apc-aa-coders — Discussion only useful to active developers.

[Apc-aa-coders] mgettext questions

[Pep T. <pe...@tu...>]

Hi,

I was trying to improve the spanish translation of ActionApps, but I
found some things that I would like to ask / comment before commiting
changes to the CVS.

1. The _output_lang.php3 file, which as I understand contains text to be
   shown by view.php3 and slice.php3, does not get included. Therefore,
   things like the scroller allways appear in English. I think that they
   should appear in the language defined for the slice... The solution I
   found for this is to modify slice.php3 and include/view.php3 adding
   something like this:

	require_once "./include/mgettext.php3";
	$lang_file = substr ($slice_info[lang_file], 0, 2);
	if (!$LANGUAGE_NAMES [$lang_file])
	  $lang_file = "en";
	bind_mgettext_domain ($GLOBALS["AA_INC_PATH"]."lang/".$lang_file."_output_lang.php3");

   This works, but I am not sure it is the proper way to go.

2. A similar problem: the _param_wizard_lang file is not properly
   included. E.g. when editing a field, the translations for "validate"
   and "publish" values are not showing. I fixed this temporarily by
   including _param_wizard_lang from the _news_lang, but I know this is
   not the way to go...

3. I would like to add two new strings to translate: the names of the
   new nodes ('spot' / 'option') created by the site module. So I
   changed the corresponding strings to _m("..") in site/index.php3. As
   I understand, now I should run the "misc/mgettext/index.php3" script
   to update all the language files... should I do this and commit all
   of them to CVS?

A side note about the translation files. There are several strings that
are repeated along several of the _lang files, so one must translate the
same string several times... is this normal? 

pep

[Apc-aa-coders] Bug fixed - so now how to change a view to a new slice?

[Mitra <mi...@ea...>]

Hi

There used to be a nasty bug where if you changed slice while editing 
a view, then the view ended up in the new slice.

Its great ... someones fixed it ... but


Is there a way to change the slice a view refers to now?   (Hacks are 
accepted !)

Its not really important, but I just created a view in the wrong 
slice which is why I wanted to do this.

- Mitra


-- 
Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182

Life is a Mystery to be Lived, not a Problem to be Solved

[Apc-aa-coders] CVS updates: Calendar and Site

[Mitra <mi...@ea...>]

I've uploaded some changes to the CVS version

The first set are to do with Calendars, I would be surprised if any 
Calendars will run on the Release version since there were at least 
two bugs one of which I think would have stopped any calendars 
running.

The second set are with the site module, I've allowed adding a 
parameter &sitefix=1  that will do some fixing on broken site, you 
should see an error message if it needs fixing.   later - when I'm 
more sure of the causes - I'll automate the process.

honza - I think the problem is that when a element in an array gets 
deleted the next created element gets slotted into the vacant slot, 
not at the end, so you can end up with array s where the elements are 
NOT in the order of the keys, and so some of your loops aren't 
happening in key order.  This was also what I think was causing 
problems with Move-Up and Move-Down on occasions.

- Mitra

-- 
Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182

Life is a Mystery to be Lived, not a Problem to be Solved

Re: [Apc-aa-coders] eb_functions syntax, sliceobj.php

[Mitra <mi...@ea...>]

Hi Honza

I didn't hear back on this one ....

I was thinking about this a bit, and figured that using an array 
"eb_functions" might not be the best idea, although I still think the 
syntax {functionname:arg1:arg2:arg3:arg4} is the way to go.

Instead ... I was wondering if we should use something like the 
insert_fnc_name  method.

We would define  a set of functions like
expand_alias
expand_switch
expand_user
expand_math

and the site designer in their modules/site/sites/site_xxx.php3 could 
also define any functions they want.

Then the string expansion becomes simple
1: Parse the syntax  into functionname,arg1, arg2 etc
2:check "is_callable(functionname)"
3: If so call it
4: If not, then look for a field or file or the other examples by nam.
5: If still not, then parse the legacy syntaxes like 
switch(arg1)arg2:arg3:arg4:arg5

The advantage would be that the most common cases would be at the 
beginning of the parse, and caught by the is_callable.

What do you think?  I'm assuming that "is_callable" is fairly quick.

