Issues reported for version 3.2.3 in Redmine that were not fixed in version 3.3.3 #547

enov · 2014-09-17T10:38:10Z

This is a placeholder issue for all Issues reported for version 3.2.3 in Redmine that were not fixed in version 3.3.3.

A new issue will be created and linked here, for each item below, with its title and description from Redmine.

The letter R is prefixed to the Redmine's issue number to avoid confusion between the issues of Github.

If you want to take an issue, drop me a line here, so that I assign it to you.

Hopefully we get through this and have a good 3.3.3 release :)

- [x] [R4791](http://dev.kohanaframework.org/issues/4791) Request_Client_Curl adds extra content-type header to all requests - [x] [R4778](http://dev.kohanaframework.org/issues/4778) PHPDoc comments - license location wrong - [x] [R4678](http://dev.kohanaframework.org/issues/4678) send_file() error on mime type - [x] [R4651](http://dev.kohanaframework.org/issues/4651) Request::uri() throws exception instead of returning a valid uri (docs obsolete) - [x] [R4627](http://dev.kohanaframework.org/issues/4627) shutdown_handler always uses 'Kohana_Exception::handler' instead of what's set in exception_handler() - [x] [R4625](http://dev.kohanaframework.org/issues/4625) Miscellaneous spelling errors - [x] [R4622](http://dev.kohanaframework.org/issues/4622) StdOut and StdErr Log Writers invalid log line format causes them not to output levels - [x] [R4613](http://dev.kohanaframework.org/issues/4613) Request class converts . to _ in url query string - [x] [R4607](http://dev.kohanaframework.org/issues/4607) Innacurate secure connection detection behind a load balancer - [x] [R4592](http://dev.kohanaframework.org/issues/4592) Response::send_file broken when no mime type specified - [x] [R4587](http://dev.kohanaframework.org/issues/4587) Request::redirect calls URL::site using the wrong parameters - [x] [R4568](http://dev.kohanaframework.org/issues/4568) DebugTest::test_dump fails because file has crlf line endings - [x] [R4522](http://dev.kohanaframework.org/issues/4522) Response does not respect Request's protocol - [x] [R4482](http://dev.kohanaframework.org/issues/4482) Array to string conversion in Arr::merge - [x] [R4287](http://dev.kohanaframework.org/issues/4287) Validation memory leak - [x] [R4241](http://dev.kohanaframework.org/issues/4241) It Must return original url with external link in request - [x] [R4235](http://dev.kohanaframework.org/issues/4235) Request::process_uri() must ignore external routes - [x] [R4201](http://dev.kohanaframework.org/issues/4201) Validation: field lable value - [x] [R4145](http://dev.kohanaframework.org/issues/4145) Translated messages should not rely on the default language being English - [x] [R4108](http://dev.kohanaframework.org/issues/4108) i18n __ function ignore $lang parameter - [x] [R4101](http://dev.kohanaframework.org/issues/4101) Cache, memcache, lifetime - [x] [R4079](http://dev.kohanaframework.org/issues/4079) Route::uri() should encode parameters - [x] [R3967](http://dev.kohanaframework.org/issues/3967) Kohana_RequestTest::test_url fails and Kohana_RequestTest::test_uri_only_trimed_on_internal fail for some custom routes - [x] [R3941](http://dev.kohanaframework.org/issues/3941) PHPDoc error for IntelliJ IDEA - [x] [R3896](http://dev.kohanaframework.org/issues/3896) Error message for 'password confirm' is incorrect in Auth module - [x] [R3788](http://dev.kohanaframework.org/issues/3788) Kohana 3.1 User model e-mail validation - [x] [R3604](http://dev.kohanaframework.org/issues/3604) Kohana_Session_Native should optionally Cookie Parameters when calling session_set_cookie_params - [x] [R3373](http://dev.kohanaframework.org/issues/3373) Kohana profiler Division by zero ErrorReflection [Warning] - [x] [R3931](http://dev.kohanaframework.org/issues/3931) Kohana_Exception::handler() doesn't exit with proper error codes - [x] [R4794](http://dev.kohanaframework.org/issues/4794) In online documentation, Error Handling page is throwing errors :) - [x] [R4275](http://dev.kohanaframework.org/issues/4275) kohanaframework.org/3.2 Guide Error - [x] [R4820](http://dev.kohanaframework.org/issues/4820) doc: incorrect mention of the missing method "Request::instance()" - [x] [R4814](http://dev.kohanaframework.org/issues/4814) Duplicate line of code wants to be removed - [x] [R4813](http://dev.kohanaframework.org/issues/4813) Kohana_UTF8Test::test_clean with data set number 3 array("\xFF", '') does not pass - [x] [R4706](http://dev.kohanaframework.org/issues/4706) Infinite loop in some of the assertion methods - [x] [R4322](http://dev.kohanaframework.org/issues/4322) More irregular pluralizations - [x] [R4092](http://dev.kohanaframework.org/issues/4092) Sub-requests don't inherit some properties properly from the parent - [x] [R3985](http://dev.kohanaframework.org/issues/3985) Kohana 3.1 User model e-mail validation - [x] [R3962](http://dev.kohanaframework.org/issues/3962) URL::query() should be consistent - [x] [R3946](http://dev.kohanaframework.org/issues/3946) Illegal offset type when call ArrayObject->offsetExists(arguments) - [x] [R3849](http://dev.kohanaframework.org/issues/3849) Failed 3 tests in Kohana_NumTest::test_format - [x] [R3716](http://dev.kohanaframework.org/issues/3716) Enforce uppercase constants and comparison operators - [x] [R3577](http://dev.kohanaframework.org/issues/3577) system/view/kohana/error.php should not be dependent on other classes - [x] [R4819](http://dev.kohanaframework.org/issues/4819) Replace preg_replace with /e modifier with preg_replace_callback - [x] [R4818](http://dev.kohanaframework.org/issues/4818) Add PHP 5.5 to Travis CI configuration - [x] [R4817](http://dev.kohanaframework.org/issues/4817) Add PHP 5.4 to Travis CI configuration - [x] [R4717](http://dev.kohanaframework.org/issues/4717) Why OAuth removed - [x] [R4676](http://dev.kohanaframework.org/issues/4676) Security::token() - [x] [R4675](http://dev.kohanaframework.org/issues/4675) Security::token() - [x] [R4623](http://dev.kohanaframework.org/issues/4623) Remove phpunit 3.6 and lower support - [x] [R4088](http://dev.kohanaframework.org/issues/4088) htaccess FastCgi - [x] [R4022](http://dev.kohanaframework.org/issues/4022) Arr::pluck() should use array_key_exists() - [x] [R4021](http://dev.kohanaframework.org/issues/4021) Arr::extract() should use array_key_exists() - [x] [R3973](http://dev.kohanaframework.org/issues/3973) Kohana_Cache_File Driver errors - [x] [R3925](http://dev.kohanaframework.org/issues/3925) Cookie::delete should uses Cookie::set - [x] [R3512](http://dev.kohanaframework.org/issues/3512) Conflicting standards dealing with whitespace - [x] [R3049](http://dev.kohanaframework.org/issues/3049) delete_all() - Error with 'file' driver - [x] [R3047](http://dev.kohanaframework.org/issues/3047) Cache error on __construct() - Linux machine - [x] [R4640](http://dev.kohanaframework.org/issues/4640) Template documentation broken and obsolete - [x] [R4624](http://dev.kohanaframework.org/issues/4624) Unsalted Cookies or Ingredients Docs - [x] [R4615](http://dev.kohanaframework.org/issues/4615) Arr::get should use both isset and array_key_exists - [x] [R4605](http://dev.kohanaframework.org/issues/4605) Unable to pass body params in external PUT requests (Windows) - [x] [R4601](http://dev.kohanaframework.org/issues/4601) Kohana::$magic_quotes always true on PHP 5.3.X

