# code search

Code search is the practice and tooling for searching, browsing, and understanding source code, including finding files, symbols, definitions, usages, and code history.

This is one page of public article previews, not the complete archive. Follow Next page to continue. Summaries are not the original full articles.

## Your Coding Agent Can Read the Code. Give It the History with Relate by Neo4j

DevFeed: [Your Coding Agent Can Read the Code. Give It the History with Relate by Neo4j](<https://devfeed.tech/articles/your-coding-agent-can-read-the-code-give-it-the-history-with-relate-by-neo4j-144105.md>)

Original publisher: [Read original article](<https://neo4j.com/blog/developer/your-coding-agent-can-read-the-code-give-it-the-history-with-relate-by-neo4j/>)

Author: Kate Smith

Published: 2026-11-18T21:00:00Z

Content type: tutorial

Language: en

Sources: [Neo4j](<https://devfeed.tech/sources/graph-database-technology-neo4j-blog.md>)

Topics: [Neo4j](<https://devfeed.tech/topics/neo4j.md>), [code search](<https://devfeed.tech/topics/code-search.md>), [Open Source](<https://devfeed.tech/topics/open-source.md>), [codex](<https://devfeed.tech/topics/codex.md>), [MSP MCP](<https://devfeed.tech/topics/msp-mcp.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agent-skill](<https://devfeed.tech/tags/agent-skill.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [api-key](<https://devfeed.tech/tags/api-key.md>), [cli](<https://devfeed.tech/tags/cli.md>), [coding](<https://devfeed.tech/tags/coding.md>), [coding-agent](<https://devfeed.tech/tags/coding-agent.md>), [developer](<https://devfeed.tech/tags/developer.md>), [extension](<https://devfeed.tech/tags/extension.md>), [git](<https://devfeed.tech/tags/git.md>), [graph](<https://devfeed.tech/tags/graph.md>), [graph-database](<https://devfeed.tech/tags/graph-database.md>), [history](<https://devfeed.tech/tags/history.md>), [load](<https://devfeed.tech/tags/load.md>), [neo4j](<https://devfeed.tech/tags/neo4j.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [relate](<https://devfeed.tech/tags/relate.md>), [repository](<https://devfeed.tech/tags/repository.md>), [security](<https://devfeed.tech/tags/security.md>), [source](<https://devfeed.tech/tags/source.md>)

### AI overview

Relate combines a repository's Git history and working tree in one graph, letting coding agents find co-changed files, prior change patterns, and contributors that source code alone may not reveal. In the FastAPI example, `api_key.py` repeatedly changed with three security modules that do not reference it; all three tested agents found those siblings with the graph, while none did from source code alone. The post shows importing history with `git-to-graph` and file text with `fs-to-graph`, then querying the store through read-only MCP or the CLI. The comparison also found that agents could identify tests and imported files from code, but needed history to identify relevant past contributors. File-text matching for unreferenced pairs is approximate, and `fs-to-graph` defaults to a 64 KB content cap, which can omit large files unless raised.

### Source excerpt

Load a repository's git history into a graph with Relate's git-to-graph extension, serve it read-only, and see what a coding agent finds with it that it cannot find in the source tree. In a popular open source codebase, security/api_key.py has... Read more ->

## Wayfinder: Frontier-level Code Search at a Fraction of the Cost

DevFeed: [Wayfinder: Frontier-level Code Search at a Fraction of the Cost](<https://devfeed.tech/articles/wayfinder-frontier-level-code-search-at-a-fraction-of-the-cost-141039.md>)

Original publisher: [Read original article](<https://blog.voyageai.com/2026/11/18/wayfinder-code-search/>)

Author: Voyage AI

Published: 2026-11-18T15:30:00Z

Content type: article

Language: en

Sources: [Voyage AI](<https://devfeed.tech/sources/voyage-ai.md>)

Topics: [code search](<https://devfeed.tech/topics/code-search.md>), [hybrid-search](<https://devfeed.tech/topics/hybrid-search.md>), [AI-driven information retrieval](<https://devfeed.tech/topics/ai-driven-information-retrieval.md>), [AI-assisted coding](<https://devfeed.tech/topics/ai-assisted-coding.md>), [Frontier AI](<https://devfeed.tech/topics/frontier-ai.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [hierarchical agent systems](<https://devfeed.tech/topics/hierarchical-agent-systems.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [code](<https://devfeed.tech/tags/code.md>), [code-search](<https://devfeed.tech/tags/code-search.md>), [cost](<https://devfeed.tech/tags/cost.md>), [frontier](<https://devfeed.tech/tags/frontier.md>), [grep](<https://devfeed.tech/tags/grep.md>), [news](<https://devfeed.tech/tags/news.md>), [search](<https://devfeed.tech/tags/search.md>), [wayfinder](<https://devfeed.tech/tags/wayfinder.md>)

### AI overview

Wayfinder pairs a small fine-tuned code-search model with a local hybrid index to make agent search cheaper while finding more relevant files. The index combines semantic search (80%) with BM25 (20%); its results are merged and reranked with grep results, then the model runs follow-up greps. On 200 held-out tasks, open models with this approach found about 74% of the right files for roughly $0.01 per search, compared with 76% for Sonnet 5.5 high effort at $0.12 and 78% for GPT-6.1 Sol medium effort at $0.07. Two pooled Wayfinder runs reached 78% for about $0.02. In a 42-task coding evaluation, competitive search methods had similar pass rates; Wayfinder reached 55%, while pooled runs reached 61%, but the sample was too small to establish a clear lead. Indexing also has ongoing costs: the authors estimate a typical repository pays back after about two searches monthly. Wayfinder is an open-source research project, not a MongoDB product; the authors recommend evaluating it on individual codebases.

### Source excerpt

Author: Henry Weller Before a coding agent can fix a bug, it has to find the code that matters. The coding agent industry has generally embraced grep as the way agents search code, but repeated searches with grep can become slow and expensive as codebases grow. For customers building coding agents, we wanted to explore [...]

## Revision-Consistent Code Context for Agents That Edit Their Environment

DevFeed: [Revision-Consistent Code Context for Agents That Edit Their Environment](<https://devfeed.tech/articles/revision-consistent-code-context-for-agents-that-edit-their-environment-139576.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.21616>)

Author: Meng Shi

Published: 2026-11-18T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [AI Context Management](<https://devfeed.tech/topics/ai-context-management.md>), [code search](<https://devfeed.tech/topics/code-search.md>), [Agent Harness](<https://devfeed.tech/topics/agent-harness.md>), [async](<https://devfeed.tech/topics/async.md>)

Tags: [abstract](<https://devfeed.tech/tags/abstract.md>), [after](<https://devfeed.tech/tags/after.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [all](<https://devfeed.tech/tags/all.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [boundary](<https://devfeed.tech/tags/boundary.md>), [call](<https://devfeed.tech/tags/call.md>), [can](<https://devfeed.tech/tags/can.md>), [code](<https://devfeed.tech/tags/code.md>), [code-intelligence](<https://devfeed.tech/tags/code-intelligence.md>), [comparison](<https://devfeed.tech/tags/comparison.md>), [complete](<https://devfeed.tech/tags/complete.md>), [composition](<https://devfeed.tech/tags/composition.md>), [condition](<https://devfeed.tech/tags/condition.md>), [content](<https://devfeed.tech/tags/content.md>), [context](<https://devfeed.tech/tags/context.md>), [contribution](<https://devfeed.tech/tags/contribution.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [environment](<https://devfeed.tech/tags/environment.md>), [revision](<https://devfeed.tech/tags/revision.md>)

### AI overview

The paper proposes a complete-revision publication contract to prevent code-intelligence queries from combining repository facts from different update states. A service must wait, report unavailability, or return a complete old revision before publication; afterward, it returns the complete new revision, with a shared version identity across all views in each response. In Codetrip, which joins source, symbol, and graph facts, this approach produced no unsafe responses in 150 Gin v1.12.0 update trials, compared with two for live-source composition. Journaled, idempotent publication recovered to context equivalent to a clean rebuild in all 300 evaluable trials across frozen Gin and Cobra, with median restart-to-context latency of 77-131 ms. A separate rename probe found mixed context in 0/30 Codetrip trials and 30/30 trials each for CodeGraph and codebase-memory-mcp before clean-full-equivalent convergence. These results are limited to the reported controlled trials.

### Source excerpt

arXiv:2611.21616v1 Announce Type: new Abstract: Editing agents query repositories while simultaneously changing them. Source content, symbols, graph relations, and retrieval state may be maintained by different stores and refreshed at different times. During an update, a successful query can therefore combine facts that never belonged to the same repository state. We call this condition mixed-revision code context. This paper proposes an Agent-visible complete-revision publication contract for composite code-intelligence queries. Before a query-visible publication point, a service must wait, report unavailability, or serve a complete old revision; after that point, it serves the complete new revision. The contract requires a shared version identity and publication boundary across every view returned in one response. We instantiate the method in Codetrip, whose brief operation joins source, symbol, and graph facts. In a controlled comparison on frozen Gin v1.12.0, live-source composition produced two unsafe responses in 150 update trials, whereas complete-revision publication produced none under the same five edit scenarios. We further implement journaled, idempotent publication and terminate the real MCP process at five durable boundaries. All 300 evaluable recovery trials on frozen Gin and Cobra returned complete-new context equivalent to a clean full rebuild, with median restart-to-context latency of 77--131 ms. In a separate rename probe, mixed context occurred in 0/30 Codetrip trials and 30/30 trials for both CodeGraph and codebase-memory-mcp before clean-full-equivalent convergence. The contribution is an Agent-facing safety contract, implementation pattern, and measurement protocol for joining live repository bytes with asynchronously derived facts. Every successful composite response consequently has a testable repository-state meaning, including after interruption.

