API / Apache Velocity Interview questions
Explain the lifecycle of a Velocity template from load to render?
Rendering a template goes through a consistent sequence of stages inside the engine, whether it's the first request or the thousandth.
flowchart TD
A[getTemplate name called] --> B{Cached and unmodified?}
B -- Yes --> F[Return cached parsed Template/AST]
B -- No --> C[ResourceLoader reads raw .vm file]
C --> D[Parser (generated by JavaCC) builds an AST of nodes]
D --> E[Template object wraps the AST and is cached]
E --> F
F --> G[template.merge context, writer called]
G --> H[Visitor walks the AST]
H --> I[References resolved against Context via Uberspector]
I --> J[Directives (#if/#foreach/#macro) executed]
J --> K[Output streamed to the Writer]
The expensive work, reading the file and building the AST, only happens once per template as long as caching is enabled and the source file's timestamp hasn't changed; every subsequent render reuses that cached AST and just walks it against a fresh Context. This is why keeping file.resource.loader.cache enabled matters so much for throughput: without it, the parse step repeats on every single render.
What is cached after a template is first loaded and parsed, assuming caching is enabled?
During template.merge(), what resolves each $reference against the supplied data?
More Related questions...