acoulton · 2014-09-17T11:29:19Z

@enov thanks for this. There's a couple that I fixed already but Redmine was down at the time so I couldn't comment and I don't have access to close.

R4817 (Add PHP5.4 to Travis) R4818 (Add PHP5.5 to Travis) are done - configured and passing for kohana/kohana and will be included as standard when I set up Travis for the other modules.
R4819 (preg_replace /e) is fixed, backported from 3.4 in 6b519fb
R4623 remove PHPUnit 3.6 support - is done AFAIK and closed on Redmine, we vendor 3.7 with composer.

I think there's a few others that show as closed on Redmine already - eg R3973 - I've not done a complete crosscheck.

I can try and take a couple next week, I have a few bits to get on with for the rest of this week.

acoulton · 2014-09-17T16:01:55Z

@enov while working on the build/test setup am seeing there's quite a few places where unit tests are skipped - in particular CookieTest per kohana/kohana#47 and a few others.

Should we (I) try to get them working again in the 3.3 series? It's likely to require a bit of refactoring of the classes themselves (as well as the tests) to make them more testable and the tests more useful, but I think that can be done without changing anything about the public API.

enov · 2014-09-17T20:31:53Z

Should we (I) try to get them working again in the 3.3 series? It's likely to require a bit of refactoring of the classes themselves (as well as the tests) to make them more testable and the tests more useful, but I think that can be done without changing anything about the public API.

I am not sure about this @acoulton . I looked around and I did not find a Kohana module or a OSS project in Kohana universe that extended Kohana_Cookie, except maybe just to define the cookie salt within the class. But still, I find it dangerous and I would say it should not be done for 3.3 if it requires refactoring.

