Feature request: log level support for Pawn scripts (print/printf)

Summary

Expose log severity levels (e.g. Info / Warning / Error / Debug) to Pawn scripts, instead of the current single-level print() / printf().

Context

The core server already emits leveled log lines internally, visible in the console output, for example:

[Info] OI
[Error] Unable to start legacy network on port 7777. Port in use?

That [Info] / [Error] prefixing exists in the engine/component side, but it is not reachable from gamemode/filterscript Pawn code. From Pawn, the only natives available are print() and printf() (see omp_core.inc), which write a plain, unleveled line to both the console and log.txt. Any severity distinction today is a script-side convention (developers manually prefixing strings like "[WARNING] ..."), which is inconsistent across gamemodes and doesn't hook into anything the server already understands.

Proposal

Add Pawn-side natives that reuse the same leveled logging the engine already has, for example:

native LogInfo(const format[], {Float,_}:...);
native LogWarning(const format[], {Float,_}:...);
native LogError(const format[], {Float,_}:...);
// or a single native: native Log(LOG_LEVEL:level, const format[], {Float,_}:...);

Each would print with the same [Info] / [Warning] / [Error] prefix (and console colour, where supported) used internally, and write to log.txt the same way printf() does today.

What this would gain us

• Faster tracing/debugging: grepping log.txt for [Error]/[Warning] immediately isolates real problems from routine info output, instead of scrolling through everything printf() ever wrote.
• Consistency between engine and gamemode logs: right now the server's own startup/network messages are leveled, but everything a gamemode prints is flat text, which is a visible inconsistency in the same log file.
• No more ad-hoc conventions: every gamemode currently invents its own prefix scheme (or none at all) to fake severity; a native solution removes the need for that boilerplate and keeps output uniform across projects.
• Enables tooling downstream: leveled, parseable log lines make it realistic to build log filtering, alerting, or dashboards (e.g. tailing log.txt and reacting only to [Error] lines) without fragile string-matching against whatever prefix a given script author chose.
• Optional filtering by level: could pair naturally with a config.json setting (e.g. minimum log level) to reduce console/file verbosity in production without touching gamemode code.

This is purely additive: print()/printf() would keep working exactly as they do now.

Start by reading the Pawn natives in omp_core.inc and tracing how print() and printf() reach the core server's existing leveled logging. Compare the proposed separate natives with the single-level alternative, then verify that the chosen API preserves current output behavior while producing consistent severity prefixes in the console and log.txt.