- Mitra



At 7:52 AM +1000 3/7/03, Mitra wrote:
>At 2:19 PM +0200 2/7/03, Honza Malik wrote:
>>Hi Mitra (and Norbert),
>>
>>1) *eb_functions*
>>
>>I checked the eb_function syntax:
>>
>>{<fnctn>:xxx:yyyy}, where <fnctn> is php function permitted in
>>$eb_function[] array.
>>
>>I want to ask you, if it will not be so big problem, if I change the
>>syntax to:
>>
>>{function:<fnctn>:xxx:yyyy} or something similar. The reason is, that
>>current syntax needs regular expressions for parsing (which is slower)
>>and it could also be a problem, when we will be rewriting alias parser
>>(stringexpand) in the future.
>>
>>eb_functions are new - not documented yet, so I hope it should not be
>>problem for other AA users.
>
>Are regular expressions really that much slower? But if regular 
>expressions are slower then I don't see that this syntax takes any 
>more complex parsing than {function:aaaa:xxxx:yyyy}, they both need 
>a string broken up at the colons ?  So I think speed of parsing is 
>independent of the syntax chosen.
>
>We use this same format in lots of places - e.g. {scroller:xxxxx}  I 
>would guess that if we could get the syntax consistent - e.g. 
>{xxx:yyy:zzz} then a single expression at the top would make it much 
>simpler.
>
>I'm not sure that it is a good idea to change the syntax  from 
>{aaa:bbb:ccc}  to {function:aaa:bbbb:cccc}  , part of my idea of 
>doing this is to make it easier for the code to be readable and 
>obvious, instead of having to remember the syntax of all the 
>possible things.  So for example I have a function {morefull} 
>which is defined in site_xxx.php3 to be a quite complex switch 
>statement  based on whether there is more full text or not.   Being 
>able to put {morefull} inside a view, is quite a bit more elegant 
>than {function:morefull} ,  also it allows easy expansion without 
>extending expand_bracketed   e.g.  {substr:{headline........}:1:40} 
>which is quite a bit more elegant than 
>{function:substr:{headline........}:1:40}
>
>stringexpand is quite intelligent, it will check for "morefull"  in 
>eb_functions, and in aliases etc etc, it might even give us quicker 
>code if we turned for example include and scroller and alias into 
>eb_functions so that they didn't have to be parsed individually?
>
>Also we have the legacy issue ... if we change it we'll have to 
>support the old syntax as well - like we still have to support the 
>{include(file)}  syntax for include.   Maybe I'm the only one to use 
>eb_functions,  but there are examples of this scattered all through 
>my sites,  so at least I would need a commented in version of the 
>old syntax.
>
>
>>
>>2) *slice class*
>>
>>I'm going to use slice objects defined in sliceobj.php3. It is great we
>>have such object - very useful. I also love the lazy evaluation approach
>>which saves time OR memory.
>>
>>In sliceobj you use lazy evaluation, which (by default) saves memory,
>>but is database intensive (one query for each field in slice table).
>>Do you use slice object with huge amount of slices, where such memory
>>savings is crucial and if, is it somewhere outside of 'slices' class?
>>If not, I plan to rewrite the class to load all slice records in one DB
>>query by default.
>
>I think that is a good idea.  I think this lazy evaluation actually 
>makes it slower in this case. What I suggest is to still make it 
>lazy, but load all the slice records in a DB query when the FIRST 
>field is accessed. That way a sliceobj can be used in some place 
>where the fields are never accessed without any penalty on 
>performance.
>
>- Mitra
>
>--
>Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
>home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182
>
>Life is a Mystery to be Lived, not a Problem to be Solved


-- 
Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182

Life is a Mystery to be Lived, not a Problem to be Solved

[Apc-aa-coders] Empty drop-down for Calendar type in views

[Mitra <mi...@ea...>]

Anyone know what the Empty drop-down for Calendar type when editing a 
Calendar View  is for ?

Is this hang-over from some older version of the code?

- Mitra

-- 
Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182

Life is a Mystery to be Lived, not a Problem to be Solved

[Apc-aa-coders] RE: [Apc-aa-general] Reader Management - Unique Email

[Marek T. <ma...@gn...>]

Hi, 

 
 
> I'm just creating a number of mailing lists on our sites ...
> 
> I'm wondering about hte "Unique Email" requirement in Reader Management Slices.
> 
> The validation is set to "con_email......:1"  i.e. check that the 
> email address is Unique across ALL SLICES
This is a bug - it should only test it accross All slices of Reader Management type.
 
> This would seem to me to prevent someone being a member of multiple 
> reader management domains on the same apc-aa installation. 
I could give you some reasons why we were forced during the process of planning RM module to make USERNAME unique. 
I can't give you an answer why the email address has to be unique right now.
I'll reply lter - need to look at it more properly when I have time. 

Sorry for any troubles. 
 Marek

[Apc-aa-coders] Alerts - how to access your own subscription

[Mitra <mi...@ea...>]

I'm trying to figure out how to access a User record in the anonymous form,

For example ..... I have a Subscribe link on the site, I'd like a 
user to be able to find their existing record, but if I retype the 
email address, even with the correct password,  it complains that the 
email address is in use.  (this is expected since fillform only 
allows the HTTP access, or access code mechanism, not password).

I can see a number of ways to change the code to do this - for 
example recognizing an existing email address and password, but I 
wonder if someone has already got a way to do this elegantly with the 
existing code?

- Mitra

-- 
Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182

Life is a Mystery to be Lived, not a Problem to be Solved

[Apc-aa-coders] Changes committed to CVS (many)

[Mitra <mi...@ea...>]

I've committed a number of changes to CVS that I've been holding back 
until the release,

The significant ones in terms of functionality are
* NOT sending confirmations when a new reader is added IF they are 
already confirmed (e.g. if you've set the form that way, or imported 
and defaulted them all to confirmed).
* added post2shtml to the keystr so that site.php3 caching cooperates with it.

Then .... a big structural change, I've split fillform.php3 into 
include/fillform.php3, this allows me to use it from site.php3,  i.e. 
anonymous forms now work in Site.php3 as well as Slice.php3

(this won't be in the documentation, since  anonymous forms are 
documented in DocBook.)

The rest of the changes are listed below (from the CHANGES file)

- Mitra


07/05/03 - added defaults and default functions and random  to importer.php3
(mitra)  - don't send welcome inviting confirmation if already confirmed
          - removed $db paramater from "new itemview"
          - added link to editsliceinfo and colors in summarize
          - used newer unaliasing in mail.php3
          - removed $db paramater from new manager
          - use getDB/freeDB in alerts_sending alerts/event
          - extended timelimit in alerts/event to 600 secs
          - added post2shtml_id to key_str when caching site.php3
          - removed warnings on unable to expand {0,4}

-- 
Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182

Life is a Mystery to be Lived, not a Problem to be Solved

[Apc-aa-coders] fix to config.php3 in CVS

[Mitra <mi...@ea...>]

I've committed a fix to config.php3 - it checked AA_INC_PATH before 
setting it, so any site that is built on it will output the error 
message.