I ticked a dozen issues above. These are the ones you mentioned, plus the ones that are already fixed or marked wontfix|worksforme.

Kindly, if someone can review the ones I ticked, before I move them to the bottom of the list and start working on the ones that need fixing. Just make sure that these do not require any fixing.

acoulton · 2014-09-17T21:07:25Z

@enov I've reviewed your ticked issues and I agree that they're all valid to close.

Regarding:

I find it dangerous and I would say it should not be done for 3.3 if it requires refactoring.

Just to be clear, I'm talking tiny amounts of internal refactoring. For example for Kohana_Cookie just extracting the two setcookie calls to something like:

class Kohana_Cookie {
    public static function delete() {
        //return setcookie($name, NULL, -86400, Cookie::$path, Cookie::$domain, Cookie::$secure, Cookie::$httponly);
        return static::setcookie($name, NULL, -86400, Cookie::$path, Cookie::$domain, Cookie::$secure, Cookie::$httponly);
    }

    protected static function setcookie($name, $value, $expire, $path, $domain, $secure, $httponly)
    {
        return setcookie($name, $value, $expire, $path, $domain, $secure, $httponly);
    }
}

I can't really see any way that could affect anyone even if they've extended and wrapped/replaced the existing methods? I guess only if they've added a static setcookie method of their own for some reason - we could pick a more obscure name like static::_kohana_internal_mockable_setcookie to really minimise the risk.

That would allow us to extend the class for unit testing and mock out the setcookie call so that it's possible to test those methods after headers have been sent. Then I'd rewrite/expand the tests so that they cover some of the actual expected behaviour - for eg at the moment the Cookie::set tests basically just assert that setcookie returns true, they don't seem to cover any of the expiration, salting, whatever.

I think this is less dangerous than having these tests permanently skipped and will help ensure that any more major refactoring into 3.4 doesn't break anything that worked in 3.3.

enov · 2014-09-18T07:16:33Z

I think this is less dangerous than having these tests permanently skipped and will help ensure that any more major refactoring into 3.4 doesn't break anything that worked in 3.3.

This is a valid point to make an exception. _kohana_internal_mockable_setcookie is ugly, I'd go for setcookie with the hope we don't break anything :) But we should be abusive with exceptions though ;)

enov · 2014-09-18T14:41:46Z

@acoulton Except for the first two issues that you merged, I have ticked 9 items that do not require any action from us. Could you kindly review them?

The cookie tests were all being skipped because headers are always sent by the PHPUnit test runner (fixes kohana/kohana#47). Additionally even when they ran they were covering very little of the actual logic of the class, mostly just asserting that setcookie returns TRUE. Rewrote the tests completely and slightly refactored the Cookie class to allow mocking of the timestamp and setcookie calls. Refs #547.

acoulton · 2014-09-18T16:01:49Z

enov · 2014-09-19T05:42:22Z

In order not to loose track, I have reopened the bugs R4092 and R3962 in Redmine, with status Feedback and rescheduled them for 3.4, as they are still valid bugs.

The cookie tests were all being skipped because headers are always sent by the PHPUnit test runner (fixes kohana/kohana#47). Additionally even when they ran they were covering very little of the actual logic of the class, mostly just asserting that setcookie returns TRUE. Rewrote the tests completely and slightly refactored the Cookie class to allow mocking of the timestamp and setcookie calls. Refs #547.

enov · 2014-11-11T10:14:57Z

@acoulton I have reviewed and ticked the remaining issues that are on the upper part of the list. There are few that I am not sure, I left them unchecked. Could you kindly review the checked ones?

acoulton · 2014-11-11T11:23:03Z

@enov will try and work through this list tomorrow

enov · 2014-11-19T08:42:25Z

Where art thou, @acoulton?

acoulton · 2014-11-19T11:11:51Z

@enov sorry things have been crazy.

Have looked over them just now and these all seem fine to me:

Quite a few were either worksforme, wontfix, or had fixes that were already merged to 3.3 (sometimes 3.2) and I couldn't see anything outstanding.

enov · 2014-11-19T11:17:28Z

We missed to see you around. I will take those issues you checked down in the list.

enov · 2014-11-26T09:37:54Z

@acoulton

I have checked the remaining issues:

Regarding R4601 I have also removed version_compare in e5ce1a3.
Regarding R4654 I have added few comments in kohana/kohana@6107011

R3860 and R3859 both rely on R3552 which was set to Unscheduled in Redmine.