## Does a Code-Tuned Embedding Model Pay Off? papr-embed-code-1 in Qdrant

DevFeed: [Does a Code-Tuned Embedding Model Pay Off? papr-embed-code-1 in Qdrant](<https://devfeed.tech/articles/does-a-code-tuned-embedding-model-pay-off-papr-embed-code-1-in-qdrant-140947.md>)

Original publisher: [Read original article](<https://qdrant.tech/blog/papr-embed-code/>)

Author: info@qdrant.tech (Andrey Vasnetsov)

Published: 2026-11-18T00:00:00Z

Content type: article

Language: en

Sources: [Qdrant Blog on Qdrant - Vector Search Engine](<https://devfeed.tech/sources/qdrant-blog-on-qdrant-vector-search-engine.md>)

Topics: [code search](<https://devfeed.tech/topics/code-search.md>), [Qdrant](<https://devfeed.tech/topics/qdrant.md>), [Fine-tuning](<https://devfeed.tech/topics/fine-tuning.md>), [quantization](<https://devfeed.tech/topics/quantization.md>), [Benchmark](<https://devfeed.tech/topics/benchmark.md>), [Open Source Models & Datasets](<https://devfeed.tech/topics/open-source-models-datasets.md>)

Tags: [approximate-nearest-neighbor-search](<https://devfeed.tech/tags/approximate-nearest-neighbor-search.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [bert](<https://devfeed.tech/tags/bert.md>), [card](<https://devfeed.tech/tags/card.md>), [code](<https://devfeed.tech/tags/code.md>), [code-search](<https://devfeed.tech/tags/code-search.md>), [data](<https://devfeed.tech/tags/data.md>), [embedding](<https://devfeed.tech/tags/embedding.md>), [embedding-model](<https://devfeed.tech/tags/embedding-model.md>), [embeddings](<https://devfeed.tech/tags/embeddings.md>), [fasttext](<https://devfeed.tech/tags/fasttext.md>), [general](<https://devfeed.tech/tags/general.md>), [hnsw](<https://devfeed.tech/tags/hnsw.md>), [image-search](<https://devfeed.tech/tags/image-search.md>), [knn-algorithm](<https://devfeed.tech/tags/knn-algorithm.md>), [leaderboards](<https://devfeed.tech/tags/leaderboards.md>), [matching](<https://devfeed.tech/tags/matching.md>), [model](<https://devfeed.tech/tags/model.md>), [neural-network](<https://devfeed.tech/tags/neural-network.md>), [open](<https://devfeed.tech/tags/open.md>), [open-weights](<https://devfeed.tech/tags/open-weights.md>), [qdrant](<https://devfeed.tech/tags/qdrant.md>), [qwen3](<https://devfeed.tech/tags/qwen3.md>), [recommender-system](<https://devfeed.tech/tags/recommender-system.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [saas](<https://devfeed.tech/tags/saas.md>), [search](<https://devfeed.tech/tags/search.md>), [simaes-networks](<https://devfeed.tech/tags/simaes-networks.md>), [similarity](<https://devfeed.tech/tags/similarity.md>), [test-case](<https://devfeed.tech/tags/test-case.md>), [top](<https://devfeed.tech/tags/top.md>), [training](<https://devfeed.tech/tags/training.md>), [training-data](<https://devfeed.tech/tags/training-data.md>), [transformer](<https://devfeed.tech/tags/transformer.md>), [vector-search](<https://devfeed.tech/tags/vector-search.md>), [vector-search-engine](<https://devfeed.tech/tags/vector-search-engine.md>), [vectors](<https://devfeed.tech/tags/vectors.md>), [word2vec](<https://devfeed.tech/tags/word2vec.md>)

### AI overview

On code retrieval Papr's papr-embed-code-1 improves over its Qwen3-Embedding base on code it did not train on: average nDCG@10 gains are 0.034 for the 4B model and 0.026 for the 0.6B. Both score slightly below the base on prose. In tests across seven code tasks, 1-bit TurboQuant retained 99.7% of the 4B's float32 score without rescoring and all of it with 2x oversampling; the compressed 4B outscored the 0.6B in float32 using 12x less RAM, though it embeds about four times slower. On roughly one million CodeSearchNet snippets, rescored 1-bit search retained 99.5% of exact-search score at 1.0 ms per query versus 1.8 ms for float32 at hnsw_ef 128. Raising hnsw_ef to 256 improved every setup; language filtering also helped. BM25 alone lagged dense search, and fusion generally hurt, with only negligible gains at low weight. The authors recommend the base model for prose-heavy collections and suggest testing compressed 4B vectors, rescoring, and higher hnsw_ef for code search.

### Source excerpt

If you search code, you've probably asked this: is a code-specific embedding model worth it, or is a strong general one good enough? Papr's papr-embed-code-1 is a good test case. It's Qwen3-Embedding trained further on code retrieval, with open weights in two sizes: 0.6B and 4B. Papr reports top scores on CoIR, a code retrieval benchmark. But it trained on most of CoIR's training data, and its model card says so. Papr also reports the 4B at #1 on MTEB's two held-out code leaderboards, RTEB (Code) and CodeRAG.

## Embeddings vs. Grep: Rethinking the Dense Advantage in Code Search

DevFeed: [Embeddings vs. Grep: Rethinking the Dense Advantage in Code Search](<https://devfeed.tech/articles/embeddings-vs-grep-rethinking-the-dense-advantage-in-code-search-138001.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.20455>)

Author: Siqiao Xue, Ziyin Zhang, Zihan Liao, Jin Qin, Yixiang Mu, Hang Yu

Published: 2026-11-17T05:00:00Z

Content type: news

Language: en

Sources: [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [code search](<https://devfeed.tech/topics/code-search.md>), [hybrid-search](<https://devfeed.tech/topics/hybrid-search.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>)

Tags: [benchmark](<https://devfeed.tech/tags/benchmark.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [bm25](<https://devfeed.tech/tags/bm25.md>), [code](<https://devfeed.tech/tags/code.md>), [code-search](<https://devfeed.tech/tags/code-search.md>), [coding-agents](<https://devfeed.tech/tags/coding-agents.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [embeddings](<https://devfeed.tech/tags/embeddings.md>), [experiments](<https://devfeed.tech/tags/experiments.md>), [go](<https://devfeed.tech/tags/go.md>), [java](<https://devfeed.tech/tags/java.md>), [javascript](<https://devfeed.tech/tags/javascript.md>), [python](<https://devfeed.tech/tags/python.md>), [queries](<https://devfeed.tech/tags/queries.md>), [raw](<https://devfeed.tech/tags/raw.md>), [repositories](<https://devfeed.tech/tags/repositories.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [typescript](<https://devfeed.tech/tags/typescript.md>)

### AI overview

Repo-Premium evaluates file-level code search in more than 91 repositories across six languages, using curated and raw queries and pull requests from 2022-2026. The authors attribute dense retrieval's benchmark advantage partly to evaluation choices favoring it, including chunking files for BM25 and treating edited tests as relevant, and partly to queries that omit rare code identifiers. Adding LLM-generated identifiers to queries ("generate-then-grep") improved BM25 more than dense retrieval and matched it in every Python pool, including on pull requests opened after the LLM's release. The excerpt says no tested method could predict which retriever would win for an individual query. The authors recommend fair lexical baselines, combining dense retrieval with BM25, and adding code terms to queries that lack them.

### Source excerpt

arXiv:2611.20455v1 Announce Type: new Abstract: Code-search benchmarks show dense encoders well ahead of BM25, yet coding agents do well with \texttt{grep}. What does this dense advantage measure? Existing benchmarks cannot tell: they use curated queries, old code and a handicapped BM25. To find out, we build Repo-Premium, a realistic repository-level code-search benchmark. It ranks files in real repositories, with curated and raw queries in Python, JavaScript, TypeScript, Go, Rust and Java, pull requests from 2022 to 2026, and a preregistered test pool over $91$ repositories. Our experiments trace the advantage to two sources: evaluation choices that favor dense retrieval, such as splitting files into chunks for BM25 and counting edited tests as relevant, and queries that lack the code's rare identifiers. Expanding the query with LLM-written identifiers (generate-then-grep) helps BM25 far more than dense retrieval, letting it match dense retrieval in every Python pool, even on pull requests opened after the LLM was released. No method we tried predicts which retriever will win on a given query. From these results we derive three guidelines: measure dense gains against a fair lexical baseline, fuse dense retrieval with BM25, and add the code's words to queries that lack them.

## GraphScent: Calibrating Directional Relevance Signals for Bounded-Context Code Localization

DevFeed: [GraphScent: Calibrating Directional Relevance Signals for Bounded-Context Code Localization](<https://devfeed.tech/articles/graphscent-calibrating-directional-relevance-signals-for-bounded-context-code-localization-135076.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.17527>)

Author: Ramesh Arvind Naagarajan

Published: 2026-11-16T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Artificial Intelligence](<https://devfeed.tech/sources/arxiv-artificial-intelligence.md>), [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [code search](<https://devfeed.tech/topics/code-search.md>), [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [call hierarchy](<https://devfeed.tech/topics/call-hierarchy.md>)

Tags: [ast](<https://devfeed.tech/tags/ast.md>), [bounded-context](<https://devfeed.tech/tags/bounded-context.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [false-positive](<https://devfeed.tech/tags/false-positive.md>), [gemma](<https://devfeed.tech/tags/gemma.md>), [gpu](<https://devfeed.tech/tags/gpu.md>), [language-model](<https://devfeed.tech/tags/language-model.md>), [lexical](<https://devfeed.tech/tags/lexical.md>), [lightweight](<https://devfeed.tech/tags/lightweight.md>), [localization](<https://devfeed.tech/tags/localization.md>), [overlap](<https://devfeed.tech/tags/overlap.md>), [python](<https://devfeed.tech/tags/python.md>), [relevance](<https://devfeed.tech/tags/relevance.md>), [repair](<https://devfeed.tech/tags/repair.md>), [ripple](<https://devfeed.tech/tags/ripple.md>), [runtime](<https://devfeed.tech/tags/runtime.md>), [scopes](<https://devfeed.tech/tags/scopes.md>), [search](<https://devfeed.tech/tags/search.md>), [similarity](<https://devfeed.tech/tags/similarity.md>), [training](<https://devfeed.tech/tags/training.md>), [upstream](<https://devfeed.tech/tags/upstream.md>)

### AI overview

Across 211 Python issues in 86 repositories, Gemma 4 relevance scores were informative for enclosing scopes and upstream callers, while downstream callees, peer methods, historical co-change partners, and dense node embeddings performed near chance. Masked-scope probes and controls for lexical similarity and snippet length suggest the directional signal reflects functional role rather than leakage or surface overlap. GraphScent uses relation-conditioned likelihood ratios on runtime Typed-AST multigraphs, then a listwise comparison; AST Ripple checks after edits for missed co-edits. The authors report that calibration helps smaller models avoid false-positive callers and raises exact localization on lexically hard targets from 12.8% to 23.4% at 31B, within a single-GPU budget.