This might be worth fixing in the release?

- Mitra

-- 
Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182

Life is a Mystery to be Lived, not a Problem to be Solved

Re: [Apc-aa-coders] switch behavior

[Norbert B. <br...@ch...>]

From: "Honza Malik" <hon...@ec...>
> you are right - if you use {switch({variable}).+:NOTEMPTY:EMPTY} when
> variable is not defined, NOTEMPTY string is displayed, but in current
> setting it IS expected behavior.

Did not realized that. Thanks (also to Mitra) for the explanation.=20


> {variable} string is expanded only if variable is defined (or if it is
> listed in eb_functions). If variable is not defined, the string stays
> unchanged. I think it is not so bad behavior - sometimes we do not =
want
> to eat all content in curly braces - say in javascript. We never know,
> what user fills in the view/site definition and what content (s)he =
want
> to achieve. The expansion of only known variables is good approach, I
> think.


n.

Re: [Apc-aa-coders] switch behavior

[Honza M. <hon...@ec...>]

Hi Norbert,

you are right - if you use {switch({variable}).+:NOTEMPTY:EMPTY} when
variable is not defined, NOTEMPTY string is displayed, but in current
setting it IS expected behavior.

{variable} string is expanded only if variable is defined (or if it is
listed in eb_functions). If variable is not defined, the string stays
unchanged. I think it is not so bad behavior - sometimes we do not want
to eat all content in curly braces - say in javascript. We never know,
what user fills in the view/site definition and what content (s)he want
to achieve. The expansion of only known variables is good approach, I
think.

Another thing is, that we should mention it in documentation.

> What about new version of AA? 

July 18-th I will go for holiday, so it will be before this date. On the
other hand there will not be much changes from current CVS, so if you
need it right now, you can install it form CVS (and run sql_update.php3)

   Honza




03. 07. 2003 v 11:54, Norbert Brazda napsal:

> I used what you have suggested. Result is: "NOT{variable}EMPTY". The
> problem is that variable is NOT DEFINED (i.e. it is not empty, but
> does not exist at all). I think that switch should show "other"
> option, not "not empty" option in such case.
> 
> I have solved this in our case, but would like to say that this
> behavior is not expected and might cause problems to other people
> using switch function.
> 
> 
> norbert
> br...@ch...
> 
> 
> 
> From: "Mitra" <mi...@ea...>
> > I use this syntax all the time (actually I use
> > 
> > {switch({variable}).+:NOTEMPTY::EMPTY}
> > 
> > but I think they behave the same way. I would suggest outputing 
> > {variable} and checking that its really empty. Like
> > 
> > {switch({variable}).+:NOT{variable}EMPTY::EMPTY}
> > 
> > 
> > - Mitra
> > 
> > 
> > At 5:22 PM +0200 2/7/03, Norbert Brazda wrote:
> > >We have found that {switch():} function is not behaving as expected...
> > >
> > >{switch({variable}).+:NOTEMPTY:EMPTY}
> > >should output NOTEMPTY string if the variable is not empty and EMPTY 
> > >string if variable is empty. This works.
> > >
> > >One would expected that if variable is not defined, switch function 
> > >would output EMPTY string. But it outputs NOTEMPTY string.
> > >
> > >Please, correct it.
> > >
> > >
> > >
> > >     Norbert Brazda
> > >
> > >     CHANGENET.SK
> > >     Mlynske nivy 41, 821 09  Bratislava
> > >     tel/fax: 02-55560026, 0905-729359
> > >
> > >
> > >     SPAJAME LUDI, KTORI MENIA SVET
> > >     ________________________________________________________________
> > >     Prinasame pohlad na svet ocami aktivnych obcanov. Vytvarame 
> > >nezavisly priestor
> > >     pre kriticke nazory a reflexiu diania okolo nas. Poskytujeme 
> > >nekomercne internetove
> > >     sluzby pre neziskove organizacie a samospravy.
> > >
> > >
> > >
> > >-------------------------------------------------------
> > >This SF.Net email sponsored by: Free pre-built ASP.NET sites including
> > >Data Reports, E-commerce, Portals, and Forums are available now.
> > >Download today and enter to win an XBOX or Visual Studio .NET.
> > >http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
> > >_______________________________________________
> > >Apc-aa-coders mailing list
> > >Apc...@li...
> > >https://lists.sourceforge.net/lists/listinfo/apc-aa-coders
> > 
> > 
> > -- 
> > Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
> > home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182
> > 
> > Life is a Mystery to be Lived, not a Problem to be Solved
> > 
> 
> 
> -------------------------------------------------------
> This SF.Net email sponsored by: Free pre-built ASP.NET sites including
> Data Reports, E-commerce, Portals, and Forums are available now.
> Download today and enter to win an XBOX or Visual Studio .NET.
> http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
> _______________________________________________
> Apc-aa-coders mailing list
> Apc...@li...
> https://lists.sourceforge.net/lists/listinfo/apc-aa-coders
> 

Re: [Apc-aa-coders] switch behavior

[Mitra <mi...@ea...>]

{variable}  expands to "{variable}" when it doesn't exist to avoid 
unecessarily catching it when people inadvertantly put { or } in 
something that they don't want expanded. This happens most often with 
Javascript in a view.

One reminder ..... if you have a problem, add &errcheck=1  to the URL 
, it catches many errors (including undefined variables).

