Do "ALL" session vars expire at the same time?

[Fulltimer]

I have several locations within my app which use eval_valid() to test the validity of certain session vars. Each validity check determines whether certain menu options or other user options or events are made available.

For some reason some session vars seem to pass the eval test while my main session var seems to have expired. For example if my main session var is assigned at the opening of the app ( session.main_office_id ). Maybe ten minutes later, I assign the vars ( session.did_printletter_1 ) and (session.have_selected_color).

Shouldn't eval_valid() return true for session.main_office_id WHENEVER eval_valid() returns true for session.have_selected_color ?

This only seems to be a concern when the app is allowed to sit "idle" for some time (maybe 60 minutes etc - my setting).

In other word's - Do "ALL" session vars expire at the same time?

Also, under what circumstances would my a5 code show up in my browser window literally-
<%a5 ? eval_valid("session.main_office_id") %>

This kind of code exposure seems fairly reckless and dangerous to the developer. Any thoughts?

[Steve Wood]

Re: Do &quot;ALL&quot; session vars expire at the same time?

You may be getting false readings due to how your application is configured, how you set session variables. The best way to determine what Vars exist is to create an A5W page with ONLY the following:

<%a5
?a5w_info()
%>

and fire that page to view all session vars that exist at the time of your test.

All session vars are deleted when the session expires, but depending on your configuration, some of them can be recreated the moment you refresh a page in your application. The purpose of setting the above function on a blank A5W page is to ensure the page is not also setting other session vars when it is refreshed.

Also, under what circumstances would my a5 code show up in my browser window literally-
<%a5 ? eval_valid("session.main_office_id") %>

This kind of code exposure seems fairly reckless and dangerous to the developer. Any thoughts?

It may appear if your page was not named with the A5W extension. It would appear if you were opening the HTML or A5W page directly in a browser with Alpha Five WAS not running.

[Fulltimer]

Re: Do &quot;ALL&quot; session vars expire at the same time?

All session vars are deleted when the session expires, but depending on your configuration, some of them can be recreated the moment you refresh a page in your application.

Steve,

Thanks for your verbose reply to my question, however it inadvertently may be pointing out the "potential problem" that I thought I may be experiencing.

Your statement:

"some of them can be recreated the moment you refresh a page in your application" is what is at issue.

Let me try to explain:

If session.var1 and session.var2 are 20 minutes "old" and a new session var is NOW created, do ALL the session vars become the "SAME AGE". In other words, are session.var1 and session.var2 suddenly "refreshed" with new life because a server event has taken place? (ie. no new values are stored to these vars at this event)

Tom W.

[Fulltimer]

Re: Do &quot;ALL&quot; session vars expire at the same time?

It may appear if your page was not named with the A5W extension. It would appear if you were opening the HTML or A5W page directly in a browser with Alpha Five WAS not running.

Steve,

On the second part of my question, here is added info.

I have literally seen pages rendered where a grid or grids where displayed properly, YET my <%a5 .... %> code was ALSO written to the page in plain text.

Obviously the WAS must be running if the grids displayed. This only occurs when some session var has expired or event has taken place that is rare but is repeatable. All I have to do is go to a page deep within my app and let the app just sit there for 4 hours or more. If I then come back home and click on a button on that "viewable" page, the action may cause a partial display of the good with the bad.

Tom W.

[Steve Wood]

Re: Do &quot;ALL&quot; session vars expire at the same time?

Tom,

You may have to show a video or share the at issue grid and provide an example. Exactly where is that code? It must be in some freeform area, field labor or such, or perhaps a text area field as part of a table record.

[Steve Wood]

Re: Do &quot;ALL&quot; session vars expire at the same time?

If session.var1 and session.var2 are 20 minutes "old" and a new session var is NOW created, do ALL the session vars become the "SAME AGE". In other words, are session.var1 and session.var2 suddenly "refreshed" with new life because a server event has taken place? (ie. no new values are stored to these vars at this event)

Well, to "create" a new session variable you have to hit the server. So in an indirect way, Yes, if you create a new session var you will have "refreshed" the session and all variables will have a new lifespan equal to the session timeout set on the server config.