Please review whenever you have time.

acoulton · 2014-11-26T10:16:51Z

@enov these look good to me, I think the two empty_rules issues should be closed and fixed in 3.4 with the R3552 solution or similar.

enov · 2014-11-26T10:21:25Z

Created a new issue #575 so that we do not forget about it.

acoulton · 2014-11-26T10:26:41Z

Ace. Possibly late in the day to suggest, but should we consider just running a one-time export from Redmine to Github - then we could be sure that nothing gets lost and stop using R... links? Someone must have built a tool for that, no?

enov · 2014-11-26T10:28:01Z

Good idea @acoulton, but don't know of any tool for that.

acoulton · 2014-11-26T10:29:05Z

@enov there's this - http://blogs.law.harvard.edu/rprasad/2014/07/10/moving-from-redmine-to-github-issues/ - I could try it out (on a dummy repo) and see what happens...

enov · 2014-11-26T10:29:31Z

In Github, remaining issues to release 3.3.3:

Documentation fixed by @KostyaTretyak in Fixed errors in docs #404
implemented a caching config reader cache#49 caching config reader
Added Cache_Xcache driver for Kohana 3 cache#30 Xcache driver
Redis driver cache#41 Redis driver

enov · 2014-11-26T10:30:43Z

Would you like me to add those above the big list on the top of this issue? Or should we differ those for a later release (3.3.4)?

enov · 2014-11-26T10:34:07Z

Ah, there is also kohana/cache#50. I think, since all modules follow one versioning, we need to add the same Github milestones on all repos.

acoulton · 2014-11-26T10:35:59Z

@enov I think those other github issues should go in 3.3.4 - they need a bit more work and aren't critical.

There's also a few open issues on Redmine that I didn't see on your list above? I think some may be invalid though. http://dev.kohanaframework.org/projects/kohana3/issues?query_id=66

enov · 2014-11-26T11:01:20Z

Alright, will create a new issue as those are not reported for version 3.2.3.

Do you concur with me that we should close this one? 🎉

enov added this to the 3.3.3 milestone Sep 17, 2014

This was referenced Sep 18, 2014

R4820 incorrect mention of the missing method Request::instance #550

Merged

Resolves R4814 by removing duplicate line of code kohana/userguide#64

Merged

related to R4813, removes redundant notice-disabling lines #551

Merged

acoulton mentioned this issue Sep 18, 2014

Fix skipped/incomplete core tests for 3.3 series #552

Merged

This was referenced Sep 19, 2014

R3931 - Register exit(1) as a shutdown function, to allow Kohana shutdown properly #553

Closed

Don't set empty body on external curl requests [Fixes R4791] #554

Closed

enov closed this as completed Nov 28, 2014

enov mentioned this issue Jan 21, 2015

add MySQLi driver kohana/database#91

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Issues reported for version 3.2.3 in Redmine that were not fixed in version 3.3.3 #547

Issues reported for version 3.2.3 in Redmine that were not fixed in version 3.3.3 #547

enov commented Sep 17, 2014

acoulton commented Sep 17, 2014

acoulton commented Sep 17, 2014

enov commented Sep 17, 2014

acoulton commented Sep 17, 2014

enov commented Sep 18, 2014

enov commented Sep 18, 2014

acoulton commented Sep 18, 2014

enov commented Sep 19, 2014

enov commented Nov 11, 2014

acoulton commented Nov 11, 2014

enov commented Nov 19, 2014

acoulton commented Nov 19, 2014

enov commented Nov 19, 2014

enov commented Nov 26, 2014

acoulton commented Nov 26, 2014

enov commented Nov 26, 2014

acoulton commented Nov 26, 2014

enov commented Nov 26, 2014

acoulton commented Nov 26, 2014

enov commented Nov 26, 2014

enov commented Nov 26, 2014

enov commented Nov 26, 2014

acoulton commented Nov 26, 2014

enov commented Nov 26, 2014

Issues reported for version 3.2.3 in Redmine that were not fixed in version 3.3.3 #547

Issues reported for version 3.2.3 in Redmine that were not fixed in version 3.3.3 #547

Comments

enov commented Sep 17, 2014

acoulton commented Sep 17, 2014

acoulton commented Sep 17, 2014

enov commented Sep 17, 2014

acoulton commented Sep 17, 2014

enov commented Sep 18, 2014

enov commented Sep 18, 2014

acoulton commented Sep 18, 2014

enov commented Sep 19, 2014

enov commented Nov 11, 2014

acoulton commented Nov 11, 2014

enov commented Nov 19, 2014

acoulton commented Nov 19, 2014

enov commented Nov 19, 2014

enov commented Nov 26, 2014

acoulton commented Nov 26, 2014

enov commented Nov 26, 2014

acoulton commented Nov 26, 2014

enov commented Nov 26, 2014

acoulton commented Nov 26, 2014

enov commented Nov 26, 2014

enov commented Nov 26, 2014

enov commented Nov 26, 2014

acoulton commented Nov 26, 2014

enov commented Nov 26, 2014