Fix: Use unsigned integer reading methods for PUSHDATA opcode lengths

[ALiberalVoluntarist]

Fix PUSHDATA opcode length parsing

Problem

The Script._build_chunks method was using read_int to parse the length bytes for
PUSHDATA1, PUSHDATA2, and PUSHDATA4 opcodes. Since these values should always be treated
as unsigned integers, using a method that interprets them as signed could result in
negative values when the high bit is set. This would cause incorrect parsing of the
script chunks, potentially leading to runtime errors or incorrect script execution.

Solution

This PR modifies the _build_chunks method to use the appropriate unsigned integer
reading methods:

• read_uint8() for PUSHDATA1
• read_uint16_le() for PUSHDATA2
• read_uint32_le() for PUSHDATA4

This ensures that length values are always treated as unsigned, allowing for
the correct handling of large data chunks.

Testing

Added comprehensive tests that verify:

1. Proper handling of PUSHDATA opcodes with large length values
2. Correct parsing of edge cases (length 0, incomplete length specification)
3. Serialization/deserialization consistency with various PUSHDATA operations

The tests ensure that scripts with PUSHDATA opcodes correctly maintain their integrity
through serialization and deserialization, and that chunk parsing works as expected.Retry

Checklist:

• [ x] I have performed a self-review of my own code
• [ x] My changes generate no new warnings
• [ x] I have run the linter

[voyager1708]

This pull request introduces several improvements and fixes.