### Source excerpt

arXiv:2611.17527v1 Announce Type: cross Abstract: Running open-weight language models locally on a single consumer GPU requires replacing expanding multi-turn exploration histories with short, stateless code inspections. Stateless repository localization, however, hinges on a foundational question: when a frozen language model evaluates an unedited code entity from a compact view, which structural relationships make its relevance judgment informative about a nearby repair target? Studying Gemma 4 models across 211 real-world Python issues in 86 repositories, we uncover a systematic directional relevance asymmetry: zero-temperature relevance scores on enclosing scopes and upstream callers consistently separate true structural neighbors from hard lexical distractors, whereas scores on downstream callees, peer methods, historical co-change partners, and dense node embeddings remain near chance. Zero-leakage masked-scope probes and regressions controlling for lexical similarity and snippet length confirm that this asymmetry reflects functional role (callers expose symptom entrypoints delegating to defective routines, while downstream helpers appear benign alone) rather than target leakage or surface overlap. Guided by this finding, we introduce GraphScent, a training-free search controller that propagates out-of-repository, relation-conditioned log-likelihood ratios over runtime Typed-AST multigraphs before resolving ties via a single listwise comparison, alongside AST Ripple, a lightweight post-edit check for missed co-edits. Likelihood calibration prevents overconfident smaller models from chasing false-positive callers while lifting exact localization on lexically hard targets from 12.8 to 23.4 percent at 31B within a single-GPU budget.

## JioHotstar Benchmarks Knowledge Bases for AI Coding Agents

DevFeed: [JioHotstar Benchmarks Knowledge Bases for AI Coding Agents](<https://devfeed.tech/articles/jiohotstar-benchmarks-knowledge-bases-for-ai-coding-agents-129195.md>)

Original publisher: [Read original article](<https://www.infoq.com/news/2026/11/jiohotstar-codingagent-retrieval/>)

Author: Leela Kumili

Published: 2026-11-11T14:28:00Z

Content type: news

Language: en

Sources: [InfoQ](<https://devfeed.tech/sources/infoq.md>)

Topics: [agentic-coding](<https://devfeed.tech/topics/agentic-coding.md>), [code search](<https://devfeed.tech/topics/code-search.md>), [AI-assisted coding](<https://devfeed.tech/topics/ai-assisted-coding.md>), [LLM evaluation / benchmarking](<https://devfeed.tech/topics/llm-evaluation-benchmarking.md>), [knowledge-management](<https://devfeed.tech/topics/knowledge-management.md>)

Tags: [agentic-ai-architecture](<https://devfeed.tech/tags/agentic-ai-architecture.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-coding](<https://devfeed.tech/tags/ai-coding.md>), [ai-coding-agents](<https://devfeed.tech/tags/ai-coding-agents.md>), [ai-ml-data-engineering](<https://devfeed.tech/tags/ai-ml-data-engineering.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [architecture-design](<https://devfeed.tech/tags/architecture-design.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [code](<https://devfeed.tech/tags/code.md>), [code-search](<https://devfeed.tech/tags/code-search.md>), [coding](<https://devfeed.tech/tags/coding.md>), [coding-agents](<https://devfeed.tech/tags/coding-agents.md>), [design](<https://devfeed.tech/tags/design.md>), [development](<https://devfeed.tech/tags/development.md>), [harness](<https://devfeed.tech/tags/harness.md>), [jiohotstar-codingagent-retrieval](<https://devfeed.tech/tags/jiohotstar-codingagent-retrieval.md>), [ml-data-engineering](<https://devfeed.tech/tags/ml-data-engineering.md>), [model-context-protocol-mcp](<https://devfeed.tech/tags/model-context-protocol-mcp.md>), [news](<https://devfeed.tech/tags/news.md>), [requirements](<https://devfeed.tech/tags/requirements.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [retrieval-augmented-generation](<https://devfeed.tech/tags/retrieval-augmented-generation.md>), [search](<https://devfeed.tech/tags/search.md>), [semantic](<https://devfeed.tech/tags/semantic.md>), [software-engineering](<https://devfeed.tech/tags/software-engineering.md>), [tasks](<https://devfeed.tech/tags/tasks.md>), [token](<https://devfeed.tech/tags/token.md>), [tool](<https://devfeed.tech/tags/tool.md>), [tools](<https://devfeed.tech/tags/tools.md>), [trade](<https://devfeed.tech/tags/trade.md>)

### AI overview

JioHotstar's evaluation found no retrieval approach that worked best across all coding tasks. Filesystem tools remained competitive on concrete work; Sourcebot's indexed text and symbol search improved results on harder SWE-bench Verified tasks at lower cost, while retrieval often added overhead on simpler SWE-bench Lite tasks. On CodeScaleBench, semantic retrieval helped with ambiguous requirements that were hard to map through exact text or symbol searches. The team recommends keeping filesystem tools as a baseline, matching structural or semantic search to the task, and measuring tool-use traces alongside accuracy. The excerpt describes tests using Claude Sonnet, subsets of SWE-bench and CodeScaleBench, and an internal question-answering harness; it does not establish a universal ranking.

### Source excerpt

JioHotstar evaluated 6 approaches to code retrieval for AI coding agents using SWE, CodeScale Bench, and an internal question-answering harness. Structural search performed well on concrete tasks, while semantic retrieval helped with ambiguous requirements. The findings highlight trade-offs in accuracy, token use, and tool design, underscoring the need to match retrieval tools to task types. By Leela Kumili

## Automatic MSBuild configuration for the Microsoft C++ Language Server in Copilot CLI

DevFeed: [Automatic MSBuild configuration for the Microsoft C++ Language Server in Copilot CLI](<https://devfeed.tech/articles/automatic-msbuild-configuration-for-the-microsoft-c-language-server-in-copilot-cli-129361.md>)

Original publisher: [Read original article](<https://devblogs.microsoft.com/cppblog/automatic-msbuild-configuration-for-the-microsoft-c-language-server-in-copilot-cli/>)

Author: Sinem Akinci

Published: 2026-11-10T17:02:48Z

Content type: release

Language: en

Sources: [C++ Team Blog](<https://devfeed.tech/sources/c-team-blog.md>)

Topics: [GitHub Copilot CLI](<https://devfeed.tech/topics/github-copilot-cli.md>), [code search](<https://devfeed.tech/topics/code-search.md>)

Tags: [automatic](<https://devfeed.tech/tags/automatic.md>), [automatic-configuration](<https://devfeed.tech/tags/automatic-configuration.md>), [automatically](<https://devfeed.tech/tags/automatically.md>), [c-plus-plus](<https://devfeed.tech/tags/c-plus-plus.md>), [cli](<https://devfeed.tech/tags/cli.md>), [code-intelligence](<https://devfeed.tech/tags/code-intelligence.md>), [compilation](<https://devfeed.tech/tags/compilation.md>), [compile](<https://devfeed.tech/tags/compile.md>), [compiler](<https://devfeed.tech/tags/compiler.md>), [configuration](<https://devfeed.tech/tags/configuration.md>), [context](<https://devfeed.tech/tags/context.md>), [copilot](<https://devfeed.tech/tags/copilot.md>), [database](<https://devfeed.tech/tags/database.md>), [github-copilot-cli](<https://devfeed.tech/tags/github-copilot-cli.md>), [go](<https://devfeed.tech/tags/go.md>), [json](<https://devfeed.tech/tags/json.md>), [msbuild](<https://devfeed.tech/tags/msbuild.md>), [project](<https://devfeed.tech/tags/project.md>), [source](<https://devfeed.tech/tags/source.md>), [symbols](<https://devfeed.tech/tags/symbols.md>)

### AI overview

For MSBuild-based C++ projects, the Microsoft C++ Language Server in GitHub Copilot CLI now automatically generates a `compile_commands.json` at startup when needed, using Visual Studio's MSBuild tooling to read compiler options, include paths, and preprocessor definitions from `.vcxproj` files and property sheets. This removes the prior need to supply the database or have Copilot find and run a configuration skill, giving language features such as go-to-definition, find-references, and build-aware diagnostics project context. Generation runs in the background; its duration depends on solution size and complexity, and progress is available with `/lsp logs`. The automation covers MSBuild only: for CMake, the text directs users to ask Copilot CLI to "load project" to use the generate-compile-commands skill; custom builds can use a custom extractor skill. Installation prerequisites, including EULA acceptance, still apply. Timeout settings and disabling the automation are covered in configuration docs, with changes taking effect at the next language-server startup.

### Source excerpt

Accurate C++ code intelligence starts with understanding how your project builds. Compiler options, include paths, and preprocessor definitions all affect how the language server interprets your source files. A compilation database, such as compile_commands.json, provides that context, so that tools such as go-to-definition and find-references can resolve symbols accurately. The Microsoft C++ Language Server now [...] The post Automatic MSBuild configuration for the Microsoft C++ Language Server in Copilot CLI appeared first on C++ Team Blog.

## SWE-Time: Benchmarking Temporal Repository Understanding

DevFeed: [SWE-Time: Benchmarking Temporal Repository Understanding](<https://devfeed.tech/articles/swe-time-benchmarking-temporal-repository-understanding-127261.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.12083>)

Author: Shijun Xiao, Xinrong Chen, Hao Liu, Wensheng Zhang, Yongqiang Tang

Published: 2026-11-10T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [code search](<https://devfeed.tech/topics/code-search.md>), [Repository](<https://devfeed.tech/topics/repository.md>), [versioning](<https://devfeed.tech/topics/versioning.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [benchmarking](<https://devfeed.tech/tags/benchmarking.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [coding-agents](<https://devfeed.tech/tags/coding-agents.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [developer](<https://devfeed.tech/tags/developer.md>), [evaluation](<https://devfeed.tech/tags/evaluation.md>), [evolution](<https://devfeed.tech/tags/evolution.md>), [framework](<https://devfeed.tech/tags/framework.md>), [github](<https://devfeed.tech/tags/github.md>), [regression](<https://devfeed.tech/tags/regression.md>), [repository](<https://devfeed.tech/tags/repository.md>), [retrospective](<https://devfeed.tech/tags/retrospective.md>), [temporal](<https://devfeed.tech/tags/temporal.md>), [time](<https://devfeed.tech/tags/time.md>)

### AI overview

SWE-Time evaluates code agents on repository evolution rather than treating repositories as static snapshots. It tests retrospective understanding through historical regression analysis and prospective understanding of intended changes from pre-change states, using a framework that turns real developer queries and historical evidence into verifiable tasks. Across 1,661 cases from 65 GitHub repositories, 15 coding agents were evaluated; within Claude Code, Claude Opus 4.8 scored highest, with a SWE-Time average of 61.9. The abstract reports that historical code evidence and developer intent matter, while reasoning consistently across repository states remains challenging.

