SShortSingh.
TopTechnologyAI & MLStartupsProgrammingScienceBusinessFinanceCrypto & Web3World
Back to feed
Programming

How Stale Search Responses Break Accessible Comboboxes and How to Fix Them

0
·1 views

Accessible comboboxes built to ARIA standards can still fail when asynchronous search responses arrive out of order, leaving aria-activedescendant pointing to options that no longer exist in the DOM. A common trigger is rapid typing, where a slower earlier query overwrites results from a faster later one, breaking keyboard navigation for screen reader users. IME composition adds further risk, as partial keystrokes can dispatch network requests before the user has finished entering text. Developers can guard against this by tracking a request generation counter and discarding any response that is not the most recent, then restoring the active option by a stable key rather than an array index. Playwright tests and manual screen-reader passes are recommended to verify focus integrity, valid aria-activedescendant references, and correct announcement suppression for discarded responses.

Read the full story at DEV Community

This is an AI-generated summary. ShortSingh links to the original source for the complete article.

Discussion (0)

Log in to join the discussion and vote.

Log in

Related stories

0
ProgrammingDEV Community ·

Developer Builds Privacy-First Chrome Extension to Autofill Job Application Forms Locally

A developer has built a Chrome extension called Job Form Autofill that automatically fills out job application forms with a single click, with all data stored exclusively in the browser's local storage. The tool was created to address the repetitive nature of job hunting, where applicants must repeatedly enter the same personal details across hundreds of different company-specific forms. Unlike many existing autofill tools, this extension never transmits user data to external servers, making it suitable for sensitive job-hunting information such as names, addresses, and birth dates. Built using Manifest V3, the extension uses heuristic field inference to identify form fields even when they lack standard naming conventions, a common problem on Japanese recruiting platforms. The extension is currently available on the Chrome Web Store and requires only local storage permissions to function.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Why MCP Servers, Not Chatbots, Define Truly AI-Native Software

A growing critique in developer circles argues that most products labeled 'AI-native' are simply conventional apps with a chat widget bolted on, offering conversation about features rather than actual execution of domain tasks. The core distinction is whether an AI agent can autonomously perform state-changing operations — such as creating invoices or filing returns — not merely summarize or describe them. The Model Context Protocol (MCP) is presented as a standardized solution, defining a structured interface of typed, callable tools that AI agents can discover and invoke directly against live systems. By exposing domain logic as named MCP tools with strict input and output schemas, developers give agents a safe, repeatable surface to operate software without human clicks or bespoke integrations. The argument concludes that true AI-nativeness lives at the domain layer, not the presentation tier, and the test is simple: could an agent still operate the product if the chat widget were removed.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Engineering Team Cuts RAG Pipeline Latency 40% Using Bayesian Search and Hybrid Retrieval

A development team overhauled their retrieval-augmented generation (RAG) pipeline after standard production deployments proved inadequate for legal contracts, API docs, and customer support tickets. The team replaced fixed 512-token chunking with document-specific strategies — including recursive, semantic, and agentic chunking — achieving recall@10 scores between 91% and 97% across content types. They combined vector search with BM25 keyword search using Reciprocal Rank Fusion, then applied a cross-encoder reranker to narrow 50 candidates down to 5, adding only 50ms while boosting recall by 15%. A Bayesian optimization layer was used to tune retrieval parameters, ultimately cutting end-to-end query latency by 40%. The rebuilt pipeline replaced what the team described as a 'semantic search and hope' approach with a measurable, tunable system suited for large-scale production use.

0 comments Read more at DEV Community
0
ProgrammingDEV Community ·

Dart's built-in regex engine vulnerable to ReDoS; RE2 FFI binding offers linear-time fix

Dart's default RegExp engine uses backtracking, which can cause catastrophic slowdowns on certain inputs — a vulnerability known as ReDoS. A 31-character crafted string takes over 5 seconds to reject against a common word-list validator pattern, and some patterns never finish at all. Google's RE2 library, now available as a Dart FFI package, matches in linear time regardless of input, processing the same 31-character case in about 25 microseconds. Unlike dart:core's RegExp, RE2 rejects unsafe patterns such as backreferences at construction time rather than failing silently under load. The package also provides utilities like Re2.escape() and a compiled program size cap to further protect against untrusted input.

0 comments Read more at DEV Community
How Stale Search Responses Break Accessible Comboboxes and How to Fix Them · ShortSingh