Fastcgi and persistence in sub modules

[Anonymous Monk]

Dear monks,

In the context of web scripting, say I've a main fastcgi-enabled script that imports Module A (before the "while" loop). Module A in turn imports Module B and Module C.

Since Module A is loaded before the "while" loop in the main script, it's saved in memory. But what about Module B and Module C? Do these also become persistent as a result of loading Module A at the start of the script?

If not, is it a good practice to load them in the main script at the same time Module A is loaded i.e. before the "while" loop?

Looking forward to your replies and many thanks in anticipation.

[zwon]

If you have package A that looks like:

package A;
use B;
use C;
...

and the script

use A;
while(1) {...}

Then modules A, B, and C will be loaded during script compilation time. Even if you place use A inside the while loop or after it (see use). If it is a good practice or not depends on your particular conditions. If you have pre-forking server running on some sort of Unix it is advisable to preload modules as it will reduce total memory usage due to COW.

[Anonymous Monk]

Exactly as you have shown. Thank you so much.

[MidLifeXis]

IMO, if your script only depends on the functionality provided by module A, then only load module A. If your script peeks behind the covers of Module A and loads Module B and Module C from within the script, you have now made your script dependent on modules A, B, and C, and not just A. If you change the implementation of module A to use modules D and E instead, you now have to maintain the script as well.

Only load what you explicitly depend on.

From a persistence standpoint, fastcgi will have the dependencies also loaded.