### Source excerpt

arXiv:2611.12083v1 Announce Type: new Abstract: Repository-level code-understanding benchmarks have progressively expanded the spatial scope of reasoning, from local code to entire repositories, while largely treating each repository as a static snapshot. However, how models reason about repository evolution over time remains largely unexplored. We introduce SWE-Time, a benchmark for temporal repository understanding that makes repository evolution itself the reasoning target. SWE-Time evaluates two complementary temporal directions: retrospective understanding of repository evolution through historical regression analysis and prospective understanding of intended changes from pre-change repository states. To enable realistic yet reliable evaluation in these settings, we introduce the RealWorld-to-Verifiable Construction Framework, which transforms real-world developer queries and historical evidence into verifiable evaluation tasks. We evaluate 15 coding agents on 1,661 cases from 65 GitHub repositories. Within Claude Code, Claude Opus 4.8 achieves the highest SWE-Time Avg. of 61.9, leaving substantial headroom. Controlled analyses highlight the importance of historical code evidence and developer intent, while consistent reasoning across repository states remains challenging. These results highlight repository evolution as a distinct and challenging dimension of code understanding.

## SWHForest: A Structural Curation and Search Layer for Large-Scale Code Corpora

DevFeed: [SWHForest: A Structural Curation and Search Layer for Large-Scale Code Corpora](<https://devfeed.tech/articles/swhforest-a-structural-curation-and-search-layer-for-large-scale-code-corpora-125379.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2611.11439>)

Author: Federico Pennino, Andrea Gurioli, Maurizio Gabbrielli, Paolo Ferragina

Published: 2026-11-10T05:00:00Z

Content type: news

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>), [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [code search](<https://devfeed.tech/topics/code-search.md>)

Tags: [api](<https://devfeed.tech/tags/api.md>), [archive](<https://devfeed.tech/tags/archive.md>), [ast](<https://devfeed.tech/tags/ast.md>), [auditing](<https://devfeed.tech/tags/auditing.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [boilerplate](<https://devfeed.tech/tags/boilerplate.md>), [code](<https://devfeed.tech/tags/code.md>), [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [curation](<https://devfeed.tech/tags/curation.md>), [dataset](<https://devfeed.tech/tags/dataset.md>), [download](<https://devfeed.tech/tags/download.md>), [downloads](<https://devfeed.tech/tags/downloads.md>), [file](<https://devfeed.tech/tags/file.md>), [files](<https://devfeed.tech/tags/files.md>), [framework](<https://devfeed.tech/tags/framework.md>), [index](<https://devfeed.tech/tags/index.md>), [internet](<https://devfeed.tech/tags/internet.md>), [intersection](<https://devfeed.tech/tags/intersection.md>), [language-models](<https://devfeed.tech/tags/language-models.md>), [layer](<https://devfeed.tech/tags/layer.md>), [search](<https://devfeed.tech/tags/search.md>)

### AI overview

SWHForest adds a structural search layer to pointer-based code corpora so researchers can identify files to download without first retrieving and parsing the full archive. It parses each file into an AST in one pass, discards the source text, and retains a structural summary. Queries run locally with SQL or index intersection, or can be exposed by archive providers as a server-side filtering API; only matching candidates are then downloaded. On the ten largest Stack-Edu languages (52.1 million unique files and 341 million definitions), the authors report millisecond query times and 46-72% fewer downloads. They also report that live structural searches fell from hours of per-language streaming to seconds or minutes.

### Source excerpt

arXiv:2611.11439v1 Announce Type: cross Abstract: Pre-training code language models relies on massive code corpora usually streamed over the internet. However, researchers rarely need a multi-terabyte archive in its entirety; instead, their workflow centers on curated subsets: selecting well-documented files, filtering out repetitive boilerplate, auditing for benchmark contamination, or retrieving specific code patterns. Yet, because modern corpora are typically published as file pointers and basic attributes rather than source text, no general structural layer exists to select which files to download: curating a subset requires downloading and parsing millions of files first just to discard most of them. We present SWHForest, an open-source framework and public dataset that provides this missing structural layer for any pointer-based corpus without redistributing source code. A single pass over the archive parses each file into an AST and immediately discards the source text, keeping only its structural summary. Queried locally or exposed by archive providers as a server-side filtering API, SWHForest selects matching files first (via SQL or index intersection) and downloads only surviving candidates. Across the ten largest languages of Stack-Edu (52.1M unique files, 341M definitions), SWHForest answers curation queries in milliseconds while cutting downloads by 46-72%, reducing live structural searches from hours of per-language streaming to seconds or minutes.

## Building a RAG Pipeline for Semantic Code Search: A Developer Diary and Field Notes

DevFeed: [Building a RAG Pipeline for Semantic Code Search: A Developer Diary and Field Notes](<https://devfeed.tech/articles/building-a-rag-pipeline-for-semantic-code-search-a-developer-diary-and-field-notes-119339.md>)

Original publisher: [Read original article](<https://blog.jetbrains.com/ai/2026/11/building-a-rag-pipeline-for-semantic-code-search-a-developer-diary-and-field-notes-part-2/>)

Author: Eduard Gurskiy

Published: 2026-11-05T14:54:27Z

Content type: article

Language: en

Sources: [JetBrains](<https://devfeed.tech/sources/the-jetbrains-blog.md>)

Topics: [code search](<https://devfeed.tech/topics/code-search.md>), [Retrieval Augmented Generation (RAG)](<https://devfeed.tech/topics/retrieval-augmented-generation-rag.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [search-engines](<https://devfeed.tech/topics/search-engines.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agentic](<https://devfeed.tech/tags/agentic.md>), [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [algorithm](<https://devfeed.tech/tags/algorithm.md>), [algorithms](<https://devfeed.tech/tags/algorithms.md>), [ann](<https://devfeed.tech/tags/ann.md>), [approximate-nearest-neighbor](<https://devfeed.tech/tags/approximate-nearest-neighbor.md>), [approximation](<https://devfeed.tech/tags/approximation.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [code](<https://devfeed.tech/tags/code.md>), [code-search](<https://devfeed.tech/tags/code-search.md>), [pipeline](<https://devfeed.tech/tags/pipeline.md>), [production](<https://devfeed.tech/tags/production.md>), [rag](<https://devfeed.tech/tags/rag.md>), [rag-pipeline](<https://devfeed.tech/tags/rag-pipeline.md>), [search](<https://devfeed.tech/tags/search.md>), [semantic](<https://devfeed.tech/tags/semantic.md>), [storage](<https://devfeed.tech/tags/storage.md>), [tooling](<https://devfeed.tech/tags/tooling.md>)

### AI overview

Air Context's code-search RAG pipeline combines approximate nearest-neighbor search with an exact comparison over a smaller candidate set, balancing scale and precision. For filtered search, pre-filtering gives full recall but can become costly on large subsets; metadata-aware indexes such as Qdrant and turbopuffer support filters within ANN, while post-filtering can return too few valid results. Incremental indexing reuses unchanged Merkle-tree subtrees and deduplicates chunks; HNSW builds run asynchronously with controlled concurrency, and the service scales across customer-pinned shards. Developers can also use a local USearch index, with results merged from local and remote search. After retrieval, the system removes duplicates, groups file fragments, and reranks candidates. A bounded retrieval agent can call search and read tools, while result detail varies with relevance to limit context. The excerpt says a follow-up will evaluate agent impact; its teaser reports up to 70% fewer turns, without describing the evaluation here.

### Source excerpt

Part 2: Storage, Search, and Agent Tooling In this blog series, we aim to share some of the lessons we learned over the past six months as we built and rolled out Air Context, a production-grade RAG system for code search. In the first post of the series, we covered chunking, parsing, and vectorization. Turning [...]

## OpenAI to acquire Pierre Computer Company, Inc.

DevFeed: [OpenAI to acquire Pierre Computer Company, Inc.](<https://devfeed.tech/articles/openai-to-acquire-pierre-computer-company-inc-117792.md>)

Original publisher: [Read original article](<https://openai.com/index/openai-to-acquire-pierre>)

Published: 2026-11-04T14:30:00Z

Content type: news

Language: en

Sources: [OpenAI News](<https://devfeed.tech/sources/openai-news.md>)

Topics: [OpenAI](<https://devfeed.tech/topics/openai.md>), [codex](<https://devfeed.tech/topics/codex.md>), [Tool](<https://devfeed.tech/topics/tool.md>), [code search](<https://devfeed.tech/topics/code-search.md>)

Tags: [acquisition](<https://devfeed.tech/tags/acquisition.md>), [agent](<https://devfeed.tech/tags/agent.md>), [api](<https://devfeed.tech/tags/api.md>), [apps](<https://devfeed.tech/tags/apps.md>), [branch](<https://devfeed.tech/tags/branch.md>), [codex](<https://devfeed.tech/tags/codex.md>), [coding](<https://devfeed.tech/tags/coding.md>), [coding-agents](<https://devfeed.tech/tags/coding-agents.md>), [company](<https://devfeed.tech/tags/company.md>), [developer-tools](<https://devfeed.tech/tags/developer-tools.md>), [developers](<https://devfeed.tech/tags/developers.md>), [diffs](<https://devfeed.tech/tags/diffs.md>), [git](<https://devfeed.tech/tags/git.md>), [highlighter](<https://devfeed.tech/tags/highlighter.md>), [highlights](<https://devfeed.tech/tags/highlights.md>), [libraries](<https://devfeed.tech/tags/libraries.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [openai](<https://devfeed.tech/tags/openai.md>), [pierre](<https://devfeed.tech/tags/pierre.md>), [tools](<https://devfeed.tech/tags/tools.md>)

### AI overview

OpenAI announced plans to acquire Pierre Computer Company, the company behind Code Storage and open-source developer tools used in Codex. Trees supports fast project-file navigation, while Diffs helps reviewers inspect changes; Diffs also offers inline editing, and Highlights provides fast syntax highlighting. OpenAI says it aims to build on Pierre's Git workflow tools for coding agents and advance source control. Pierre's open-source projects will remain open source, and customers can keep using Code Storage after closing while OpenAI works to bring its features into Codex. The acquisition remains subject to customary closing conditions, including regulatory approvals; the companies remain separate until closing.

### Source excerpt

OpenAI will acquire Pierre Computer Company, expanding Codex's tools for building, understanding, and reviewing code.

