Re: [PHP-CVS] cvs: php-src(PHP_5_2) /ext/standard pack.c /ext/standard/tests/strings pack.phpt
On 04/13/2007 06:17 AM, Brian Shire wrote: FYI: Reverted my bad pack test changes, just verified that the latest gcc-4.2 branch on Darwin returns 0 rather than -1 for the following code: #include stdio.h int main() { double foo = (double)4294967296LL; printf(%d\n, (unsigned long)(foo)); } Thanks! -- Wbr, Antony Dovgal -- PHP CVS Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-CVS] cvs: php-src(PHP_5_2) /ext/standard pack.c /ext/standard/tests/strings pack.phpt
FYI: Reverted my bad pack test changes, just verified that the latest gcc-4.2 branch on Darwin returns 0 rather than -1 for the following code: #include stdio.h int main() { double foo = (double)4294967296LL; printf(%d\n, (unsigned long)(foo)); } - Shire [EMAIL PROTECTED] [EMAIL PROTECTED] -- PHP CVS Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-CVS] cvs: php-src(PHP_5_2) /ext/standard pack.c /ext/standard/tests/strings pack.phpt
On 04/03/2007 11:50 PM, Brian Shire wrote: shire Tue Apr 3 19:50:40 2007 UTC Modified files: (Branch: PHP_5_2) /php-src/ext/standard pack.c /php-src/ext/standard/tests/strings pack.phpt Log: correction for previous fix to bug #38770 (pack/unpack is broken on 64bit) fix pack test, correct space to tabs Brian, this breaks the test on Linux/FreeBSD i386. Though somehow it's working on Mac/PPC. diff -u php-src/ext/standard/tests/strings/pack.phpt:1.1.2.2 php-src/ext/standard/tests/strings/pack.phpt:1.1.2.3 --- php-src/ext/standard/tests/strings/pack.phpt:1.1.2.2Sat Dec 30 02:00:51 2006 +++ php-src/ext/standard/tests/strings/pack.phptTue Apr 3 19:50:40 2007 @@ -145,7 +145,7 @@ ) Array ( -[1] = 0 +[1] = -1 ) Array ( @@ -185,7 +185,7 @@ ) Array ( -[1] = 0 +[1] = -1 ) Array ( @@ -233,7 +233,7 @@ ) Array ( -[1] = 0 +[1] = -1 ) Array ( @@ -305,7 +305,7 @@ ) Array ( -[1] = 0 +[1] = -1 ) Array ( -- Wbr, Antony Dovgal -- PHP CVS Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-CVS] cvs: php-src(PHP_5_2) /ext/standard pack.c /ext/standard/tests/strings pack.phpt
Yeah, discussed briefly with Jani and Ilia. I can revert it (or temporarily remove those lines of the test) until I can figure out why this doesn't work on Linux if you like? On Darwin the code intval (4294967296) will return -1, while on other systems it appears to be 0. I can reproduce this on the C level by replicating the double- long cast that's happening within PHP. Any suggestions on this would be welcome. Antony, you said below that you tested on PPC? Seems like this is a Darwin specific problem then (I tested on Darwin/Intel), perhaps I should just remove those tests for Darwin for the time being? Thanks, -shire On Apr 10, 2007, at 2:34 PM, Antony Dovgal wrote: On 04/03/2007 11:50 PM, Brian Shire wrote: shire Tue Apr 3 19:50:40 2007 UTC Modified files: (Branch: PHP_5_2) /php-src/ext/standard pack.c /php-src/ext/standard/tests/ strings pack.phpt Log: correction for previous fix to bug #38770 (pack/unpack is broken on 64bit) fix pack test, correct space to tabs Brian, this breaks the test on Linux/FreeBSD i386. Though somehow it's working on Mac/PPC. diff -u php-src/ext/standard/tests/strings/pack.phpt:1.1.2.2 php- src/ext/standard/tests/strings/pack.phpt:1.1.2.3 --- php-src/ext/standard/tests/strings/pack.phpt:1.1.2.2 Sat Dec 30 02:00:51 2006 +++ php-src/ext/standard/tests/strings/pack.phpt Tue Apr 3 19:50:40 2007 @@ -145,7 +145,7 @@ ) Array ( -[1] = 0 +[1] = -1 ) Array ( @@ -185,7 +185,7 @@ ) Array ( -[1] = 0 +[1] = -1 ) Array ( @@ -233,7 +233,7 @@ ) Array ( -[1] = 0 +[1] = -1 ) Array ( @@ -305,7 +305,7 @@ ) Array ( -[1] = 0 +[1] = -1 ) Array ( -- Wbr, Antony Dovgal -- PHP CVS Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-CVS] cvs: php-src(PHP_5_2) /ext/standard pack.c /ext/standard/tests/strings pack.phpt
On 04/11/2007 02:45 AM, Brian Shire wrote: Yeah, discussed briefly with Jani and Ilia. I can revert it (or temporarily remove those lines of the test) until I can figure out why this doesn't work on Linux if you like? On Darwin the code intval (4294967296) will return -1, while on other systems it appears to be 0. I can reproduce this on the C level by replicating the double- long cast that's happening within PHP. Any suggestions on this would be welcome. Well, if it's just this Darwin/Intel difference then it's not that critical, I believe. Antony, you said below that you tested on PPC? Seems like this is a Darwin specific problem then (I tested on Darwin/Intel), perhaps I should just remove those tests for Darwin for the time being? Let me see if I can reproduce it on Darwin/Intel, that'll take some time.. -- Wbr, Antony Dovgal -- PHP CVS Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php
Re: [PHP-CVS] cvs: php-src(PHP_5_2) /ext/standard pack.c /ext/standard/tests/strings pack.phpt
On Apr 10, 2007, at 3:56 PM, Antony Dovgal wrote: On 04/11/2007 02:45 AM, Brian Shire wrote: Yeah, discussed briefly with Jani and Ilia. I can revert it (or temporarily remove those lines of the test) until I can figure out why this doesn't work on Linux if you like? On Darwin the code intval (4294967296) will return -1, while on other systems it appears to be 0. I can reproduce this on the C level by replicating the double- long cast that's happening within PHP. Any suggestions on this would be welcome. Well, if it's just this Darwin/Intel difference then it's not that critical, I believe. Antony, you said below that you tested on PPC? Seems like this is a Darwin specific problem then (I tested on Darwin/Intel), perhaps I should just remove those tests for Darwin for the time being? Let me see if I can reproduce it on Darwin/Intel, that'll take some time.. Thanks, I can reproduce and so can Ilia on Darwin/Intel so if you're getting a -1 on Darwin/PPC too then it tells me it's probably some silly Darwin difference/bug rather than a 64-bit extension problem on the Intel 2 Duo chips. If someone has Linux/non-Darwin OS on an Intel 2 Duo chip that would be informative as well. - Shire [EMAIL PROTECTED] [EMAIL PROTECTED] -- PHP CVS Mailing List (http://www.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php