Edit report at https://bugs.php.net/bug.php?id=66242&edit=1

 ID:                 66242
 Comment by:         ogatak at jp dot ibm dot com
 Reported by:        thuhc at yahoo dot com
 Summary:            CLI server failure on 64-bit ARM (Aarch64)
 Status:             Assigned
 Type:               Bug
 Package:            Testing related
 Operating System:   Linux AARCH64
 PHP Version:        5.5.6
 Assigned To:        ardbiesheuvel
 Block user comment: N
 Private report:     N

 New Comment:

Is there any progress on this issue?

I got the same failure in PHP 5.6.0 on Linux PPC64 (RHEL 6.4).

I also verified this failure can be solved by adding "-fsigned-char" GCC option 
in CFLAGS.  Could you modify configure.in to include this option?


Previous Comments:
------------------------------------------------------------------------
[2013-12-20 05:11:36] thuhc at yahoo dot com

Hi ardbiesheuvel,

About mimetype issue on test.mp3, I found an issue in 
ext/fileinfo/libmagic/apprentice.c
protected uint64_t file_signextend(struct magic_set *ms, struct magic *m, 
uint64_t v)
...
             case FILE_BYTE:
                        v = (char) v;
                        break;
...
char is unsigned char by default on ARM (on Intel char is signed char by 
default) --> invalid mimetype with test.mp3
If I changed code: v = (signed char) v --> it's ok.

------------------------------------------------------------------------
[2013-12-12 08:47:24] ardbiesheu...@php.net

The mimetype issue you are seeing is not caused by PHP but by the underlying 
OpenEmbedded installation that you are using.

I have built and tested PHP/CLI on Ubuntu Saucy under a QEMU/arm64 chroot, and 
I get the following failures (out of 7051 executed):

function with many parameters [tests/func/010.phpt]
-> Times out, due to execution overhead of the QEMU emulator

Check cli_process_title support on Unix 
[sapi/cli/tests/cli_process_title_unix.phpt]
-> Not supported under QEMU

file upload greater than 2G [sapi/cli/tests/upload_2G.phpt]
-> Fails on out of memory condition (on my 8GB machine)

RFC: DateTime and Daylight Saving Time Transitions (zone type 3, ba) 
[ext/date/tests/rfc-datetime_and_daylight_saving_time-type3-ba.phpt]
RFC: DateTime and Daylight Saving Time Transitions (zone type 3, bs) 
[ext/date/tests/rfc-datetime_and_daylight_saving_time-type3-bs.phpt]
-> Not quite clear to me, but reproducible also on my x86 Ubuntu Saucy 
installation

So as far as I can tell, there are no arm64 specific issues remaining in this 
report that are caused by bugs in PHP.

Thanks for the report.

------------------------------------------------------------------------
[2013-12-12 02:37:33] thuhc at yahoo dot com

================================================================================
/home/root/php-src/ext/fileinfo/tests/finfo_file_002.phpt
================================================================================
array(8) {
  ["/home/root/php-src/ext/fileinfo/tests/resources/dir.zip"]=>
  string(15) "application/zip"
  ["/home/root/php-src/ext/fileinfo/tests/resources/test.awk"]=>
  string(10) "text/plain"
  ["/home/root/php-src/ext/fileinfo/tests/resources/test.bmp"]=>
  string(14) "image/x-ms-bmp"
  ["/home/root/php-src/ext/fileinfo/tests/resources/test.gif"]=>
  string(9) "image/gif"
  ["/home/root/php-src/ext/fileinfo/tests/resources/test.jpg"]=>
  string(10) "image/jpeg"
  ["/home/root/php-src/ext/fileinfo/tests/resources/test.mp3"]=>
  string(24) "application/octet-stream"
  ["/home/root/php-src/ext/fileinfo/tests/resources/test.pdf"]=>
  string(15) "application/pdf"
  ["/home/root/php-src/ext/fileinfo/tests/resources/test.png"]=>
  string(9) "image/png"
}
================================================================================
013+   string(24) "application/octet-stream"
013-   string(10) "audio/mpeg"
================================================================================

================================================================================
/home/root/php-src/ext/standard/tests/array/bug65304.phpt
================================================================================
int(-9223372036854775808)
float(9.2233720368548E+18)
================================================================================
001+ int(-9223372036854775808)
002- float(%s)
================================================================================

------------------------------------------------------------------------
[2013-12-10 11:15:24] ardbiesheu...@php.net

OK, that looks a whole lot better. I have pushed another fix to address the 
#65304 regression (Zend: fix overflow handling bug in non-x86 
fast_add_function), so the only remaining failure shared between the two runs 
whose results you have posted is ext/fileinfo/tests/finfo_file_002.phpt.

Any background on those? Care to share the .log?

------------------------------------------------------------------------
[2013-12-10 10:05:14] thuhc at yahoo dot com

I tested with  current source:
git clone http://git.php.net/repository/php-src.git

Bug #53879 (DateTime::createFromFormat() fails to parse cookie expiration date)
[ext/date/tests/bug53879.phpt]
RFC: DateTime and Daylight Saving Time Transitions (zone type 3, ba) [ext/date/t
ests/rfc-datetime_and_daylight_saving_time-type3-ba.phpt]
RFC: DateTime and Daylight Saving Time Transitions (zone type 3, bs) [ext/date/t
ests/rfc-datetime_and_daylight_saving_time-type3-bs.phpt]
finfo_file(): Testing mime types [ext/fileinfo/tests/finfo_file_002.phpt]
Bug #65304 (Use of max int in array_sum) [ext/standard/tests/array/bug65304.phpt
]
Test strncmp() function : usage variations - binary safe(binary values) [ext/sta
ndard/tests/strings/strncmp_variation6.phpt]

------------------------------------------------------------------------


The remainder of the comments for this report are too long. To view
the rest of the comments, please view the bug report online at

    https://bugs.php.net/bug.php?id=66242


--
Edit this bug report at https://bugs.php.net/bug.php?id=66242&edit=1

-- 
PHP Quality Assurance Mailing List <http://www.php.net/>
To unsubscribe, visit: http://www.php.net/unsub.php

Reply via email to