Phantasmal MUD Lib for DGD

Phantasmal Site
Phantasmal API
Archive
DGD
Getting DGD
Game Design Issues
The Kernel Library
LPC
Writing a Library
Changing LPC
Object Cleanup
⁄cmd vs wiztool
Conditional Inheritance
Object Database
Distribute a State Dump?
Designing Driver Objs
FTPD
GurbaLib
HTTPD
State Dumps
Heaven7
InterMUD3
Interrupt Call
Designing AUTO Objs
Misc Issues
NFS
Object Binding
Object Management
Other Net Services
Persistent MUDLibs
Player vs Body
Precompiling
Reimplementing From Scratch
Releasing Code
Rlimits
Script Delays
Misc Security
Outgoing Email
So You Want To...
Start from Scratch?
Supplementary Documentation
Telnet Protocol
Using the Kernel
Using Melville
Using Phantasmal
WebDAV
Which License
What Does It Do?
DGD LPC Reference
Running a MUD
Skotos
CSharp vs DGD
Contributing to DGD
DGD Glossary
Java vs DGD
DGD MUDs
Miscellaneous DGD
MudOS vs DGD
Slush Bucket
Why Use DGD?
Design
Development
Innsmouth MUD
Phantasmal Operation
Setup
Test module index
Phantasmal Tutorials
Comparison to Other Libs
Credits
Current Features
History
Installing Baseline Phantasmal
About

Precompiling LPC Objects

If you're planning to precompile LPC objects, take a good look at dgd/src/lpc/Makefile in the DGD driver source. There are some commented-out lines in there that you can uncomment (while commenting others). Doing so will precompile the kernel library if you compile DGD. It's fairly easy to figure out.

If you're looking to precompile for the speed advantage, measure the code performance first. Using a mostly-hardcoded tool in LPC like parse_string, implode(), or explode() can vastly improve your performance. They're often better than a precompiling your LPC, and they're far more portable. Measure, measure, measure.

Similarly, before moving functionality into the driver itself and requiring all your users to run a patched DGD, first try an LPC solution, and then a precompiled LPC solution. You're not likely to get even a two times difference between one and another. DGD is already quite efficient, even when interpreted. So when in doubt, measure. When optimizing, optimize your algorithm first. It will make much, much more difference than optimizing what your code is compiled into.