## MediaWiki Code2Code Search: Neural Retrieval for the Semantic Discovery of Open-Source Software Entities

DevFeed: [MediaWiki Code2Code Search: Neural Retrieval for the Semantic Discovery of Open-Source Software Entities](<https://devfeed.tech/articles/mediawiki-code2code-search-neural-retrieval-for-the-semantic-discovery-of-open-source-software-entities-113473.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2607.26766>)

Author: Francesco Tosoni

Published: 2026-11-03T05:00:00Z

Content type: article

Language: en

Sources: [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [code search](<https://devfeed.tech/topics/code-search.md>), [search-engines](<https://devfeed.tech/topics/search-engines.md>), [AI-driven information retrieval](<https://devfeed.tech/topics/ai-driven-information-retrieval.md>), [Open Source](<https://devfeed.tech/topics/open-source.md>)

Tags: [2](<https://devfeed.tech/tags/2.md>), [abstract](<https://devfeed.tech/tags/abstract.md>), [all](<https://devfeed.tech/tags/all.md>), [alpha](<https://devfeed.tech/tags/alpha.md>), [apache](<https://devfeed.tech/tags/apache.md>), [apache-2-0](<https://devfeed.tech/tags/apache-2-0.md>), [api](<https://devfeed.tech/tags/api.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [arxiv](<https://devfeed.tech/tags/arxiv.md>), [cs-ai](<https://devfeed.tech/tags/cs-ai.md>), [cs-cl](<https://devfeed.tech/tags/cs-cl.md>), [cs-ir](<https://devfeed.tech/tags/cs-ir.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [discovery](<https://devfeed.tech/tags/discovery.md>), [mediawiki](<https://devfeed.tech/tags/mediawiki.md>), [neural](<https://devfeed.tech/tags/neural.md>), [open](<https://devfeed.tech/tags/open.md>), [open-source](<https://devfeed.tech/tags/open-source.md>), [open-source-software](<https://devfeed.tech/tags/open-source-software.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [search](<https://devfeed.tech/tags/search.md>), [semantic](<https://devfeed.tech/tags/semantic.md>), [software](<https://devfeed.tech/tags/software.md>), [source](<https://devfeed.tech/tags/source.md>)

### AI overview

MediaWiki Code2Code Search uses neural retrieval to find functions, types, and templates by computational intent across 1.29 million entities in more than 2,500 MediaWiki repositories. Its split-build architecture handles GPU-intensive indexing offline and serves queries on CPU; the FAISS IVF-PQ index uses 168.6 MiB, 96.6% less than a flat float32 baseline, with median query latency of 1.85 seconds on commodity hardware under Toolforge's 6 GiB RAM limit. In a blinded 27-query evaluation, a three-model LLM jury rated its nDCG@10 at 0.81 versus 0.63 for BM25, with gains across all judges, especially for name-obfuscated and cross-language queries. The abstract reports an Apache 2.0 licence and an open REST API.

### Source excerpt

arXiv:2607.26766v2 Announce Type: replace-cross Abstract: Code search in large-scale ecosystems is often hindered by the lexical gap between user queries and implementation details, alongside the trade-off between the low latency of traditional Information Retrieval (IR) and the precision of Deep Learning (DL). We present MediaWiki Code2Code Search, a neural retrieval system for semantic code-to-code discovery. By indexing 1.29 million structural entities (functions, types, and templates) across 2,500+ MediaWiki repositories, our system enables retrieval based on computational intent rather than surface tokens. We employ a split-build architecture, decoupling GPU-intensive offline indexing from a CPU-only serving layer; our FAISS IVF-PQ index occupies 168.6 MiB: a 96.6\% reduction compared to a flat float32 baseline, and achieves a median query latency of 1.85 seconds on commodity hardware, satisfying the 6 GiB RAM constraint of Wikimedia Toolforge. In a blinded evaluation graded by a three-model LLM jury (inter-judge Krippendorff's $\alpha = 0.74$), our system improves ranking quality over a BM25 baseline across a 27-query benchmark, raising the jury-mean nDCG@10 (to 0.81 from 0.63), a positive and consistent gain across all judges. Gains are most pronounced in name-obfuscated and cross-language tasks where lexical methods fail. The system is available at https://code2codesearch.toolforge.org under the Apache 2.0 licence and provides an open RESTful API.

## What Is a Code Knowledge Graph? Graphify Explained with a City Map

DevFeed: [What Is a Code Knowledge Graph? Graphify Explained with a City Map](<https://devfeed.tech/articles/what-is-a-code-knowledge-graph-graphify-explained-with-a-city-map-106216.md>)

Original publisher: [Read original article](<https://ihatereading.in/blogs/what-is-a-code-knowledge-graph-graphify-explained-with-a-city-map>)

Author: Shrey Vijayvargiya

Published: 2026-10-29T09:00:00Z

Content type: article

Language: en

Sources: [iHateReading](<https://devfeed.tech/sources/ihatereading.md>)

Topics: [Graphs](<https://devfeed.tech/topics/graphs.md>), [graph theory](<https://devfeed.tech/topics/graph-theory.md>), [Graph Mining](<https://devfeed.tech/topics/graph-mining.md>), [code search](<https://devfeed.tech/topics/code-search.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [Agent Skills](<https://devfeed.tech/topics/agent-skills.md>), [MSP MCP](<https://devfeed.tech/topics/msp-mcp.md>), [Documentation](<https://devfeed.tech/topics/documentation.md>)

Tags: [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [ai](<https://devfeed.tech/tags/ai.md>), [ai-agents](<https://devfeed.tech/tags/ai-agents.md>), [ai-assistant](<https://devfeed.tech/tags/ai-assistant.md>), [algorithm](<https://devfeed.tech/tags/algorithm.md>), [apache-2-0](<https://devfeed.tech/tags/apache-2-0.md>), [api](<https://devfeed.tech/tags/api.md>), [ast](<https://devfeed.tech/tags/ast.md>), [backend](<https://devfeed.tech/tags/backend.md>), [blockchain](<https://devfeed.tech/tags/blockchain.md>), [class](<https://devfeed.tech/tags/class.md>), [code](<https://devfeed.tech/tags/code.md>), [code-knowledge-graph](<https://devfeed.tech/tags/code-knowledge-graph.md>), [code-knowledge-graph-for-claude-code](<https://devfeed.tech/tags/code-knowledge-graph-for-claude-code.md>), [code-knowledge-graph-mcp](<https://devfeed.tech/tags/code-knowledge-graph-mcp.md>), [coding](<https://devfeed.tech/tags/coding.md>), [convert-codebase-to-knowledge-graph](<https://devfeed.tech/tags/convert-codebase-to-knowledge-graph.md>), [development](<https://devfeed.tech/tags/development.md>), [frontend](<https://devfeed.tech/tags/frontend.md>), [function](<https://devfeed.tech/tags/function.md>), [graph](<https://devfeed.tech/tags/graph.md>), [graphify-alternatives](<https://devfeed.tech/tags/graphify-alternatives.md>), [graphify-claude-code](<https://devfeed.tech/tags/graphify-claude-code.md>), [javascript](<https://devfeed.tech/tags/javascript.md>), [knowledge-graph](<https://devfeed.tech/tags/knowledge-graph.md>), [map](<https://devfeed.tech/tags/map.md>), [node-js](<https://devfeed.tech/tags/node-js.md>), [product](<https://devfeed.tech/tags/product.md>), [programming](<https://devfeed.tech/tags/programming.md>), [react](<https://devfeed.tech/tags/react.md>), [repo](<https://devfeed.tech/tags/repo.md>), [search](<https://devfeed.tech/tags/search.md>), [typescript](<https://devfeed.tech/tags/typescript.md>), [vector](<https://devfeed.tech/tags/vector.md>), [vector-search](<https://devfeed.tech/tags/vector-search.md>), [web-development](<https://devfeed.tech/tags/web-development.md>)

### AI overview

A code knowledge graph maps files, functions and classes to nodes and relationships such as calls and imports to directed edges. The article reports running Graphify on a 100-file JavaScript and React repository: its tree-sitter code-only build produced 749 nodes and 1,760 edges in 2.7 seconds, locally and without an LLM. Walking call edges backward with `graphify affected` traced a change to `parseFrontmatter()` to 16 functions across two files, including sync operations that keyword or similarity search did not find. The article distinguishes graph traversal for impact and connections from vector search for finding code by resemblance, and suggests using both for different questions. Graphify can miss dynamic calls, reflection, string-built imports and cross-service calls; the author advises keeping the graph current and confirming findings with tests. The plain-language `query` command returned a fuzzy neighborhood in the reported example, while `affected` and `path` were more precise for impact questions.

### Source excerpt

A code knowledge graph is a map of your codebase. I ran Graphify on a real repo to see what breaks when I change one function, and why vector search can't answ...

## GitHub intelligence puts your code, context, and history to work

DevFeed: [GitHub intelligence puts your code, context, and history to work](<https://devfeed.tech/articles/github-intelligence-puts-your-code-context-and-history-to-work-105828.md>)

Original publisher: [Read original article](<https://github.blog/changelog/2026-10-28-github-intelligence-puts-your-code-context-and-history-to-work>)

Author: Allison

Published: 2026-10-28T15:40:00Z

Content type: release

Language: en

Sources: [GitHub Changelog](<https://devfeed.tech/sources/github-changelog.md>)

Topics: [GitHub Copilot](<https://devfeed.tech/topics/github-copilot.md>), [AI Context Management](<https://devfeed.tech/topics/ai-context-management.md>), [code search](<https://devfeed.tech/topics/code-search.md>), [GitHub](<https://devfeed.tech/topics/github.md>)

Tags: [agents](<https://devfeed.tech/tags/agents.md>), [code](<https://devfeed.tech/tags/code.md>), [context](<https://devfeed.tech/tags/context.md>), [copilot](<https://devfeed.tech/tags/copilot.md>), [github](<https://devfeed.tech/tags/github.md>), [history](<https://devfeed.tech/tags/history.md>), [intelligence](<https://devfeed.tech/tags/intelligence.md>), [release](<https://devfeed.tech/tags/release.md>), [universe26](<https://devfeed.tech/tags/universe26.md>)

### AI overview

GitHub intelligence adds team context to Copilot through three capabilities: Copilot Memory carries personal preferences and repository conventions across sessions; Dreaming (public preview, opt-in) updates memories and proposes changes to checked-in skills or instructions through pull requests; and code search finds relevant code across related repositories without a local clone, using exact text, symbols, patterns, filters, or semantic queries. Memory is generally available on paid Copilot plans and users can view or delete personal memories. Dreaming has separate user and administrator settings, repository access follows Copilot Memory settings, and dream runs consume AI credits. In reviewed public-preview CLI sessions that used memory, 93% showed better adherence to preferences and 42% showed more efficient workflows attributed to the included memories; these figures describe the reviewed sessions, not all usage.

