# Android string-scanning bug breaks apps handling non-ASCII text

_Published Wednesday, October 14, 2026 at 3:06 AM EDT · Products · Latest · Tier 2 — Notable_

![Description On Android 17 builds from September 2026, wcscspn() and wcspbrk() stop at the first character of the string that is >= U+0100 whenever the reject set has two or more characters that are... — Primary](https://opengraph.githubassets.com/faa19897607d84aad613f66ea165a6363ff4d54160d84554507628d175f561b1/android/ndk/issues/2260)

An Android bug report identifies a regression in two string-scanning functions that can break applications handling emoji and other non-ASCII text. A response in the Android NDK issue tracker acknowledges the diagnosis, says the issue persists in the latest development code, and proposes additional tests and a backport of the fix.

The functions incorrectly treat characters at or above U+0100 as matches when searching for two or more delimiters below U+0100. The reporter traced the behavior through disassembly of the shipped library because public source code differed from September builds. Visual Pinball stopped loading tables with emoji in VBScript comments; its folder-moving code also misclassified affected paths as wildcard patterns.

## Sources

- [GitHub](https://github.com/android/ndk/issues/2260)

---
Canonical: https://techandbusiness.org/newswire/fuj7ebUPp2dVwjrMjMoyxb
Published: 2026-10-14T07:06:18.262Z
Story chronology: 2026-10-14T06:18:01.000Z
Retrieved: 2026-10-14T09:03:30.091Z
Publisher: Tech & Business (techandbusiness.org)