Remi Collet | 16 Aug 19:38 2014

More QA in Fedora for PHP stack, and Horde


Hi,

I should probably have tell you about this before my previous message.

See
http://blog.famillecollet.com/post/2014/08/12/Koschei-continuous-integration-of-PHP-stack-in-Fedora

Which means, that now, any change in the PHP stack, which could have
impact on Horde stack, will raise a test build, and so will allow us
to detect ASAP any regression / breakage.

Exemple:
http://koschei.cloud.fedoraproject.org/package/php-horde-Horde-Kolab-Storage

Notice, for 7302107, build was raised by the update from PHP 5.6.0RC3
to PHP 5.6.0RC4 (RC3, in Fedora, have a temporary patch to keep 5.5
behavior of unserialize, dropped in RC4, which explain the failure)

I really think this should helps us to make the "PHP stack" and "Horde
stack" much better.

For now, only "rawhide" (Fedora devel branch, with php 5.6) is
monitored... but this could change in the future.

Remi.
Remi Collet | 16 Aug 18:58 2014

Horde-Kolab-Storage - unserialize issue


Hi,

With PHP 5.6.0 test suite fails :

There were 51 errors:

1) Horde_Kolab_Storage_Unit_Data_BaseTest::testDefaultType
Erroneous data format for unserializing 'Horde_Mime_Part'

/dev/shm/BUILD/php-horde-Horde-Kolab-Storage-2.1.0/Horde_Kolab_Storage-2.1.0/lib/Horde/Kolab/Storage/Driver/Mock/Data.php:103
...

Digging in this issue.

First: you should be awarer that "serialized" data is a very bad
pratice for storage... no warranty to work accross various versions.

See the recent discussion about the "unserialize" regression/issue...

The test/Horde/Kolab/Storage/fixtures/event.struct contains:
O:15:"Horde_Mime_Part":...

Which is wrong, as Horde_Mime_Part implements Serializable.
(should be C:... with PHP > 5.4)

Proof.

<?php
require 'Horde/Autoloader/Default.php';
(Continue reading)

Remi Collet | 16 Aug 19:08 2014

Horde-Kolab-Storage - unserialize issue


Hi,

With PHP 5.6.0 test suite fails :

There were 51 errors:

1) Horde_Kolab_Storage_Unit_Data_BaseTest::testDefaultType
Erroneous data format for unserializing 'Horde_Mime_Part'

/dev/shm/BUILD/php-horde-Horde-Kolab-Storage-2.1.0/Horde_Kolab_Storage-2.1.0/lib/Horde/Kolab/Storage/Driver/Mock/Data.php:103
...

Digging in this issue.

First: you should be awarer that "serialized" data is a very bad
pratice for storage... no warranty to work accross various versions.

See the recent discussion about the "unserialize" regression/issue...

The test/Horde/Kolab/Storage/fixtures/event.struct contains:
O:15:"Horde_Mime_Part":...

Which is wrong, as Horde_Mime_Part implements Serializable.
(should be C:... with PHP > 5.4)

Proof.

<?php
require 'Horde/Autoloader/Default.php';
(Continue reading)

Michael M Slusarz | 15 Aug 07:19 2014

Git repo splitting

Now that the x.2 releases have been made, and several months of  
bugfixes have occurred, we need to turn our attention to starting to  
plan the git repo split.

The straw that broke the camels back (or at least my back) is that as  
of right now Travis is completely unusable - at least from a practical  
perspective.  You make a small change in Horde_Foo, and Travis returns  
an error because of a completely unrelated SQL error in Horde_Bar.   
This error may very well be tied into the fact that we are running  
such a monster PHPUnit test that is stressing their setup too badly.   
(Another issue: I commit something that breaks Horde_Foo. Other devs  
should not have to worry about this broken behavior if they later  
commit something to Horde_Bar).

So this is brainstorming on the proper path to do this.

I have (mostly) finished the script necessary to do the actual  
splitting of any given repository.  It correctly filters the repo and  
transfers all tags - and it rewrites the tags to the form used by  
Composer.  Currently it only transfers the master branch.  Need to add  
support so that applications automatically keep FRAMEWORK_* branches  
also.  And add option to allow additional branches to be kept since we  
have a few of those lying around.

I'm really not too concerned about developer checkout.  We can  
worry/talk about things like git subtree later IMHO.  The most  
important tool will be something to update all git repos at once, but  
this is something that can be written in 5 minutes.  And even that is  
not critical.

(Continue reading)

Picon

invitation to participate in a PhD dissertation survey on requirements engineering practices used in open source software development

Dear sir/madam,

You are invited to participate in a survey titled "requirements engineering
practices used in the development of open source software projects". This
survey is part of a PhD dissertation study on requirements gathering in
open source software development.