### Source excerpt

Agents should understand more than your codebase--they should understand how your team works. Your preferences, repository conventions, skills, and custom instructions guide how changes should be made, while code in... The post GitHub intelligence puts your code, context, and history to work appeared first on The GitHub Blog.

## HIDE: Hierarchical Isolation via Documentation and Evidence for Rustc Bugs

DevFeed: [HIDE: Hierarchical Isolation via Documentation and Evidence for Rustc Bugs](<https://devfeed.tech/articles/hide-hierarchical-isolation-via-documentation-and-evidence-for-rustc-bugs-103658.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.37343>)

Author: Tianrong Zhang, Yakun Zhang, Dan Hao

Published: 2026-10-28T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [code search](<https://devfeed.tech/topics/code-search.md>), [AI Chat](<https://devfeed.tech/topics/ai-chat.md>), [Static code analysis](<https://devfeed.tech/topics/static-code-analysis.md>), [Documentation](<https://devfeed.tech/topics/documentation.md>), [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>)

Tags: [bugs](<https://devfeed.tech/tags/bugs.md>), [challenges](<https://devfeed.tech/tags/challenges.md>), [competition](<https://devfeed.tech/tags/competition.md>), [compiler](<https://devfeed.tech/tags/compiler.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [documentation](<https://devfeed.tech/tags/documentation.md>), [isolation](<https://devfeed.tech/tags/isolation.md>), [language-models](<https://devfeed.tech/tags/language-models.md>), [rank](<https://devfeed.tech/tags/rank.md>), [reduce](<https://devfeed.tech/tags/reduce.md>), [revision](<https://devfeed.tech/tags/revision.md>), [source-code](<https://devfeed.tech/tags/source-code.md>)

### AI overview

HIDE ranks suspicious Rust compiler source files by combining compiler documentation with bug evidence. Its Layered Documentation Aggregation stage summarizes individual files and how similarly purposed files work together. Hierarchical Bug Isolation then narrows candidates, using Scoped Candidate Competition to limit files considered and Evidence-Guided Code Slicing to limit code included in prompts. The authors evaluate HIDE on 160 reproducible rustc bugs against Agentless, SoapFL, and GraphLocator. With GPT-5.4, it reports 42.08% Accuracy@1 and 80.42% Accuracy@10, respectively 5.83 and 6.04 percentage points above the strongest baseline results. The evidence provided is the abstract, not the full paper.

### Source excerpt

arXiv:2610.37343v1 Announce Type: new Abstract: Rustc bugs can prevent valid programs from compiling or cause incorrect compilation results, making the identification of responsible source files essential for repair. Using large language models (LLMs) for rustc bug isolation raises three challenges. First, distributed compiler functionality requires understanding both what individual files do and how they work together. Second, a large source file search space makes exhaustive analysis of all files costly. Third, including documentation and source code for many candidate files in LLM prompts leads to context overload, which can reduce bug isolation accuracy. We propose HIDE, a hierarchical LLM-based approach that combines compiler documentation with bug evidence to rank suspicious rustc source files. Given a rustc revision and a bug context, HIDE consists of two stages that generate compiler documentation and narrow the search space. To understand distributed compiler functionality, the Layered Documentation Aggregation stage summarizes what individual files do and how files with similar functions work together. To reduce the large source file search space, the Hierarchical Bug Isolation stage uses hierarchical candidate selection to reduce the number of files requiring detailed analysis. To mitigate context overload, this stage uses Scoped Candidate Competition and Evidence-Guided Code Slicing to limit the number of candidates and the amount of source code included in each prompt, respectively. We evaluate HIDE on 160 reproducible rustc bugs and compare it with Agentless, SoapFL, and GraphLocator. With GPT-5.4, HIDE achieves 42.08% Accuracy@1 and 80.42% Accuracy@10, exceeding the strongest baseline results by 5.83% and 6.04%, respectively. These results demonstrate that generating compiler documentation, narrowing candidates hierarchically, and limiting analysis context enable effective rustc bug isolation.

## OneCodeReviewBench: Typed Decisions for Code Review and Search

DevFeed: [OneCodeReviewBench: Typed Decisions for Code Review and Search](<https://devfeed.tech/articles/onecodereviewbench-typed-decisions-for-code-review-and-search-101742.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.33389>)

Author: Mohamad Mohebifar, Alex Bit

Published: 2026-10-27T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [Code review](<https://devfeed.tech/topics/code-review.md>), [ai code review](<https://devfeed.tech/topics/ai-code-review.md>), [code search](<https://devfeed.tech/topics/code-search.md>), [Memory Safety](<https://devfeed.tech/topics/memory-safety.md>), [software security audit](<https://devfeed.tech/topics/software-security-audit.md>)

Tags: [bugs](<https://devfeed.tech/tags/bugs.md>), [code](<https://devfeed.tech/tags/code.md>), [code-review](<https://devfeed.tech/tags/code-review.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [errors](<https://devfeed.tech/tags/errors.md>), [languages](<https://devfeed.tech/tags/languages.md>), [memory](<https://devfeed.tech/tags/memory.md>), [memory-safety](<https://devfeed.tech/tags/memory-safety.md>), [model](<https://devfeed.tech/tags/model.md>), [openai](<https://devfeed.tech/tags/openai.md>), [python](<https://devfeed.tech/tags/python.md>), [search](<https://devfeed.tech/tags/search.md>), [yaml](<https://devfeed.tech/tags/yaml.md>)

### AI overview

OneCodeReviewBench evaluates bounded code-review decisions across 1,968 synthetic cases in 12 programming languages and YAML, using explicit contracts for runtime assumptions, source scope, and answer criteria. Among ten model-adapter systems, OpenAI Decisions scores 81.3% Boolean balanced accuracy; Kev-27B scores 76.8% Choice accuracy and 0.367 Score mean absolute error. A bounded Python import resolver matches all 40 references tested within its supported profile. The benchmark also includes raw symbol-search comparisons, engineer-reviewed cases, and a versioned artifact. Remaining reviewer disagreements and repaired inputs need resolution; transfer to repositories and matched latency and cost evaluation remain future work.

### Source excerpt

arXiv:2610.33389v1 Announce Type: new Abstract: Code review requires finding functional bugs, memory-safety errors, security weaknesses, and performance problems across languages and software stacks. Tools can express bounded parts of this work as decisions about supplied code. We present OneCodeReviewBench, 1,968 synthetic cases across 12 programming languages and YAML, with Boolean, categorical, and ordered answers. Explicit contracts specify runtime assumptions, source scope, and answer criteria; separate evidence records support reference review. One author reviewed 150 cases; two additional engineers reviewed 100 of those cases. We compare ten model--adapter systems on the same cases and contracts: OpenAI Decisions reaches 81.3% Boolean balanced accuracy; Kev-27B reaches 76.8% Choice accuracy and Score mean absolute error of 0.367. Excluding two families with incomplete contracts preserves these leaders. Additional comparisons cover 586 cases with raw symbol search and 40 with a bounded Python import resolver; the resolver matches all 40 references within its supported profile. A versioned artifact reproduces these analyses and retains the 85 cases with matching engineer answers; the reported review subset excludes two incomplete-contract cases. Remaining reviewer disagreements and repaired inputs require resolution; repository-scale transfer and matched latency/cost evaluation remain future work.

## Context7 vs Exa 3.0 on WebCode: The Same Accuracy in a Fifth of the Tokens

DevFeed: [Context7 vs Exa 3.0 on WebCode: The Same Accuracy in a Fifth of the Tokens](<https://devfeed.tech/articles/context7-vs-exa-3-0-on-webcode-the-same-accuracy-in-a-fifth-of-the-tokens-103671.md>)

Original publisher: [Read original article](<https://upstash.com/blog/context7-search-webcode>)

Author: Ramazan Onur Acar

Published: 2026-10-27T00:00:00Z

Content type: news

Language: en

Sources: [Upstash](<https://devfeed.tech/sources/upstash.md>)

Topics: [Benchmark](<https://devfeed.tech/topics/benchmark.md>), [code search](<https://devfeed.tech/topics/code-search.md>), [Machine learning](<https://devfeed.tech/topics/machine-learning.md>), [Claude](<https://devfeed.tech/topics/claude.md>)

Tags: [adapter](<https://devfeed.tech/tags/adapter.md>), [agent](<https://devfeed.tech/tags/agent.md>), [agents](<https://devfeed.tech/tags/agents.md>), [api](<https://devfeed.tech/tags/api.md>), [api-key](<https://devfeed.tech/tags/api-key.md>), [bazel](<https://devfeed.tech/tags/bazel.md>), [benchmark](<https://devfeed.tech/tags/benchmark.md>), [browser](<https://devfeed.tech/tags/browser.md>), [claude](<https://devfeed.tech/tags/claude.md>), [claude-opus](<https://devfeed.tech/tags/claude-opus.md>), [claude-opus-5-5](<https://devfeed.tech/tags/claude-opus-5-5.md>), [client](<https://devfeed.tech/tags/client.md>), [code](<https://devfeed.tech/tags/code.md>), [coding](<https://devfeed.tech/tags/coding.md>), [coding-agents](<https://devfeed.tech/tags/coding-agents.md>), [compute](<https://devfeed.tech/tags/compute.md>), [context7](<https://devfeed.tech/tags/context7.md>), [curl](<https://devfeed.tech/tags/curl.md>), [search](<https://devfeed.tech/tags/search.md>), [tokens](<https://devfeed.tech/tags/tokens.md>)

### AI overview

In the 307-question WebCode RAG benchmark, Context7 Search and Exa 3.0 produced similar correct-answer rates: 85.7% for Context7 and 86.2% for Exa Lite and Auto. With the same answer and judge models, Context7 responses averaged about 1,050 tokens and led to fewer wrong answers (2.3%, versus 7.3% for Lite and 8.0% for Auto); Exa searches were faster, at 0.6 seconds for Lite and 1.3 for Auto versus 3.7 for Context7. The authors attribute the wrong-answer gap to Context7 results being shorter and more likely to contain the answer, and to the model more often declining when they did not. Results are averages across three production runs; the authors build Context7, and the comparison covers this benchmark and configuration.