In fact I have this code in my config.php3 ...

if ($GLOBALS[REMOTE_ADDR] == "203.43.123.456") {
     $errcheck = 1;
#    $debugcache = 1;
#    $debugupload = 1;
#    $debugsi = 1; // Debug Storeitem
#    $debugfill = 1;
#    $debugimport = 1;
       $trace = 1; // On error messages try to display a trace of 
where the problem is.
}

So that if it spots the user is me (based on an IP address that I 
use, but the users don't) then it sets errcheck=1 (commented out 
alternatives for debugging other sorts of problems).

- Mitra


At 11:54 AM +0200 3/7/03, Norbert Brazda wrote:
>I used what you have suggested. Result is: "NOT{variable}EMPTY". The 
>problem is that variable is NOT DEFINED (i.e. it is not empty, but 
>does not exist at all). I think that switch should show "other" 
>option, not "not empty" option in such case.
>
>I have solved this in our case, but would like to say that this 
>behavior is not expected and might cause problems to other people 
>using switch function.
>
>
>norbert
>br...@ch...
>
>
>
>From: "Mitra" <mi...@ea...>
>>  I use this syntax all the time (actually I use
>>
>>  {switch({variable}).+:NOTEMPTY::EMPTY}
>>
>>  but I think they behave the same way. I would suggest outputing
>>  {variable} and checking that its really empty. Like
>>
>>  {switch({variable}).+:NOT{variable}EMPTY::EMPTY}
>>
>>
>>  - Mitra
>>
>>
>>  At 5:22 PM +0200 2/7/03, Norbert Brazda wrote:
>>  >We have found that {switch():} function is not behaving as expected...
>>  >
>>  >{switch({variable}).+:NOTEMPTY:EMPTY}
>>  >should output NOTEMPTY string if the variable is not empty and EMPTY
>>  >string if variable is empty. This works.
>>  >
>>  >One would expected that if variable is not defined, switch function
>>  >would output EMPTY string. But it outputs NOTEMPTY string.
>>  >
>>  >Please, correct it.
>>  >
>>  >
>>  >
>>  >     Norbert Brazda
>>  >
>>  >     CHANGENET.SK
>>  >     Mlynske nivy 41, 821 09  Bratislava
>>  >     tel/fax: 02-55560026, 0905-729359
>>  >
>>  >
>>  >     SPAJAME LUDI, KTORI MENIA SVET
>>  >     ________________________________________________________________
>>  >     Prinasame pohlad na svet ocami aktivnych obcanov. Vytvarame
>>  >nezavisly priestor
>>  >     pre kriticke nazory a reflexiu diania okolo nas. Poskytujeme
>>  >nekomercne internetove
>>  >     sluzby pre neziskove organizacie a samospravy.
>>  >
>>  >
>>  >
>>  >-------------------------------------------------------
>>  >This SF.Net email sponsored by: Free pre-built ASP.NET sites including
>>  >Data Reports, E-commerce, Portals, and Forums are available now.
>>  >Download today and enter to win an XBOX or Visual Studio .NET.
>>  >http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
>>  >_______________________________________________
>>  >Apc-aa-coders mailing list
>>  >Apc...@li...
>>  >https://lists.sourceforge.net/lists/listinfo/apc-aa-coders
>>
>>
>>  --
>>  Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
>>  home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182
>>
>>  Life is a Mystery to be Lived, not a Problem to be Solved
>>
>
>
>-------------------------------------------------------
>This SF.Net email sponsored by: Free pre-built ASP.NET sites including
>Data Reports, E-commerce, Portals, and Forums are available now.
>Download today and enter to win an XBOX or Visual Studio .NET.
>http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
>_______________________________________________
>Apc-aa-coders mailing list
>Apc...@li...
>https://lists.sourceforge.net/lists/listinfo/apc-aa-coders


-- 
Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182

Life is a Mystery to be Lived, not a Problem to be Solved

Re: [Apc-aa-coders] switch behavior

[Norbert B. <br...@ch...>]

I used what you have suggested. Result is: "NOT{variable}EMPTY". The =
problem is that variable is NOT DEFINED (i.e. it is not empty, but does =
not exist at all). I think that switch should show "other" option, not =
"not empty" option in such case.

I have solved this in our case, but would like to say that this behavior =
is not expected and might cause problems to other people using switch =
function.


norbert
br...@ch...



From: "Mitra" <mi...@ea...>
> I use this syntax all the time (actually I use
>=20
> {switch({variable}).+:NOTEMPTY::EMPTY}
>=20
> but I think they behave the same way. I would suggest outputing=20
> {variable} and checking that its really empty. Like
>=20
> {switch({variable}).+:NOT{variable}EMPTY::EMPTY}
>=20
>=20
> - Mitra
>=20
>=20
> At 5:22 PM +0200 2/7/03, Norbert Brazda wrote:
> >We have found that {switch():} function is not behaving as =
expected...
> >
> >{switch({variable}).+:NOTEMPTY:EMPTY}
> >should output NOTEMPTY string if the variable is not empty and EMPTY=20
> >string if variable is empty. This works.
> >
> >One would expected that if variable is not defined, switch function=20
> >would output EMPTY string. But it outputs NOTEMPTY string.
> >
> >Please, correct it.
> >
> >
> >
> >     Norbert Brazda
> >
> >     CHANGENET.SK
> >     Mlynske nivy 41, 821 09  Bratislava
> >     tel/fax: 02-55560026, 0905-729359
> >
> >
> >     SPAJAME LUDI, KTORI MENIA SVET
> >     ________________________________________________________________
> >     Prinasame pohlad na svet ocami aktivnych obcanov. Vytvarame=20
> >nezavisly priestor
> >     pre kriticke nazory a reflexiu diania okolo nas. Poskytujeme=20
> >nekomercne internetove
> >     sluzby pre neziskove organizacie a samospravy.
> >
> >
> >
> >-------------------------------------------------------
> >This SF.Net email sponsored by: Free pre-built ASP.NET sites =
including
> >Data Reports, E-commerce, Portals, and Forums are available now.
> >Download today and enter to win an XBOX or Visual Studio .NET.
> =
>http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01=

