Henri Bergius <[EMAIL PROTECTED]> said:
> Midgard 1.4
> nadmin M2
>
I think we may have moved from inline replacement in fast template to 'sub
modules in fast template' - has M2 got alot of style elements <[element_....]>
?
there are a few changes that M3 has done to change articles majorly..
removal of history pulldowns, removal of other article pulldowns.... and so
on... - breaking into frames like pages & styles.
> If there are any fixes for this, please let me know. I haven't yet tried
> whether upgrading to CVS Midgard and nadmin would help, but if no
> solutions
> come up, I'll try that tomorrow morning. We will also upgrade memory on
> the
> server to 256Mb.
>
These are some more fixes to M3 (awaiting 1.4.1RC1 --- v.soon) that could be
applied to M2
Snippet: / hklc_adminsite_host / lib / config_set ** new to M3! - this code
was in the code init of the newframeset and rcs
41: while (list ($k,$v) = each($newcookie)) {
42: $kk = "display_setting_". $k;
43: if ( $thisperson->$kk)
44: $oldcookie[$k] = $thisperson->$kk;
45: $newcookie[$k] = $oldcookie[$k];
46: }
stops it doing a $object->parameter to get variables - which is horribly SQL
intensive
*** shouldnt this be optomized in the back end????
Snippet: / hklc_adminsite_host / argv0 / browse / browse
524: if (ereg("^(page_edit|hosts)$", $orgtype)) {
mgd_include_snippet($set[snippetpath] . "/argv0/browse/browse_stylesheets"); }
- removed the |article_edit| it was legacy code...** probably just |article|
in M2
Snippet: / hklc_adminsite_host / argv0 / browse / browse_authors
2: $cache_person = array();
3: if ($persons = mgd_list_persons())
4: while ($persons->fetch())
5: $cache_person[] =mgd_get_person ($persons->id);
6:
replace the if and while with
11: for ($j=0;$j< count($cache_person);$j++) {
12: $person_obj = $cache_person[$j];
change all references to $persons-> to $person_obj
remove additonal closing } on line 30:
** this code is very slow as it has to get do a list/fetch/get just to get
sitegroup info!!!!
perhaps changing all the list calls to get sitegroups would kick ass for
speed!!!
It still ends up doing 140K's worth of SQL calls - I think without the
mgd_get it would probably be down to about 20/30K
another idea would be building a temporary hash of name:up -> obj id!
this would speed up snippet tree includes.......
nice trick here
tail -f /var/log/mysql.log > /tmp/page.log
to study what sqlcalls a particular page is doing ...
regards
alan
> /Bergie
>
>
--
Technical Director
Linux Center (HK) Ltd.
www.hklc.com
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