### Source excerpt

On Exa's WebCode benchmark, Context7 Search and Exa 3.0 led to as many correct answers, with about a fifth of the tokens for Context7 and a third of the wrong answers.

## CuraLoc: Repository-Level Fault Localization via Planned Code Inspection

DevFeed: [CuraLoc: Repository-Level Fault Localization via Planned Code Inspection](<https://devfeed.tech/articles/curaloc-repository-level-fault-localization-via-planned-code-inspection-99512.md>)

Original publisher: [Read original article](<https://arxiv.org/abs/2610.31809>)

Author: Sangeeta Warrier, Anush Sankaran, Ryen White

Published: 2026-10-26T04:00:00Z

Content type: article

Language: en

Sources: [arXiv: Machine Learning](<https://devfeed.tech/sources/arxiv-machine-learning.md>), [arXiv: Software Engineering](<https://devfeed.tech/sources/arxiv-software-engineering.md>)

Topics: [code analysis](<https://devfeed.tech/topics/code-analysis.md>), [code search](<https://devfeed.tech/topics/code-search.md>), [llm-reasoning](<https://devfeed.tech/topics/llm-reasoning.md>), [AI-driven information retrieval](<https://devfeed.tech/topics/ai-driven-information-retrieval.md>), [gpt-6-astra](<https://devfeed.tech/topics/gpt-6-astra.md>)

Tags: [cs-lg](<https://devfeed.tech/tags/cs-lg.md>), [cs-se](<https://devfeed.tech/tags/cs-se.md>), [inspection](<https://devfeed.tech/tags/inspection.md>), [language-models](<https://devfeed.tech/tags/language-models.md>), [localization](<https://devfeed.tech/tags/localization.md>), [machine-learning](<https://devfeed.tech/tags/machine-learning.md>), [planning](<https://devfeed.tech/tags/planning.md>), [repository](<https://devfeed.tech/tags/repository.md>), [retrieval](<https://devfeed.tech/tags/retrieval.md>), [swe-bench](<https://devfeed.tech/tags/swe-bench.md>)

### AI overview

CuraLoc plans repository inspection before retrieving source: it selects candidate files, summarizes symbols, then uses one model decision to choose starting symbols, related code to inspect, and likely edit hypotheses. Deterministic retrieval gathers that code for a final ranking, avoiding iterative model-controlled navigation. On SWE-bench Verified with GPT-5.2, it reports Func@1 of 59.3%, versus LocAgent's 53.8%, RepoSearcher's 56.2%, and Agentless's 39.6%, with 90.3% fewer recorded tokens than LocAgent. On SWE-bench Lite, it reports 83.8% Func@5, over 15% relative improvement on Agentless. The abstract says strong localization results were also demonstrated with GPT-6 Astra; it provides no specific Astra score.

### Source excerpt

arXiv:2610.31809v1 Announce Type: new Abstract: A bug report describes what went wrong, but rarely where the code should change. Existing repository-level localization methods often let language models iteratively decide what code to inspect next, requiring repeated model calls and substantial token usage. We investigate whether these inspection decisions can instead be planned before source retrieval. In this research, we introduce CuraLoc, which first selects candidate files and represents their symbols using compact behavioral and structural descriptions. In a single planning decision, the model jointly selects starting symbols, chooses code relationships to inspect from each, and forms hypotheses about likely edits. Deterministic retrieval collects the requested source for a final ranking informed by those hypotheses, concentrating model reasoning on choosing and comparing code rather than directing each retrieval step. We demonstrate the efficacy of our proposed approach on the SWE-Bench Lite and Verified datasets. On SWE Bench Verified with GPT-5.2, CuraLoc raises Func@1 from LocAgent's 53.8% to 59.3% while using 90.3% fewer recorded tokens, and also outperforms RepoSearcher (56.2%) and Agentless (39.6%) on Func@1. On SWE-bench Lite, CuraLoc achieves a high 83.8% Func@5 accuracy improving on Agentless by more than 15% in relative accuracy. Additionally, CuraLoc supports different model backbones, with strong localization accuracy demonstrated using both GPT-5.2 and GPT-6 Astra. These results provide evidence that repository-level code inspection can be planned effectively without iterative model-controlled navigation.

## Qdrant 1.20 - Full-Text Search Improvements

DevFeed: [Qdrant 1.20 - Full-Text Search Improvements](<https://devfeed.tech/articles/qdrant-1-20-full-text-search-improvements-95827.md>)

Original publisher: [Read original article](<https://qdrant.tech/blog/qdrant-1.20.x/>)

Author: info@qdrant.tech (Andrey Vasnetsov)

Published: 2026-10-22T01:00:00Z

Content type: release

Language: en

Sources: [Qdrant Blog on Qdrant - Vector Search Engine](<https://devfeed.tech/sources/qdrant-blog-on-qdrant-vector-search-engine.md>)

Topics: [Qdrant](<https://devfeed.tech/topics/qdrant.md>), [search-engines](<https://devfeed.tech/topics/search-engines.md>), [code search](<https://devfeed.tech/topics/code-search.md>), [Embeddings](<https://devfeed.tech/topics/embeddings.md>), [quantization](<https://devfeed.tech/topics/quantization.md>), [AI search](<https://devfeed.tech/topics/ai-search.md>), [hybrid-search](<https://devfeed.tech/topics/hybrid-search.md>)

Tags: [approximate-nearest-neighbor-search](<https://devfeed.tech/tags/approximate-nearest-neighbor-search.md>), [bert](<https://devfeed.tech/tags/bert.md>), [bm25](<https://devfeed.tech/tags/bm25.md>), [code](<https://devfeed.tech/tags/code.md>), [code-search](<https://devfeed.tech/tags/code-search.md>), [embeddings](<https://devfeed.tech/tags/embeddings.md>), [faster](<https://devfeed.tech/tags/faster.md>), [fasttext](<https://devfeed.tech/tags/fasttext.md>), [full-text-search](<https://devfeed.tech/tags/full-text-search.md>), [hnsw](<https://devfeed.tech/tags/hnsw.md>), [image-search](<https://devfeed.tech/tags/image-search.md>), [knn-algorithm](<https://devfeed.tech/tags/knn-algorithm.md>), [matching](<https://devfeed.tech/tags/matching.md>), [nested](<https://devfeed.tech/tags/nested.md>), [neural-network](<https://devfeed.tech/tags/neural-network.md>), [payload](<https://devfeed.tech/tags/payload.md>), [qdrant](<https://devfeed.tech/tags/qdrant.md>), [quantization](<https://devfeed.tech/tags/quantization.md>), [recommender-system](<https://devfeed.tech/tags/recommender-system.md>), [saas](<https://devfeed.tech/tags/saas.md>), [search](<https://devfeed.tech/tags/search.md>), [simaes-networks](<https://devfeed.tech/tags/simaes-networks.md>), [similarity](<https://devfeed.tech/tags/similarity.md>), [storage](<https://devfeed.tech/tags/storage.md>), [transformer](<https://devfeed.tech/tags/transformer.md>), [vector](<https://devfeed.tech/tags/vector.md>), [vector-search](<https://devfeed.tech/tags/vector-search.md>), [vector-search-engine](<https://devfeed.tech/tags/vector-search-engine.md>), [vectors](<https://devfeed.tech/tags/vectors.md>), [version](<https://devfeed.tech/tags/version.md>), [word2vec](<https://devfeed.tech/tags/word2vec.md>)

### AI overview

Qdrant 1.20 adds BM25 ranking directly on full-text indexes, deriving scoring statistics from payload text and allowing query-level tuning without generating embeddings; the new `code` tokenizer splits identifiers on underscores and case changes while retaining full identifiers. Per-element payload indexes can answer nested filters without payload reads when all conditions use indexed fields on the same array; a benchmark on 100,000 points found them 7-38 times faster than regular indexes with payload reads. New `turbo16` and `turbo8` datatypes reduce vector storage: reported tests found `turbo16` matched `float32` recall@10 within 0.001, while `turbo8` approached `float16` recall at half its size. The release also adds `bits8` quantization, memory-tier controls for the ID tracker, and internal cluster authentication by default when API keys are configured. Upgrades require attention to read-only-key clusters and memory tiers for inline storage, whose vectors now follow the HNSW index tier.

### Source excerpt

Qdrant 1.20.0 is out! Let's look at the main features for this version: BM25 on Full-Text Indexes and Code Search: Rank by BM25 straight from your payload text, with no embeddings to generate. Plus, better code search with a new code tokenizer. Faster Nested Filters: A new per-element payload index answers nested filters without reading payloads, making them faster. 16-Bit and 8-Bit TurboQuant: A 16-bit TurboQuant datatype that halves vector storage without accuracy loss, and an 8-bit TurboQuant datatype and quantization option with recall close to float16.