This survey contains questions about which requirements engineering
practices are being used during open source software development, how
useful requirements engineering practices are for open source software
development and the problems and issues that open source software
developers face while carrying out requirements gathering activities during
open source software development. The survey would take around twenty
minutes to complete. The questions are not mandatory. If you are interested
in participating in this survey, please fill out the survey by clicking on
the link below:

http://www.surveygizmo.com/s3/1582452/ba155dc1677a

Researcher information: My name is Jaison Kuriakose and I am a PhD
candidate at memorial university, Canada. My contact email is jk5573 <at> mun.ca

*The proposal for this research has been reviewed by the Interdisciplinary
Committee on Ethics in Human Research and found to be in compliance with
Memorial University's ethics policy. If you have ethical concerns about the
research, such as the way you have been treated or your rights as a
participant, you may contact the Chairperson of the ICEHR at icehr <at> mun.ca
<icehr <at> mun.ca> or by telephone at 709-864-2861. *

Thank you
(Continue reading)

Michael J Rubinsky | 11 Aug 21:34 2014

testing, ignore.

mike 
Sent from mobile 
--

-- 
dev mailing list
Frequently Asked Questions: http://wiki.horde.org/FAQ
To unsubscribe, mail: dev-unsubscribe <at> lists.horde.org
Jasper Olbrich | 11 Aug 13:02 2014
Picon

Horde_Util 2.5.0/contact list for new email/pear packages

Hello,

one of our users experienced a change in the "Adress Book" page when you 
click the "To" link in the email composition window.
In Horde 5.1.6, the initial view showed all entries in the default 
adress book, but in Horde 5.2.1, the view is empty, probably because of 
this commit

https://github.com/horde/horde/commit/a58ffee8f7430324467a11b018446019888acead

With current git master code, everything works again, I think because of 
this

https://github.com/horde/horde/commit/0df158de3fec8c94ae9e89e14d3495be273b2428#diff-038fa079bf9dcfd9343a0881e2f3e8ce

This commit has a tag for Horde_Util-2.5.0, but the change is not 
contained in the current PEAR package. Is this because PEAR packages are 
built from the FRAMEWORK_X_Y-branches?
When will this change be installable via PEAR?

--

-- 
dev mailing list
Frequently Asked Questions: http://wiki.horde.org/FAQ
To unsubscribe, mail: dev-unsubscribe <at> lists.horde.org

Jasper Olbrich | 30 Jul 17:11 2014
Picon

Bug #13319: possible fix

Hello,

I think I've implemented a fix for http://bugs.horde.org/ticket/13319.
The main problem is that external calendars are no longer working if 
they expect HTTP Basic Auth.

The problem is that in kronolith/lib/Driver/Ical.php the variable 
$sabreOptions['userName'] is set, but this breaks the request method in
pear/php/Horde/Dav/Client.php because the authType of the SabreDAV 
client is AUTH_BASIC | AUTH_DIGEST. That leads to the DIGEST method 
being used in the Horde_Dav_Client class.

My idea is to use the HTTP 401 response from the external calendar 
provider when the user is adding the external calendar, because that's 
the place where the provider tells us which authentication mode is required.

In my (AFAI can tell, working) solution, this information is collected 
in Horde/Dav/Client.php and passed along using the Horde_Dav_Exception 
and the following Kronolith_Exception back to the Browser.

Then the user enters his credentials and the desired authentication mode 
is sent back to the horde server as well. There it is used to access the 
calendar provider and finally stored in the database.

I don't know why the user name is set in the sabreOptions in the first 
place (the change was introduced here

https://github.com/horde/horde/commit/477f2c46041bde1b3bfa572d7fbac535046c3be6#diff-2a209d195528a6339e273da67af8fd86), 
but now something has to be done to fix this. If my approach is not 
completely wrong, I could provide a patch. Should I attach it to the bug 
(Continue reading)

Hiromi Kimura | 29 Jul 04:34 2014
Picon

Horde_Imap_Client-2.24.0 fatal error

Horde_Imap_Client-2.24.0 update() function has this bug:

EMERG: HORDE52 [imp] Call to a member function isEnabled() on a non-object \
[pid 1989 on line 216 of "/usr/local/www/horde52/libs/Horde/Imap/Client/Socket.php"]

-=-=-=-=-
Hiromi Kimura, Tandem Accelerator Complex, University of Tsukuba, JAPAN

--

-- 
dev mailing list
Frequently Asked Questions: http://wiki.horde.org/FAQ
To unsubscribe, mail: dev-unsubscribe <at> lists.horde.org

Jasper Olbrich | 21 Jul 11:35 2014
Picon

Is "return true" the sane fallback for hasAppPermission?

Hi,

I just stumbled upon the function hasAppPermission in Horde/Core/Perms.php:

     if ($this->_perms->exists($app . ':' . $permission)) {
         // determine permission
     }
     return true;

Is this "return true;" intentional? In our case, this made everyone 
"Resource Manager" in Kronolith untill we created the corresponding 
permission and set it up appropriately.

-- 
Regards,
Jasper
--

-- 
dev mailing list
Frequently Asked Questions: http://wiki.horde.org/FAQ
To unsubscribe, mail: dev-unsubscribe <at> lists.horde.org

Ralf Lang | 17 Jul 22:17 2014
Picon

horde pirum composer repository available?

Hi,

Does the current pear.horde.org create a composer repository and how can
I get its (meta) data?

I'm currently retooling and refactoring the horde distribution updater
to make it more modular, less suse-specific and possibly more
future-oriented.

I currently retrieve most data from pear.horde.org/feed.xml to inject it
into rpm .spec files (production) and debian .dsc (wip) files.

During the last months, notable projects have been shifting away from
pear/pirum to composer. I read that pirum as of version 1.1 creates a
composer repository. OTOH pirum upstream says it's deprecated, so
reading composer files and installing the stuff to some global "system"
repository might be the way to go.

Another approach I am investigating is directly monitoring git
repositories for tags that look like semantic versions, but pending the
long-awaited horde repository split I don't think this is the right time
to do it.

--

-- 
Ralf Lang
Linux Consultant / Developer
Tel.: +49-170-6381563
Mail: lang <at> b1-systems.de
B1 Systems GmbH
Osterfeldstraße 7 / 85088 Vohburg / http://www.b1-systems.de
(Continue reading)


Gmane