29 posts
AgentSkills CLI started in May to install CritterStack AI Skills from a NuGet package. 1.1.0 adds package-level updates, restore, private feeds and package source mapping. Here is how it compares with NuGet's new dotnet-package-skills, and the plan for 2.0.
ReverseMarkdown more than doubled its NuGet downloads this year, from 3.7M in January to 7.5M. The 6.3.0 release cuts allocations by up to 58% on real-world pages by making the unclosed script and style cleanup linear, and adds TextWriter and async output overloads.
Pin the .NET SDK and the Infisical CLI with mise, inject secrets at process start with infisical run, and let .NET configuration pick them up as environment variables. The same tasks on a laptop and in CI, verified against a self-hosted Infisical.
Polecat 5.29 added keyword search to the Critter Stack's SQL Server store, and it does not use SQL Server's full-text engine to do it. The inverted index Polecat maintains itself, the LINQ operators over it, and BM25 ranking derived from the token rows.
SQL Server 2025 added a VECTOR type. Polecat 5.28 reaches it with a persisted computed column over the JSON document body, which means no write path changes and no backfill. Exact k-nearest-neighbour, three metrics that are all distances, and the one real limit.
Keyword search and vector search fail in different places, so a document both rank well beats either one's favourite. How Polecat 5.29 fuses the two rankings with RRF, why only the ranks are fused and never the scores, and why candidate depth has to exceed the limit.
A Fisher store is one SQLite file, which makes backup sound trivial. Copying that file while the application runs silently loses committed writes. Here is the demonstration, the SQLite primitive that gets it right, and a hosted service that snapshots, verifies, compresses and ships every database to S3.
Event sourcing in Fisher: a sale as a stream, committing in the same transaction as your documents, projections across the lifecycles, and the async daemon. Plus the two places SQLite genuinely changes the semantics, where the exclusive append methods fail instead of waiting and where contiguous sequences make the high-water mark trivial.
Fisher is the Marten and Polecat programming model on SQLite, running inside your own process with no database server at all. A tour of the document side: identity, sessions, LINQ including joins and paging, expression indexes against generated columns, patching, optimistic concurrency, and the single-writer constraint everything else follows from.
Real-world HTML arrives wrapped in navigation chrome, tracking attributes and inline CSS. ReverseMarkdown 6.2.1 adds Config.Preprocess, an ordered pipeline of DOM transformations applied to the source HTML before conversion: remove, unwrap, rename, recover CSS-only formatting, and resolve relative URLs.
Turn the same Marten + pgvector + Ollama stack from Part 1 into a question-answering assistant over the Marten documentation. The new ingredient is chunking: splitting long markdown pages into retrievable passages, then citing the exact page each answer draws from.
Build semantic search and a small RAG pipeline over 42,000 movie plots using nothing but Postgres. Marten's new pgvector package stores and searches the embeddings, Ollama generates them locally, and there is not an API key in sight.
The structural problem DCB solves: business rules that span multiple aggregates and cannot be defended by per-stream optimistic concurrency.
We build the coupon redemption command from Part 1 using Marten's DCB API: tag types, BoundaryAggregate, FetchForWritingByTags, and the explicit ConcurrencyException retry cycle.
We rebuild the redemption command from Part 2 using Wolverine.HTTP with a [WolverinePost] endpoint, where the HTTP route, the DCB consistency boundary, and the business decision live in one static method. Includes an Alba HTTP test.
Tag storage modes, the global-lock failure mode of over-broad queries, tag governance, schema evolution, incremental migration from existing aggregates, and when DCB is the wrong tool.
ReverseMarkdown 5.0.0 is out! It’s our biggest leap in years - dramatically faster, more resilient on edge cases, and friendlier .NET HTML to Markdown converter library. Upto to 200× faster on huge HTML via TextWriter refactor. Fixed nested <p> exp...
Protecting Personally Identifiable Information (PII) has become essential as organizations handle increasing volumes of sensitive data. Regulations like the GDPR (General Data Protection Regulation) mandate strict controls over PII to safeguard indiv...
Protecting Personally Identifiable Information (PII) has become essential as organizations handle increasing volumes of sensitive data. Regulations like the GDPR (General Data Protection Regulation) mandate strict controls over PII to safeguard indiv...
Protecting Personally Identifiable Information (PII) has become essential as organizations handle increasing volumes of sensitive data. Regulations like the GDPR (General Data Protection Regulation) mandate strict controls over PII to safeguard indiv...
This is a continuation of the Marten native partial updates patching post. All the native patch operations are driven by a bunch of custom PL/pgSQL functions with prefix mt_jsonb_. Marten uses JSONB for all its JSON storage hence all th...
An interesting issue was raised in my ReverseMarkdown GitHub library repository wherein conversion of HTML table containing colspan in cells does not convert properly to Markdown. Let me show you a representative example of the HTML table content for...
Partial update or patching JSON involves making changes to specific parts of a JSON document without replacing the entire document. This is particularly useful when you only need to modify certain fields or elements within a large JSON object, rather...
I was quite intrigued by the recent version of [Bun](https://bun.sh/docs/bundler/executables#embedding-sqlite-databases) supporting a one liner call to embed a SQLite file and use it as part of a single file binary. This is super cool. See the code b...
MysticMind.PostgresEmbed v4.0 released with support for .NET 8 and few enhancements. This library provides minimum Postgres binaries for development or unit testing purposes. This library has had around 92K NuGet downloads so far. https://github.com/...
VitePress uses Shiki for syntax highlighting which is a fantastic library. Shiki automatically fetches language grammar from VSCode so the language names etc for the fenced code block comes from it. For .NET, the language names captured are csharp a...
Maintaining and managing documentation for software products or libraries require a good amount of time and discipline to keep it up to date with the latest code base. Typically code samples are the ones which can quickly go out of sync when it is ma...
NAnt was quite a popular build tool used by projects of all sizes targeting .NET Framework on Windows. I have written many complex build systems using NAnt in the past. But XML based DSL was quite clunky to use and maintain. Besides, it was always a ...
I have been a co-maintainer of .NET OSS library Marten and largely involved in taking care of the DevOps stuff pertaining to CI pipeline, build and docs system. In the early days, the build was driven using a RAKE build script written by Jeremy Mille...