure on SPARC64 / Solaris:
|
I do not know why it didn't find a path. Question: could this be an endian issue, or is it specific to sparc64? |
No, Apple is little endian. But I have some more big endian machines (PowerPC, MIPS64) and can try those. |
MIPS64 fails for a different reason (not only maze_reg, but many other tests, too, require bmp format):
|
* The written bmp files are truncated on big-endian MIPS64.
I don't know what causes the truncation in the written bmp file. |
|
I don't think that the error is an endianess issue. The error seems to be caused by a full filesystem for /tmp which can only store 100 MB on the test machine. The Leptonica tests consume this little free space very fast. |
All tests pass on a big endian PowerPC64 (tests which require gnuplot were skipped):
|
Thank you for doing all the tests, Stefan. So it appears that we are good on all platforms, except for failure to find a 1 bpp path in maze_reg on sparc64/solaris. And whatever causes that failure, it is not endian-related. |
Simplify also the code a little bit and remove a comment with test code.
convertOnBigEnd32
must not be called forihbytes
because it is already assigned the right value which is read frombmpih
in little endian order.