generating filepro HTML code using clerk !?

Bruce Easton bruce at stn.com
Tue Mar 13 09:36:12 PDT 2012


(inline response - Bruce)

On 3/13/12 11:29 AM, Henry Arredondo wrote:
> I'm currently using filePro's "HTML :id" code to generate HTML files (webpages) by running the "report file command" using index search "-i.." and it works really well and fast, but when the file gets hit real hard by many web surfers each new hit makes the report run slower and this drags all the rest running hits to a halt:
>
>
> 1)       If I stop using "report" and move it to "clerk" would the filePro's HTML code still work ?
Maybe.  Obviously our code will be triggered differently.  I rely on 
@menu for most output for browser from *clerk.  If you are using *report 
to produce web reports (and not input forms), then, although you may be 
following the same automatic index (as you've been doing with -i), 
you'll have to consider, under *clerk, how to achieve the same record 
selection and whatever final output sorting and sub-total breaking 
you've been using (where *report helps a lot with its @wbrk triggers, 
the 1-8 level sorting spec, and the ability to override any portion of 
that in selection processing).
>
>
> 2)       Is "report" mandatory to generate the HTML file using HTML :id commands ?
No.
>
>
> 3)      Would using clerk be faster that report ?
Hard to say.  I think it would really depend on the specific task.

<dream>I wish there was a 'for browser' version of *clerk (or a way to 
flag that the call is for that use), so that all of the memory and 
processing time dedicated to screen output could be eliminated.  It 
could even be dedicated to only run as @menu (so all code related to 
@when-field, @key events, @entsel, @update would be eliminated).  
Because all that is needed for that type of *clerk session is to be able 
to do lookups, have variables, call other similar processing, read & 
write files, plus the current set of commands and functions other than 
those relating to screen io. Ideally, it would be better to have a 
specific type of *clerk for this purpose rather than calling *rclerk 
with some new flag, since I suspect, removing all the unneeded code 
would reduce the size of the *clerk executable quite a bit. </dream>

>
> I have been splitting the Data in different qualifiers and this distributes the hits and speed things up.
>
> I'm using SCO Unix filepro 5.0.14.
>
> Thanks
>
> ________________________________
> Please consider the environment before printing
> -------------- next part --------------
> An HTML attachment was scrubbed...
> URL: http://mailman.celestial.com/pipermail/filepro-list/attachments/20120313/c46184f3/attachment.html
> _______________________________________________
> Filepro-list mailing list
> Filepro-list at lists.celestial.com
> Subscribe/Unsubscribe/Subscription Changes
> http://mailman.celestial.com/mailman/listinfo/filepro-list
>
>



More information about the Filepro-list mailing list