##  7 Terminal Tools That Actually Earn Their Place in Your $PATH

DevFeed: [ 7 Terminal Tools That Actually Earn Their Place in Your $PATH](<https://devfeed.tech/articles/7-terminal-tools-that-actually-earn-their-place-in-your-path-62433.md>)

Original publisher: [Read original article](<https://blog.devops.dev/%EF%B8%8F-7-terminal-tools-that-actually-earn-their-place-in-your-path-1bcb0751b7fb?source=rss----33f8b2d9a328---4>)

Author: Arvind Verma

Published: 2026-09-30T12:10:35Z

Content type: article

Language: en

Sources: [DevOps.dev](<https://devfeed.tech/sources/devops-dev.md>)

Topics: [Command-line interface](<https://devfeed.tech/topics/cli.md>), [Tool](<https://devfeed.tech/topics/tool.md>), [Developer experience](<https://devfeed.tech/topics/developer-experience.md>), [code search](<https://devfeed.tech/topics/code-search.md>), [Platform Engineering](<https://devfeed.tech/topics/platform-engineering.md>), [CI/CD](<https://devfeed.tech/topics/cicd.md>)

Tags: [ai](<https://devfeed.tech/tags/ai.md>), [aws](<https://devfeed.tech/tags/aws.md>), [aws-cli](<https://devfeed.tech/tags/aws-cli.md>), [ci-cd](<https://devfeed.tech/tags/ci-cd.md>), [cli](<https://devfeed.tech/tags/cli.md>), [cli-tools](<https://devfeed.tech/tags/cli-tools.md>), [cloud](<https://devfeed.tech/tags/cloud.md>), [command-line](<https://devfeed.tech/tags/command-line.md>), [configuration](<https://devfeed.tech/tags/configuration.md>), [delta](<https://devfeed.tech/tags/delta.md>), [devops](<https://devfeed.tech/tags/devops.md>), [devops-tools](<https://devfeed.tech/tags/devops-tools.md>), [diffs](<https://devfeed.tech/tags/diffs.md>), [docker](<https://devfeed.tech/tags/docker.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [fundamentals](<https://devfeed.tech/tags/fundamentals.md>), [fzf](<https://devfeed.tech/tags/fzf.md>), [git](<https://devfeed.tech/tags/git.md>), [kubernetes](<https://devfeed.tech/tags/kubernetes.md>), [terminal](<https://devfeed.tech/tags/terminal.md>), [terminal-tools](<https://devfeed.tech/tags/terminal-tools.md>), [tools](<https://devfeed.tech/tags/tools.md>)

### AI overview

The article presents seven CLI tools for speeding up information retrieval in DevOps: ripgrep searches text, fd locates files, fzf interactively filters lists, bat displays files with syntax highlighting, jq extracts and transforms JSON, delta improves Git diff readability, and atuin makes shell history searchable. It gives examples for Kubernetes, AWS, Terraform, Docker, CI/CD, and troubleshooting, and emphasizes combining tools into a search-find-filter-preview-parse-review workflow. The claimed benefit is less time spent locating and understanding information. It advises learning standard Unix utilities too, since production servers and minimal containers may lack these additions. Installation guidance notes that Ubuntu/Debian package availability varies and recommends project installation methods when needed.

### Source excerpt

There is a strange thing that happens when you work in DevOps long enough. You start with a terminal. Then you install Docker. Then Kubernetes. Then Terraform. Then AWS CLI. Then Helm. Then Git. Then dozens of other tools. And somehow, despite having a machine capable of doing almost everything, you still spend half your day typing commands like: grep find cat git diff followed by increasingly elaborate combinations of pipes, flags, and questionable shell gymnastics. The problem usually isn't fixing the issue. It's finding the right information quickly. That's where a few modern command-line tools can make a surprisingly large difference. They don't replace Linux fundamentals. They make the fundamentals much more pleasant to use. Here are 7 terminal tools that are genuinely useful for DevOps, cloud, Kubernetes, Terraform, Docker, CI/CD and everyday engineering work. 🚀 Section 1: The Terminal Is Still One of Your Best DevOps Tools 🧑💻 Graphical tools are convenient. Dashboards are convenient. Cloud consoles are convenient. But when something breaks in production, engineers frequently end up back at the terminal. You need to: Search logs Inspect configuration Find files Compare changes Parse JSON Search command history Inspect Kubernetes resources Investigate Terraform changes Understand what actually changed And that's where terminal productivity matters. The problem is that traditional Unix tools were designed decades ago. They're incredibly powerful, but sometimes their interfaces feel like they were created by people who believed everyone should memorize 47 flags before breakfast. Modern CLI tools improve the experience without abandoning the underlying Unix philosophy. The seven tools in this article are: 01. ripgrep (rg) Fast text search. 02. fd A simpler and faster file finder. 03. fzf Interactive fuzzy finding. 04. bat A better cat with syntax highlighting. 05. jq JSON processing. 06. delta Better Git diffs. 07. atuin A smarter shell history. The interesting p

## Using Jev as a search reranker: benchmarks and how to implement

DevFeed: [Using Jev as a search reranker: benchmarks and how to implement](<https://devfeed.tech/articles/using-jev-as-a-search-reranker-benchmarks-and-how-to-implement-78762.md>)

Original publisher: [Read original article](<https://www.elastic.co/search-labs/blog/ecommerce-search-reranking-llm-alternative-jev>)

Author: Dustin Coates

Published: 2026-09-30T00:00:00Z

Content type: article

Language: en

Sources: [Elasticsearch Labs](<https://devfeed.tech/sources/elasticsearch-labs.md>)

Topics: [code search](<https://devfeed.tech/topics/code-search.md>), [elasticsearch](<https://devfeed.tech/topics/elasticsearch.md>), [hybrid-search](<https://devfeed.tech/topics/hybrid-search.md>), [LLM evaluation / benchmarking](<https://devfeed.tech/topics/llm-evaluation-benchmarking.md>), [sentence-transformers](<https://devfeed.tech/topics/sentence-transformers.md>), [Elastic Cloud](<https://devfeed.tech/topics/elastic-cloud.md>)

Tags: [0-1](<https://devfeed.tech/tags/0-1.md>), [benchmarks](<https://devfeed.tech/tags/benchmarks.md>), [bm25](<https://devfeed.tech/tags/bm25.md>), [elasticsearch](<https://devfeed.tech/tags/elasticsearch.md>), [how-to](<https://devfeed.tech/tags/how-to.md>), [hybrid](<https://devfeed.tech/tags/hybrid.md>), [hybrid-search](<https://devfeed.tech/tags/hybrid-search.md>), [implement](<https://devfeed.tech/tags/implement.md>), [integrations](<https://devfeed.tech/tags/integrations.md>), [jev](<https://devfeed.tech/tags/jev.md>), [python](<https://devfeed.tech/tags/python.md>), [relevance](<https://devfeed.tech/tags/relevance.md>), [reranking](<https://devfeed.tech/tags/reranking.md>), [search](<https://devfeed.tech/tags/search.md>), [shopping](<https://devfeed.tech/tags/shopping.md>)

### AI overview

The article describes a benchmark of Jev, TypeSafe AI's System One model, for reranking ecommerce search results retrieved by Elasticsearch. On a 250-query Amazon Shopping Queries test set, its best composite policy raised nDCG@10 from 0.9351 to 0.9565 and Exact MRR from 0.9201 to 0.9616. The benchmark compares several Jev scoring policies with BM25, hybrid retrieval, and a Jina reranker, and provides a Python implementation. The authors caution that this is one test on one dataset, with a high proportion of Exact judgments.

### Source excerpt

We had Jev score Elasticsearch hybrid search results and let a short Python policy do the reranking, taking nDCG@10 from 0.9351 to 0.9565 on 250 Amazon Shopping Queries, and the code is all here.

## Why Coding Agents Need Dependency Signals Beyond Code Similarity

DevFeed: [Why Coding Agents Need Dependency Signals Beyond Code Similarity](<https://devfeed.tech/articles/good-architecture-deletes-the-signals-your-agent-depends-on-60595.md>)

Original publisher: [Read original article](<https://towardsdatascience.com/good-architecture-deletes-the-signals-your-agent-depends-on/>)

Author: Yonatan Sason

Published: 2026-09-27T12:00:02Z

Content type: article

Language: en

Sources: [Towards Data Science](<https://devfeed.tech/sources/towards-data-science.md>)

Topics: [code search](<https://devfeed.tech/topics/code-search.md>), [software-architecture](<https://devfeed.tech/topics/software-architecture.md>)

Tags: [agentic-ai](<https://devfeed.tech/tags/agentic-ai.md>), [ai-assistant](<https://devfeed.tech/tags/ai-assistant.md>), [ai-coding-assistants](<https://devfeed.tech/tags/ai-coding-assistants.md>), [architecture](<https://devfeed.tech/tags/architecture.md>), [coding-agent](<https://devfeed.tech/tags/coding-agent.md>), [coding-agents](<https://devfeed.tech/tags/coding-agents.md>), [context-engineering](<https://devfeed.tech/tags/context-engineering.md>), [developer-tools](<https://devfeed.tech/tags/developer-tools.md>), [engineering](<https://devfeed.tech/tags/engineering.md>), [llm](<https://devfeed.tech/tags/llm.md>), [software-engineering](<https://devfeed.tech/tags/software-engineering.md>)

### AI overview

Code retrieval tools surface code that resembles a query, but coding agents also need to identify code that may break when a change is made. The article argues that architectural boundaries can hide these dependency signals from tools that search repositories, making safe changes harder to assess.

### Source excerpt

Every boundary you draw removes a signal your tooling was relying on. That is a structure problem, not a search problem. The post Good Architecture Deletes the Signals Your Agent Depends On appeared first on Towards Data Science.

## Faster C++ code intelligence with whole codebase indexing

DevFeed: [Faster C++ code intelligence with whole codebase indexing](<https://devfeed.tech/articles/faster-c-code-intelligence-with-whole-codebase-indexing-71422.md>)

Original publisher: [Read original article](<https://github.blog/changelog/2026-09-22-faster-c-code-intelligence-with-whole-codebase-indexing>)

Author: Allison

Published: 2026-09-22T22:24:48Z

Content type: article

Language: en

Sources: [GitHub Changelog](<https://devfeed.tech/sources/github-changelog.md>)

Topics: [GitHub Copilot CLI](<https://devfeed.tech/topics/github-copilot-cli.md>), [c/c++](<https://devfeed.tech/topics/c-c-plus-plus.md>), [code search](<https://devfeed.tech/topics/code-search.md>)

Tags: [c-plus-plus](<https://devfeed.tech/tags/c-plus-plus.md>), [code-intelligence](<https://devfeed.tech/tags/code-intelligence.md>), [github-copilot-cli](<https://devfeed.tech/tags/github-copilot-cli.md>), [indexing](<https://devfeed.tech/tags/indexing.md>), [microsoft](<https://devfeed.tech/tags/microsoft.md>)

### AI overview

Whole codebase indexing gives GitHub Copilot CLI's Microsoft C++ Language Server a persistent symbol index spanning open and unopened files, using project compilation information to resolve types, symbols, includes, and file relationships. Reusing that index can speed up definitions, references, implementations, and symbol searches. WCI is enabled by default and loads when a C++ project is first opened; check progress with `/lsp logs`. Initial indexing can take longer and temporarily use more memory, especially in large or complex repositories. The index is reused and dynamically updated afterward. To disable WCI, use the indexing documentation's setting and restart the Copilot session.

### Source excerpt

C++ code intelligence in GitHub Copilot CLI is now faster with support for whole codebase indexing. C++ repositories can contain millions of lines of code across deeply connected source files...

[Next page](<https://devfeed.tech/topics/code-search.md?cursor=WyIyMDI2LTA5LTIyVDIyOjI0OjQ4KzAwOjAwIiwgIjUxYjk5NmE1LTJhYjUtNDY3ZC05YmE5LWFhMGI3ODIxY2EyYiJd>)