> >_______________________________________________
> >Apc-aa-coders mailing list
> >Apc...@li...
> >https://lists.sourceforge.net/lists/listinfo/apc-aa-coders
>=20
>=20
> --=20
> Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
> home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182
>=20
> Life is a Mystery to be Lived, not a Problem to be Solved
> 

Re: [Apc-aa-coders] switch behavior

[Mitra <mi...@ea...>]

I use this syntax all the time (actually I use

{switch({variable}).+:NOTEMPTY::EMPTY}

but I think they behave the same way. I would suggest outputing 
{variable} and checking that its really empty. Like

{switch({variable}).+:NOT{variable}EMPTY::EMPTY}


- Mitra


At 5:22 PM +0200 2/7/03, Norbert Brazda wrote:
>We have found that {switch():} function is not behaving as expected...
>
>{switch({variable}).+:NOTEMPTY:EMPTY}
>should output NOTEMPTY string if the variable is not empty and EMPTY 
>string if variable is empty. This works.
>
>One would expected that if variable is not defined, switch function 
>would output EMPTY string. But it outputs NOTEMPTY string.
>
>Please, correct it.
>
>
>
>     Norbert Brazda
>
>     CHANGENET.SK
>     Mlynske nivy 41, 821 09  Bratislava
>     tel/fax: 02-55560026, 0905-729359
>
>
>     SPAJAME LUDI, KTORI MENIA SVET
>     ________________________________________________________________
>     Prinasame pohlad na svet ocami aktivnych obcanov. Vytvarame 
>nezavisly priestor
>     pre kriticke nazory a reflexiu diania okolo nas. Poskytujeme 
>nekomercne internetove
>     sluzby pre neziskove organizacie a samospravy.
>
>
>
>-------------------------------------------------------
>This SF.Net email sponsored by: Free pre-built ASP.NET sites including
>Data Reports, E-commerce, Portals, and Forums are available now.
>Download today and enter to win an XBOX or Visual Studio .NET.
>http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
>_______________________________________________
>Apc-aa-coders mailing list
>Apc...@li...
>https://lists.sourceforge.net/lists/listinfo/apc-aa-coders


-- 
Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182

Life is a Mystery to be Lived, not a Problem to be Solved

Re: [Apc-aa-coders] eb_functions syntax, sliceobj.php

[Mitra <mi...@ea...>]

At 2:19 PM +0200 2/7/03, Honza Malik wrote:
>Hi Mitra (and Norbert),
>
>1) *eb_functions*
>
>I checked the eb_function syntax:
>
>{<fnctn>:xxx:yyyy}, where <fnctn> is php function permitted in
>$eb_function[] array.
>
>I want to ask you, if it will not be so big problem, if I change the
>syntax to:
>
>{function:<fnctn>:xxx:yyyy} or something similar. The reason is, that
>current syntax needs regular expressions for parsing (which is slower)
>and it could also be a problem, when we will be rewriting alias parser
>(stringexpand) in the future.
>
>eb_functions are new - not documented yet, so I hope it should not be
>problem for other AA users.

Are regular expressions really that much slower? But if regular 
expressions are slower then I don't see that this syntax takes any 
more complex parsing than {function:aaaa:xxxx:yyyy}, they both need a 
string broken up at the colons ?  So I think speed of parsing is 
independent of the syntax chosen.

We use this same format in lots of places - e.g. {scroller:xxxxx}  I 
would guess that if we could get the syntax consistent - e.g. 
{xxx:yyy:zzz} then a single expression at the top would make it much 
simpler.

I'm not sure that it is a good idea to change the syntax  from 
{aaa:bbb:ccc}  to {function:aaa:bbbb:cccc}  , part of my idea of 
doing this is to make it easier for the code to be readable and 
obvious, instead of having to remember the syntax of all the possible 
things.  So for example I have a function {morefull}   which is 
defined in site_xxx.php3 to be a quite complex switch statement 
based on whether there is more full text or not.   Being able to put 
{morefull} inside a view, is quite a bit more elegant than 
{function:morefull} ,  also it allows easy expansion without 
extending expand_bracketed   e.g.  {substr:{headline........}:1:40} 
which is quite a bit more elegant than 
{function:substr:{headline........}:1:40}

stringexpand is quite intelligent, it will check for "morefull"  in 
eb_functions, and in aliases etc etc, it might even give us quicker 
code if we turned for example include and scroller and alias into 
eb_functions so that they didn't have to be parsed individually?

Also we have the legacy issue ... if we change it we'll have to 
support the old syntax as well - like we still have to support the 
{include(file)}  syntax for include.   Maybe I'm the only one to use 
eb_functions,  but there are examples of this scattered all through 
my sites,  so at least I would need a commented in version of the old 
syntax.


