strings: add lorum Markov chain text generator

[mike-ward]

This PR integrates lorum into vlib/strings, refining the implementation to be suitable for the standard library.

Changes:

• Removes rand dependency: Implements a simple internal LCG (Linear Congruential Generator) to handle pseudo-randomness. This prevents circular dependencies (e.g., strings -> rand -> time), which is critical for vlib core modules.
• Deterministic by default: The generator now defaults to a fixed seed if none is provided, ensuring reproducible output for tests and simple calls.
• Renames API: Renames lorem_generate() to lorum() for consistency with the file name and conciseness.
• Updates Example: Refactors examples/lorem.v to use the new strings.lorum() API. The example now handles random seeding explicitly using rand and time to maintain the expected CLI behavior (random output by default).

Verification:

• v test vlib/strings/lorum_test.v passes.
• v run examples/lorem.v runs successfully and produces varied output.

[JalonSolov]

Just curious... why change lorem to lorum ?

[mike-ward]

Just curious... why change lorem to lorum ?

That was a goof. Fixed it up in the most recent commit.

[spytheman]

Suggested change

	t := time.now().unix_milli()
	rand.seed([u32(t), u32(t >> 32)])
	rng_seed = rand.int()
	rng_seed = rand.int()

[spytheman]

the default random number generator is already time seeded by default

[spytheman]

Suggested change

	module lorem
	import strings.lorem

[spytheman]

Suggested change

	module lorem
	import strings.lorem

Turning this into a normal (non-internal) test, has the benefit that all of the tests also become usage examples, and as a user, you can also more easily see, what the public (and well tested) APIs are.

[spytheman]

afaik only lorem_vary and lorem_tokenize are tested here, and are private?

[spytheman]

Excellent work.
@mike-ward thank you 🙇🏻 .

[mike-ward]

Thanks guys. I thought I had done my due diligence to make this easy to merge. I'll do better next time...