[Vtigercrm-developers] performance and profiling

[Conrado Maggi]

Hi,

getPicklistValues is slow because is has to check for each picklist value
whether the login users has access to it or not.

Probably there is a way to cache that info. Since only changes when you
update the picklist values.

Cheers,
Conrado

On Tue, Oct 6, 2015 at 10:36 AM, lajeesh k <lajeeshk at gmail.com> wrote:

> applied to_html changes, still detail view taking 20 seconds to load page.
> summery view loading quickly
>
> I have used xdebud and webgrind for profiling. I am a newbie in profiling
> while checking vtranslate function inside
> Vtiger_Field_Model->getPicklistValues is one of the reason for this slownes
> I have attached webgrind screenshot, it might help someone to identify the
> bottleneck
>
>
>
>
>
> Regards,
> Lajeesh
>
> On Mon, Oct 5, 2015 at 10:25 PM, lajeesh k <lajeeshk at gmail.com> wrote:
>
>> seems these suggestions from Alan is not included in vtiger 6.3.
>> when i used xdebug profiling, this to_html function is called 226985
>> times. calling this function increases as number of records increases
>> I couldn't understand how to apply these changes by Alan and Sivakumar.if
>> Anyone can provide these modified file for vtiger 6.3, then it will be
>> really helpfull
>>
>>
>>
>> Regards,
>> Lajeesh
>>
>> On Tue, Jul 22, 2014 at 6:23 PM, SivaKumar J <sivakumar.j at vtiger.com>
>> wrote:
>>
>>> Hi Alan,
>>>
>>> Thank you very much for your contribution on to_html() function calls to
>>> improve performance.
>>> After analyzing to_html() function, we got that some of those checks
>>> what we are checking in that are not required any more in Vtiger6.
>>>
>>> at line 366
>>>
>>>> *if($_REQUEST['module'] != 'Settings' && $_REQUEST['file'] !=
>>>> 'ListView' && $_REQUEST['module'] != 'Portal' && $_REQUEST['module'] !=
>>>> "Reports")// && $_REQUEST['module'] != 'Emails')*
>>>> *        $ajax_action = $_REQUEST['module'].'Ajax';*
>>>>
>>> The above check is to identify whether the current request is AJAX or
>>> not. If so, we are setting that as *$_REQUEST['module'].'Ajax'. *But,
>>> there are not AJAX file in Vtiger6 in that structure. So, we don't need
>>> this.
>>>
>>> at line 371
>>>
>>>> *if($action != 'CustomView' && $action != 'Export' && $action !=
>>>> $ajax_action && $action != 'LeadConvertToEntities' && $action !=
>>>> 'CreatePDF' && $action != 'ConvertAsFAQ' && $_REQUEST['module'] !=
>>>> 'Dashboard' && $action != 'CreateSOPDF' && $action != 'SendPDFMail' &&
>>>> (!isset($_REQUEST['submode'])) )*
>>>> *    {*
>>>> *        $doconvert = true;*
>>>> *    }*
>>>
>>> In this check, except "Export"  we don't have any actions like what we
>>> are comparing, so no need of this comparision. Earlier there is an action
>>> called Export for Exporting record in Vtiger5.4.0, but now we are using
>>> ExportData.php for this action. While Exporting record, at
>>> *modules/Vtiger/actions/ExportData.php* we are sanitizing all values
>>> which will decode those values with decode_html() on line 243. So, there is
>>> no use of encoding with to_html() for Export as well. We can remove this
>>> also. Else, we need to add this here and remove decode_html() where we are
>>> doing in sanitize values.
>>>
>>>
>>> After doing all these modification, finally my code is like this.
>>>
>>> */***
>>>> * * Function to decide whether to_html should convert values or not for
>>>> a request*
>>>> * * @global type $doconvert*
>>>> * * @global type $inUTF8*
>>>> * * @global type $default_charset*
>>>> * */*
>>>> *function decide_to_html(){*
>>>> *    global $doconvert,$inUTF8,$default_charset; *
>>>> *     $action = $_REQUEST['action']; *
>>>>
>>>> *    $inUTF8 = (strtoupper($default_charset) == 'UTF-8'); *
>>>>
>>>> *    $doconvert = true; *
>>>>
>>>>
>>>
>>>    // Need to remove decode_html() which we are doing while exporting
>>>> records from list/Detail view on  ExportData.php line 243
>>>>
>>> *  if($action == 'ExportData'){*
>>>> *        $doconvert = false; *
>>>> *    }*
>>>> *}*
>>>> // Execute this function when this file got included
>>>> *decide_to_html();*
>>>>
>>>> */** Function to convert the given string to html*
>>>> *  * @param $string -- string:: Type string*
>>>> *  * @param $ecnode -- boolean:: Type boolean*
>>>> *  * @returns $string -- string:: Type string*
>>>> *  */*
>>>> *function to_html($string, $encode=true) {*
>>>> *    // For optimization - default_charset can be either upper / lower
>>>> case.*
>>>> *    global $doconvert,$inUTF8,$default_charset,$htmlCache;*
>>>>
>>>> *    if(is_string($string)) {*
>>>> *        // In vtiger5 ajax request are treated specially and the data
>>>> is encoded*
>>>> *        if ($doconvert == true) {*
>>>> *            if(isset($htmlCache[$string])){*
>>>> *                $string = $htmlCache[$string];*
>>>> *            }else{*
>>>> *                if($inUTF8)*
>>>> *                    $string = htmlentities($string, ENT_QUOTES,
>>>> $default_charset);*
>>>> *                else*
>>>> *                    $string = preg_replace(array('/</', '/>/', '/"/'),
>>>> array('<', '>', '"'), $string);*
>>>>
>>>> *                $htmlCache[$string] = $string;*
>>>> *            }*
>>>> *        }*
>>>> *    }*
>>>> *    return $string;*
>>>> *}*
>>>>
>>>
>>> Please let me know if I missed any thing in here.
>>> Thank you
>>>
>>>
>>>
>>>
>>>
>>> On Fri, Jul 18, 2014 at 6:34 PM, Alan Bell <alan.bell at libertus.co.uk>
>>> wrote:
>>>
>>>> turns out that most of what to_html is processing is very repetitive,
>>>> if it just populates an array as it goes along and then checks in that
>>>> array it is a lot faster than doing htmlentities on the string every time.
>>>>
>>>> global $htmlcache;//store the stripped HTML as we go along, a lot of
>>>> the time we are processing the same strings
>>>> function to_html($string, $encode=true)
>>>> {
>>>>         global $doconvert;
>>>>         global $default_charset;
>>>>         global $htmlcache;
>>>>         if($doconvert){
>>>>                 if(isset($htmlcache[$string])){
>>>>                     return $htmlcache[$string];
>>>>                 }else{
>>>>                     $clean= htmlentities($string, ENT_QUOTES,
>>>> $default_charset);//we don't care if it is a string or not, faster not to
>>>> care
>>>>                     $htmlcache[$string]=$clean;
>>>>                     return $clean;
>>>>                 }
>>>>         }else{
>>>>                 return $string;
>>>>         }
>>>>
>>>>
>>>> }
>>>>
>>>> _______________________________________________
>>>> http://www.vtiger.com/
>>>>
>>>
>>>
>>>
>>> --
>>> Regards,
>>> Siva
>>> Vtiger
>>>
>>> _______________________________________________
>>> http://www.vtiger.com/
>>>
>>
>>
>
> _______________________________________________
> http://www.vtiger.com/
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.vtigercrm.com/pipermail/vtigercrm-developers/attachments/20151006/8a49a1e8/attachment.html>