>
>2) *slice class*
>
>I'm going to use slice objects defined in sliceobj.php3. It is great we
>have such object - very useful. I also love the lazy evaluation approach
>which saves time OR memory.
>
>In sliceobj you use lazy evaluation, which (by default) saves memory,
>but is database intensive (one query for each field in slice table).
>Do you use slice object with huge amount of slices, where such memory
>savings is crucial and if, is it somewhere outside of 'slices' class?
>If not, I plan to rewrite the class to load all slice records in one DB
>query by default.

I think that is a good idea.  I think this lazy evaluation actually 
makes it slower in this case. What I suggest is to still make it 
lazy, but load all the slice records in a DB query when the FIRST 
field is accessed. That way a sliceobj can be used in some place 
where the fields are never accessed without any penalty on 
performance.

- Mitra

-- 
Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182

Life is a Mystery to be Lived, not a Problem to be Solved

[Apc-aa-coders] switch behavior

[Norbert B. <br...@ch...>]

We have found that {switch():} function is not behaving as expected...

{switch({variable}).+:NOTEMPTY:EMPTY}
should output NOTEMPTY string if the variable is not empty and EMPTY =
string if variable is empty. This works.=20

One would expected that if variable is not defined, switch function =
would output EMPTY string. But it outputs NOTEMPTY string.

Please, correct it.



    Norbert Brazda

    CHANGENET.SK
    Mlynske nivy 41, 821 09  Bratislava
    tel/fax: 02-55560026, 0905-729359


    SPAJAME LUDI, KTORI MENIA SVET
    ________________________________________________________________
    Prinasame pohlad na svet ocami aktivnych obcanov. Vytvarame =
nezavisly priestor=20
    pre kriticke nazory a reflexiu diania okolo nas. Poskytujeme =
nekomercne internetove
    sluzby pre neziskove organizacie a samospravy.

Re: [Apc-aa-coders] eb_functions syntax, sliceobj.php

[Norbert B. <br...@ch...>]

From: "Honza Malik" <hon...@ec...>
> {<fnctn>:xxx:yyyy}, where <fnctn> is php function permitted in
> $eb_function[] array.
>=20
> I want to ask you, if it will not be so big problem, if I change the
> syntax to:
>=20
> {function:<fnctn>:xxx:yyyy} or something similar. The reason is, that
> current syntax needs regular expressions for parsing (which is slower)
> and it could also be a problem, when we will be rewriting alias parser
> (stringexpand) in the future.

No problem for me. Just put some warning message in the CHANGES so I =
will not forgot to change it when upgrading.=20

What about new version of AA?=20



n.

[Apc-aa-coders] eb_functions syntax, sliceobj.php

[Honza M. <hon...@ec...>]

Hi Mitra (and Norbert),

1) *eb_functions*

I checked the eb_function syntax:

{<fnctn>:xxx:yyyy}, where <fnctn> is php function permitted in
$eb_function[] array.

I want to ask you, if it will not be so big problem, if I change the
syntax to:

{function:<fnctn>:xxx:yyyy} or something similar. The reason is, that
current syntax needs regular expressions for parsing (which is slower)
and it could also be a problem, when we will be rewriting alias parser
(stringexpand) in the future.

eb_functions are new - not documented yet, so I hope it should not be
problem for other AA users.

2) *slice class*

I'm going to use slice objects defined in sliceobj.php3. It is great we
have such object - very useful. I also love the lazy evaluation approach
which saves time OR memory.

In sliceobj you use lazy evaluation, which (by default) saves memory,
but is database intensive (one query for each field in slice table). 
Do you use slice object with huge amount of slices, where such memory
savings is crucial and if, is it somewhere outside of 'slices' class?
If not, I plan to rewrite the class to load all slice records in one DB
query by default.

   Regards,
                Honza

RE: [Apc-aa-coders] Timeouts in Alerts with only 60 users.

[Mitra <mi...@ea...>]

Thanks Marek,

Interestingly the mail sending is NOT the problem,

I commented out the mail->send line in mail.php3,=20
while I was testing (at first I was just looking=20
for the problem of people recieving duplicate=20
mails) - but still got the timeout problem, it=20
looks like it is the complexity of the database=20
stuff being done to build the email, loops within=20
loops, that is causing the problem.

I increased script_timeout as a short-term=20
solution, since the user is waiting for the=20
return this is obviously not a long-term=20
solution, especially when there might easily be a=20
few hundred people on an alerts list - i.e. a few=20
hundred times an average of one second per person=20
=3D several minutes waiting EACH TIME AN ITEM IS=20
POSTED.

I'll upgrade from CVS and see if that fixes it -=20
I'm in a course this week, so dare not do it=20
today since I won't be around to fix any new bugs=20
it introduces with the new release, and I bet=20
some changes clash with ones I've made but not=20
submitted so as not to muck up the release=20
process.

