Newsgroups: php.internals
Path: news.php.net
Xref: news.php.net php.internals:23327
Mailing-List: contact internals-help@lists.php.net; run by ezmlm
Date: Sat, 13 May 2006 02:44:47 +0200
Reply-To: Marcus Boerger <helly@php.net>
Message-ID: <246971759.20060513024447@marcus-boerger.de>
To: Ron Korving <r.korving@xit.nl>
Cc: internals@lists.php.net
In-Reply-To: <9C.1A.19568.E0E64644@pb1.pair.com>
References: <321302981.20060512024659@marcus-boerger.de> <9C.1A.19568.E0E64644@pb1.pair.com>
MIME-Version: 1.0
Content-Type: text/plain; charset=us-ascii
Content-Transfer-Encoding: 7bit
Subject: Re: [PHP-DEV] Re: 5.2 interruptions
From: helly@php.net (Marcus Boerger)

Hello Ron,

 if not i am still missing something, actuall i *fear* that there are places
in the engine that do not adhere to the new semantics. If you spot one
please inform me, prefereable by writing a test case. If you don't feel able
too just let me know what you think is wrong.

best regards
marcus

Friday, May 12, 2006, 1:12:44 PM, you wrote:

> Marcus,

> Zend: __toString() Done (marcus)

> Does this mean that finally __toString acts like the whole world would want 
> it to? Can it finally be true? :)

> Ron


> "Marcus Boerger" <helly@php.net> schreef in bericht 
> news:321302981.20060512024659@marcus-boerger.de...
>> Hello internals,
>>
>>  i am now done with most bigger parts of 5.2 Todo and the remainder
>> is a bunch of fixes engine wise some folks already found during the
>> MFH phase. Right now my test results look like (see below). Whilethis
>> looks quite good on the first glance it is actually a big problem that
>> only have around 2500 tests for a beast as big as PHP.
>>
>>  A thing i am personally wondering about is that som many people
>> complained about interruptions on this development tree. Actually i
>> tried all i could to keep this interruptions as low as possible. For
>> example besides windows all built-in stuff was at least always
>> compilable. The biggest problems extension wise where two things.
>> First finding all cases of overloaded iterators since their API has
>> chnaged. This is a problem since the function pointers hide behind
>> tables and conversons, so i could not grep for them. The other major
>> issue was that for PDO we planned to MFH a bunch of stuff from HEAD
>> while unfortunatley PDO's HEAD is completley out of synch with separate
>> development in 5.1/5.2. Thus taking care of PDO had to wait until
>> the rest had settled down. Now the real question is if ppl see the
>> 5.2 already as something that goes for production quality or if
>> suddenly windows became a major developmment platform without me
>> noticing it. In the first case it is my bad, because i got our time
>> line wrong. In the second case it is still my bad, but just tell me
>> now so i can kick my windows compiler next tiem i do bigger changes.
>> Anyways sorry if you feel the interruptions have been too annoying.
>>
>> Back on the first topic: tests. While i was still happily MFHing
>> people complained about a newly detected behavior change that nobody
>> detected in HEAD so far, see Zend/tests/objects_001.phpt. I never
>> intended to change that behavior. It was simply a sideeffect of MFHing
>> new stuff and obviously we have to fix the behavior in both HEAD
>> and 5.2. Anyways i ask everybody to write more tests so that we can
>> figure out such things earlier in the future, hopefully already before
>> committing since everybody should do make test before commit anyways.
>>
>> Last but not least i ask to provide tests for stuff we do not cover
>> right now. In case you are an extension developer you can most easily
>> go to http://gcov.php.net and check your extensions for not executed
>> c code. As the developer you will know how to provide a test for the
>> identified areas.
>>
>> One good idea came up already. That is we could extend the run-tests
>> facility to handle multiple php versions per .phpt script. That way
>> we can keep track of behavior changes easily. This would for example
>> be done by extending the current block identifier by a version compare
>> string. E.g. right now expected output is identified by "--EXPECT--".
>> With the version identifier it would look like: "--[>=5.2]:EXPECT--"
>> and "--[<5.2]:EXPECT--".
>>
>> Once again sorry if you felt the interruptions too annoying.
>>
>> best regards
>> marcus
>>
>> p.s.: todo can be found here: http://oss.backendmedia.com/PhP52
>>
>>
>> TEST RESULT SUMMARY
>> ---------------------------------------------------------------------
>> Exts skipped    :   21
>> Exts tested     :   68
>> ---------------------------------------------------------------------
>>
>> Number of tests : 2591              2251
>> Tests skipped   :  340 ( 13.1%) --------
>> Tests warned    :    0 (  0.0%) (  0.0%)
>> Tests failed    :   19 (  0.7%) (  0.8%)
>> Tests passed    : 2232 ( 86.1%) ( 99.2%)
>> ---------------------------------------------------------------------
>> Time taken      :  569 seconds
>> =====================================================================
>>
>> =====================================================================
>> FAILED TEST SUMMARY
>> ---------------------------------------------------------------------
>> comparing objects to other types [Zend/tests/objects_001.phpt]
>> APC: apc_store/fetch with strings [ext/apc/tests/apc_001.phpt]
>> APC: apc_store/fetch with objects [ext/apc/tests/apc_002.phpt]
>> APC: apc_store/fetch with objects [ext/apc/tests/apc_003.phpt]
>> Test 1: Accessing single node [ext/dom/tests/dom001.phpt]
>> Test 2: getElementsByTagName() / getElementsByTagNameNS() 
>> [ext/dom/tests/dom002.phpt]
>> Test 7: DTD tests [ext/dom/tests/dom007.phpt]
>> iconv stream filter [ext/iconv/tests/iconv_stream_filter.phpt]
>> mb_strlen() [ext/mbstring/tests/mb_strlen.phpt]
>> mysqli fetch float values [ext/mysqli/tests/010.phpt]
>> mysqli fetch mixed values [ext/mysqli/tests/011.phpt]
>> mysqli fetch mixed values 2 [ext/mysqli/tests/012.phpt]
>> mysqli_bind_result (SHOW) [ext/mysqli/tests/045.phpt]
>> PDO PgSQL Large Objects [ext/pdo_pgsql/tests/large_objects.phpt]
>> Bug #31422 No Error-Logging on SoapServer-Side 
>> [ext/soap/tests/bugs/bug31422.phpt]
>> Test array_diff and array_diff_assoc behaviour 
>> [ext/standard/tests/array/007.phpt]
>> Test of the *intersect* bunch of functions (both assoc and non-assoc) 
>> [ext/standard/tests/array/array_intersect_1.phpt]
>> date_sunrise() and date_sunset() functions 
>> [ext/standard/tests/general_functions/sunfuncts.phpt]
>> Bug #22224 (implode changes object references in array) 
>> [ext/standard/tests/strings/bug22224.phpt]
>> =====================================================================
>>
>>
>>
>> Best regards,
>> Marcus 




Best regards,
 Marcus