[Phpmyadmin-devel] Bug tracking for refactoring insert/edit

[Marc Delisle]

Le 2012-06-30 13:19, Thilina Buddika Abeyrathna a écrit :
> On Sat, Jun 30, 2012 at 5:53 PM, Thilina Buddika Abeyrathna <
> thilinaabeyrathna at gmail.com> wrote:
> 
>>
>>
>> On Sat, Jun 30, 2012 at 5:26 PM, Marc Delisle <marc at infomarc.info> wrote:
>>
>>> Le 2012-06-30 04:45, Thilina Buddika Abeyrathna a écrit :
>>>> On Thu, Jun 28, 2012 at 5:18 AM, Thilina Buddika Abeyrathna <
>>>> thilinaabeyrathna at gmail.com> wrote:
>>>>
>>>>>
>>>>>
>>>>> On Wed, Jun 27, 2012 at 5:43 PM, Marc Delisle <marc at infomarc.info>
>>> wrote:
>>>>>
>>>>>> Le 2012-06-26 13:16, Thilina Buddika Abeyrathna a écrit :
>>>>>>> On Sun, Jun 24, 2012 at 4:38 AM, Marc Delisle <marc at infomarc.info>
>>>>>> wrote:
>>>>>>>
>>>>>>>> Hi Thilina,
>>>>>>>> I prefer that you fix bugs in your previous refactoring before
>>> starting
>>>>>>>> a new sub-project (you have started work on server_privileges.php)
>>>>>>>>
>>>>>>>> Please compile a list of these bugs, open artefacts on our bug
>>> tracker
>>>>>>>> and reply here references to these bugs, in one post. I will then
>>>>>> verify
>>>>>>>> to ensure we did not miss anything.
>>>>>>>>
>>>>>>>>
>>>>>>> [1] -
>>>>>>>
>>>>>>
>>> https://sourceforge.net/tracker/?func=detail&aid=3538189&group_id=23067&atid=377408
>>>>>>
>>>>>> Thilina,
>>>>>> please create artifacts in the bug tracker for these other problems:
>>>>>>
>>>>>> 1. "Edit next row" choice is no longer offered in the actions panel
>>>>>>
>>>>>> 2. "Insert another new row" produces
>>>>>> Fatal error: Call to undefined function PMA_showMessage() in
>>>>>> /opt/pma-git/phpmyadmin/tbl_change.php on line 129
>>>>>>
>>>>>> C. A logic error remains about setting $timestamp_seen to true, see
>>> [0];
>>>>>> therefore, behavior is different than version 3.5 when inserting in a
>>>>>> table that has more than one timestamp.
>>>>>>
>>>>>> [0] https://github.com/thilinaa/phpmyadmin/commit/798ccce1
>>>>>>
>>>>>> Ok Marc.
>>>>> [0] -
>>>>>
>>> https://sourceforge.net/tracker/?func=detail&aid=3538530&group_id=23067&atid=377408
>>>>>  [1] -
>>>>>
>>> https://sourceforge.net/tracker/?func=detail&aid=3538527&group_id=23067&atid=377408
>>>>> [2] -
>>>>>
>>> https://sourceforge.net/tracker/?func=detail&aid=3538524&group_id=23067&atid=377408
>>>>>
>>>>>
>>>> [1] - fixed in [3] commit.
>>>> [3] -
>>>>
>>> https://github.com/thilinaa/phpmyadmin/commit/e953deca3c50ee424d7597ce890a8a2641c05ed8
>>>
>>> This one was not completely fixed, as the message containing the INSERT
>>> statement was not displayed. I fixed it with [4].
>>>
>>> Thanks marc.
>>
>>
>>> I found another bug: After a simple INSERT (insert as new row and then
>>> go back to previous page), the generated INSERT statement is not
>>> displayed. see [5].
>>>
>>> [4]
>>>
>>> https://github.com/phpmyadmin/phpmyadmin/commit/a83f42eaa6619b8351ff67bbe528098415d53644
>>>
>>> [5]
>>>
>>> https://sourceforge.net/tracker/?func=detail&aid=3539142&group_id=23067&atid=377408
>>
>>
>> I'll look Marc.
>>
> 
> Hi Marc. I did the 'git bisect'  to figure out when brokenness introduced.
> But according to that it happen not because of my changes. Last good commit
> is [0]
> bad commit [1] - this was an upstream merge
> 
> [0] -
> https://github.com/thilinaa/phpmyadmin/commit/11776f7f116d42f5c7a7d62f7321b879e16601f3
>  [1] -
> https://github.com/thilinaa/phpmyadmin/commit/f9999253132094226fe0f30aa886a76a176b7cd6

Thilina,
when you want to "git bisect", it's better to do it on upstream master
and not on your own tree. I did it and the first bad commit is [6]

which was by mynetx, kindly fixing an undefined variable in your code,
but which unfortulately had a side effect of producing bug #3539142.
Hopefully I have fixed this with [7].

[6]
https://github.com/phpmyadmin/phpmyadmin/commit/7f6e422952454dbe02924ffecd20cb1d1c1ff06a

[7]
https://github.com/phpmyadmin/phpmyadmin/commit/2c6f80aa47fe8bc960f1e781e1197d97e30e5179

-- 
Marc Delisle
http://infomarc.info