- Mitra




At 4:19 PM +0200 1/7/03, Marek Tichy wrote:
>Hi Mitra,
>
>in fact, the emails are not (should not) being=20
>send immediatelly as you post a new item. This=20
>would be causing delays for editors
>which we wanted to avoid.
>Instead, there is a cron entry for "instant"=20
>sending that runs by default every minute.
>
>However, it used to be like that for some time=20
>in the past. He have just discovered  that=20
>StoreItem is still calling obsolete
>AlertsSentInstantAlert.
>
>You can comment it out in=20
>include/Event_Handler.php3 or it's definition in=20
>modules/alerts/event.php3 or update from cvs
>
>But anyway, I thing there may be something up=20
>with the communication to your local mail=20
>delivery agent. Could you do some debugging
>to find out where is the bootleneck ? Maybe=20
>write a simple script that uses php mail=20
>function and test how long it takes to send out
>60 emails. I don't thing that building new email=20
>for each user is clogging it to such extent.
>
>But for now, increase script_timeout for your php scripts.
>And increase the value in cron so that "instant"=20
>script doesn't run that often - can be 5 minutes=20
>or so.
>
>Best
>
>   Marek
>
>
>
>>  -----Original Message-----
>>  From: apc...@li...
>>  [mailto:apc...@li...]On Behalf Of Mitra
>>  Sent: 1. =E8ervence 2003 10:43
>>  To: apc...@so...
>>  Subject: [Apc-aa-coders] Timeouts in Alerts with only 60 users.
>>
>>
>>  I just noticed a nasty problem in Alerts.
>>
>>  I have a new list - with about 65 members on it, all set to "Instant".
>>
>>  When I add an item it sits there for a LONG time, and eventually
>>  times out, I'm not totally sure why yet, but I think it is because
>>  send_emails is doing a lot of work to figure out each reader, and/or
>>  is then calling send_email_from_table building a new email for each
>>  user.
>>
>>  I think it has to do this because of the confirmation code at the
>>  bottom of every email.
>>
>>  The worst thing is that after sending, because it times out it
>>  doesn't mark the message as sent, so the next cron task sends it
>>  again, so I've got half of my list of members who got it twice, and
>>  half didn't get it at all :-(
>>
>>  Any ideas on fixing this.
>>
>>  - Mitra
>>
>>  --
>>  Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
>>  home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182
>>
>>  Life is a Mystery to be Lived, not a Problem to be Solved
>>
>>
>>
>>  -------------------------------------------------------
>>  This SF.Net email sponsored by: Free pre-built ASP.NET sites including
>>  Data Reports, E-commerce, Portals, and Forums are available now.
>>  Download today and enter to win an XBOX or Visual Studio .NET.
>>  http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
>>  _______________________________________________
>>  Apc-aa-coders mailing list
>>  Apc...@li...
>>  https://lists.sourceforge.net/lists/listinfo/apc-aa-coders
>  >
>
>
>
>-------------------------------------------------------
>This SF.Net email sponsored by: Free pre-built ASP.NET sites including
>Data Reports, E-commerce, Portals, and Forums are available now.
>Download today and enter to win an XBOX or Visual Studio .NET.
>http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
>_______________________________________________
>Apc-aa-coders mailing list
>Apc...@li...
>https://lists.sourceforge.net/lists/listinfo/apc-aa-coders


-- 
Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182

Life is a Mystery to be Lived, not a Problem to be Solved

RE: [Apc-aa-coders] Timeouts in Alerts with only 60 users.

[Marek T. <ma...@gn...>]

Hi Mitra,

in fact, the emails are not (should not) being send immediatelly as you p=
ost a new item. This would be causing delays for editors
which we wanted to avoid.
Instead, there is a cron entry for "instant" sending that runs by default=
 every minute.

However, it used to be like that for some time in the past. He have just =
discovered  that StoreItem is still calling obsolete
AlertsSentInstantAlert.

You can comment it out in include/Event_Handler.php3 or it's definition i=
n modules/alerts/event.php3 or update from cvs

But anyway, I thing there may be something up with the communication to y=
our local mail delivery agent. Could you do some debugging
to find out where is the bootleneck ? Maybe write a simple script that us=
es php mail function and test how long it takes to send out
60 emails. I don't thing that building new email for each user is cloggin=
g it to such extent.

But for now, increase script_timeout for your php scripts.
And increase the value in cron so that "instant" script doesn't run that =
often - can be 5 minutes or so.

Best

  Marek



> -----Original Message-----
> From: apc...@li...
> [mailto:apc...@li...]On Behalf Of Mitra
> Sent: 1. =E8ervence 2003 10:43
> To: apc...@so...
> Subject: [Apc-aa-coders] Timeouts in Alerts with only 60 users.
>
>
> I just noticed a nasty problem in Alerts.
>
> I have a new list - with about 65 members on it, all set to "Instant".
>
> When I add an item it sits there for a LONG time, and eventually
> times out, I'm not totally sure why yet, but I think it is because
> send_emails is doing a lot of work to figure out each reader, and/or
> is then calling send_email_from_table building a new email for each
> user.
>
> I think it has to do this because of the confirmation code at the
> bottom of every email.
>
> The worst thing is that after sending, because it times out it
> doesn't mark the message as sent, so the next cron task sends it
> again, so I've got half of my list of members who got it twice, and
> half didn't get it at all :-(
>
> Any ideas on fixing this.
>
> - Mitra
>
> --
> Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
> home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182
>
> Life is a Mystery to be Lived, not a Problem to be Solved
>
>
>
> -------------------------------------------------------
> This SF.Net email sponsored by: Free pre-built ASP.NET sites including
> Data Reports, E-commerce, Portals, and Forums are available now.
> Download today and enter to win an XBOX or Visual Studio .NET.
> http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/0=
1
> _______________________________________________
> Apc-aa-coders mailing list
> Apc...@li...
> https://lists.sourceforge.net/lists/listinfo/apc-aa-coders
>

[Apc-aa-coders] Timeouts in Alerts with only 60 users.

[Mitra <mi...@ea...>]

I just noticed a nasty problem in Alerts.

I have a new list - with about 65 members on it, all set to "Instant".

When I add an item it sits there for a LONG time, and eventually 
times out, I'm not totally sure why yet, but I think it is because 
send_emails is doing a lot of work to figure out each reader, and/or 
is then calling send_email_from_table building a new email for each 
user.