[RichCPT]

Re: Do &quot;ALL&quot; session vars expire at the same time?

In V12 I have been noticing code that I thought used to work with eval_valid() for testing variable existence wasn't working as I thought it did under V11. I have been switching eval_valid() out with "variable_exists()". It think it works better for that purpose and Alpha says it is a lot more efficient.

I keep the timeout setting pretty low on my development application server, so I can see what happens readily on session timeout. Timeout is a huge problem because of the number of things that continue to work, perhaps due to cached pages and components. When component security is turned on, isn't there a way to configure the application server to redirect to a special page when it receives a request from a timed-out session?

On another note about session variables - I do not see why Alpha continues to demonstrate their use as a means to pass temporary information from one component to the next in some of their training videos. If the user can open two or more instances of the same component, like if everything is opened in a tab on the tabbed-UI, then the session variables are useless unless you make certain you capture and use the session variable's value as it existed when that instance was opened AND you take measures to ensure the session variable really is re-initialized and not using an old value from the cached component.
What Alpha has done in the Action Javascript for calling a Dialog2/UX component where "arguments" can be defined dynamically and then picked up in the OnDialogInitialize with "argval(e.arguments, "My_Argument")" works well for me, so far. Now if they would just make that work when calling a Grid component!

-Rich F.

[TheSmitchell]

Re: Do &quot;ALL&quot; session vars expire at the same time?

In V12 I have been noticing code that I thought used to work with eval_valid() for testing variable existence wasn't working as I thought it did under V11. I have been switching eval_valid() out with "variable_exists()". It think it works better for that purpose and Alpha says it is a lot more efficient.

A follow up to efficiency, Alpha doesn't encourage using eval_valid() for testing session variables for performance reasons. They suggested instead that you assign your session variable to a local variable and test for null:

dim test as c = session.some_var

if (.not. isNull(test)) then
  ' do things
end if

[Fulltimer]

Re: Do &quot;ALL&quot; session vars expire at the same time?

A follow up to efficiency, Alpha doesn't encourage using eval_valid() for testing session variables for performance reasons. They suggested instead that you assign your session variable to a local variable and test for null:

dim test as c = session.some_var

if (.not. isNull(test)) then
  ' do things
end if

Sarah,

I tried your code as written and found it to be a non workable example.

The reason is, if session.some_var is not yet valid. an error will occur: "subelement session.some_var not found".

Therefore unless I'm missing something, you can not use this method to test for a valid session var. I have always used eval_valid() but from the comments above, I am now trying variable_exists(). Don't know technically what the difference is, but I'd like to use the best function available.

Tom W.

[Jim Coltz]

Re: Do &quot;ALL&quot; session vars expire at the same time?

Another option, one that I use quite a bit, is to use the DEFAULT keyword and dim the variable in your code.

DIM session.myvar as c = DEFAULT ""  'if it exists, nothing happens, if not, it gets created with value you supply

If session.myvar = "" then
	'session.myvar has not been set from somewhere else
else
	'session.myvar has been set from somewhere else	
end if

[Peter.Greulich]

Re: Do &quot;ALL&quot; session vars expire at the same time?

variable_exists() is the "recommended" method last I heard from Jerry Brightbill.

[TheSmitchell]

Re: Do &quot;ALL&quot; session vars expire at the same time?

Sarah,

I tried your code as written and found it to be a non workable example.

The reason is, if session.some_var is not yet valid. an error will occur: "subelement session.some_var not found".

Therefore unless I'm missing something, you can not use this method to test for a valid session var. I have always used eval_valid() but from the comments above, I am now trying variable_exists(). Don't know technically what the difference is, but I'd like to use the best function available.

Tom W.

Yep. It was an example. If your session variable isn't defined, accessing it causes an error. You'll have to replace "some_var" with whatever your session variable is named - assuming it exists. If it doesn't exist, your back to testing session using eval_valid or variable_exists. Though, I like Jim's suggestion to use DEFAULT. I never thought to do that.