Python 2.7 & Runtime Safety — Python 2.7 data and text limits

Python 2.7 data handling

Handle unicode, byte strings, and UTF-8 explicitly.
For source-embedded multilingual UI text, use a UTF-8 coding declaration and from __future__ import unicode_literals. Resolve the client language through helpers.getClientLanguage(), normalize case and -/_, map known aliases such as Ukrainian, Russian-family, Brazilian Portuguese, and Chinese variants, then fall back from regional code to base language and finally English.
Keep every translation dictionary structurally identical to English and validate that all format placeholders can be expanded. Pass already-styled values such as hotkeys, counts, and item names through named placeholders so translations can reorder them naturally.
For binary JSON writes in dual Python 2/3 code, encode text to UTF-8; decode bytes before parsing reads.
Python 2 networking commonly uses urllib2, httplib, BaseHTTPServer, and urlparse. Python 3 fallbacks may be useful for offline tooling, but must not break the client runtime.
Avoid mutable defaults such as config_files=[]; use None and allocate a new collection per call.
When a zero-argument function captured from another class must be stored on a new-style owner class, wrap it with staticmethod(...). Storing the plain function makes later class access apply Python 2 method binding and can inject an unexpected owner argument. Restore it with the same descriptor shape.

Text limits have different units

The external string helpers implement different boundaries. normalized_unicode_trim decodes UTF-8 when needed, applies normalization (NFKC by default), slices the resulting Unicode string, then encodes it. Its length parameter is not a UTF-8 byte budget. truncate_utf8 instead expects already encoded UTF-8 and trims to a byte budget without cutting a valid multibyte sequence; it does not normalize or validate the complete input.

Input and limit	Normalized string trim	Encoded byte trim
UTF-8 for `éé`, limit 2	`éé`: four bytes	`é`: two bytes
UTF-8 for the ligature `ﬃ`, limit 2	`ff`, after normalization expands the ligature	Empty prefix: the encoded ligature needs three bytes

The outgoing text filter uses normalized trimming and then collapses whitespace; a request context uses the encoded-byte helper for its comment field. Preserve the chosen unit and processing order when adapting a boundary. Neither helper promises grapheme-based trimming or establishes server acceptance.

Selected readable helpers were checked under Python 3 using synthetic encoded strings and an adapter for Python 2 byte-index values. Valid multibyte prefixes, normalization-before-slicing and the differing output sizes were checked. No messages, requests, native APIs, installed bytecode or Python 2.7 client runtime were executed. Unicode-string slicing must not be generalized into one codepoint-count rule across narrow and wide Python 2 builds.