I think it has to do this because of the confirmation code at the 
bottom of every email.

The worst thing is that after sending, because it times out it 
doesn't mark the message as sent, so the next cron task sends it 
again, so I've got half of my list of members who got it twice, and 
half didn't get it at all :-(

Any ideas on fixing this.

- Mitra

-- 
Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182

Life is a Mystery to be Lived, not a Problem to be Solved

Re: [Apc-aa-coders] Hard coded "DropIn" stuff in events

[Mitra <mi...@ea...>]

Can I suggest we come up with a convention for site-specific code in 
distributions.

For example it could be by

if (DEFAULT_ORG_ID == "foo.com") {.... }

Better would be so that the code was never compiled, but I'm not sure 
how to do that in PHP.

I think this would be useful as often these site-specific extensions 
allow others to see how something has been done and even generalise 
it?

- Mitra


At 2:41 AM +0200 27/6/03, Honza Malik wrote:
>Hi Mitra,
>
>    in current CVS version we introduced event.class.php3 file - new
>event handling possibility in AA. It works well, I think, but the events
>are still defined inside the file, which is not final state - we plan to
>move event handlers definitions to the database.
>
>    So yes, the code is still needed (it will be never executed for you
>because it is joined with slice_id), but it will be moved to the
>database in the future.
>
>    Honza
>
>
>26. 06. 2003 v 02:16, Mitra wrote:
>
>>  Jakub ...
>>
>>  I notice in include/event.class.php3 there is some hard coded events
>>  set up by a function Event_ItemUpdated_DropIn
>>
>>  Is this something left over from testing, or is it real code still needed?
>>
>>  it appears to still be getting called (from aaevent->get_handlers
>>  from aaevent->comes  from several places)
>>
>>  Its not important ... just something odd I noticed.
>>
>>  - Mitra
>>
>
>
>
>-------------------------------------------------------
>This SF.Net email is sponsored by: INetU
>Attention Web Developers & Consultants: Become An INetU Hosting Partner.
>Refer Dedicated Servers. We Manage Them. You Get 10% Monthly Commission!
>INetU Dedicated Managed Hosting http://www.inetu.net/partner/index.php
>_______________________________________________
>Apc-aa-coders mailing list
>Apc...@li...
>https://lists.sourceforge.net/lists/listinfo/apc-aa-coders


-- 
Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182

Life is a Mystery to be Lived, not a Problem to be Solved

Re: [Apc-aa-coders] Hard coded "DropIn" stuff in events

[Honza M. <hon...@ec...>]

Hi Mitra,

   in current CVS version we introduced event.class.php3 file - new
event handling possibility in AA. It works well, I think, but the events
are still defined inside the file, which is not final state - we plan to
move event handlers definitions to the database.

   So yes, the code is still needed (it will be never executed for you
because it is joined with slice_id), but it will be moved to the
database in the future.

   Honza


26. 06. 2003 v 02:16, Mitra wrote:

> Jakub ...
> 
> I notice in include/event.class.php3 there is some hard coded events 
> set up by a function Event_ItemUpdated_DropIn
> 
> Is this something left over from testing, or is it real code still needed?
> 
> it appears to still be getting called (from aaevent->get_handlers 
> from aaevent->comes  from several places)
> 
> Its not important ... just something odd I noticed.
> 
> - Mitra
> 

[Apc-aa-coders] Hard coded "DropIn" stuff in events

[Mitra <mi...@ea...>]

Jakub ...

I notice in include/event.class.php3 there is some hard coded events 
set up by a function Event_ItemUpdated_DropIn

Is this something left over from testing, or is it real code still needed?

it appears to still be getting called (from aaevent->get_handlers 
from aaevent->comes  from several places)

Its not important ... just something odd I noticed.

- Mitra


-- 
Mitra Technology Consulting -  www.mitra.biz - mi...@mi...
home office 02-6684-8096  mobile 0414-648-722  work 02-6684-6182

Life is a Mystery to be Lived, not a Problem to be Solved

[Apc-aa-coders] Prospecting Sales Tip of the Week

[Events C. <in...@sa...>]

Change Your Objective and Change your results. 
Most salespeople have a prospecting "pitch" which is designed to interest, entice, excite, convince and persuade people to give them appointments. They have a string of questions to get the prospect involved and interested in meeting them. Every question that you ask increases their sales resistance. Thus, their negative reaction to your methods grows very quickly. 
Ask yourself this question: When you buy something, do you buy for YOUR reasons or some sales person's reasons? 
You buy for your reasons! 
Therefore START to uncover the prospect's reasons and stop forcing your own reasons on them!
Your new objective is to make appointments only with Prospects who have valid reasons to meet with you-and to disqualify everyone else preventing the time wasted when chasing non-valid prospects. How to uncover the real reasons for buying is the real key! 

For solutions to the above problem and other self created sales barriers please visit http://www.sales-help.org/w_july.htm

To learn more about how to uncover the real reasons prospects buy and to change your objectives and bottom line please visit: http://www.sales-help.org
 
For information on upcoming sales related workshops and other events please visit: http://www.sales-help.org/workshops.htm. 

Sincerely,
 
Peak Performance Sales Training, LLC 
in...@sa...


We sincerely apologize if you received this email in error.  To be automatically removed from our list please send a blank email to re...@sa....  NOTE: For automatic removal, the originating email must be the same as the address you wish to remove.  In the event an alias received the email please reply to this email and INCLUDE the email address you wish to be removed in the BODY of the email.  Please note that these requests are removed manually and take longer to remove.  Thank you.