announcements 2022-09-04

djanus16:09:15

I’ve released version 0.3.3 of https://github.com/nathell/skyscraper, a framework for structurally scraping whole sites in Clojure. Get it from Clojars: {skyscraper/skyscraper {:mvn/version "0.3.3"}} Changes in this release: • Feature: To facilitate debugging, processors can now set the :skyscraper/description key on contexts. These descriptions will be logged when downloading, instead of the URL, and won’t be propagated to child contexts. • Fix: Skyscraper now properly closes the cache when using scrape! and one of the processors throws an exception. • Fix: Skyscraper no longer complains when the server returns a quoted charset in the Content-type header. • Fix: :skyscraper.traverse/priority is no longer propagated to child contexts. • Infra: Skyscraper’s dependencies are now managed with cli-tools, with Kaocha being used for testing.

🎉 6

djanus19:09:32

Turned out I rushed with the announcement. Due to a misconfiguration (a failed attempt at migrating away from Leiningen to deps.edn), the pom.xml from 0.3.3 declared no dependencies. I've now pushed 0.3.4 which has a correct dependency listing and is otherwise the same as 0.3.3. Apologies and happy scraping!

seancorfield17:09:06

com.seancorfield/honeysql {:mvn/version "2.3.928"} -- turn Clojure data structures into SQL: • Address https://github.com/seancorfield/honeysql/issues/425 by clarifying that INTERVAL as special syntax may be MySQL-specific and PostgreSQL uses difference syntax (because INTERVAL is a data type there). • Address https://github.com/seancorfield/honeysql/issues/423 by supporting DEFAULT values and DEFAULT rows in VALUES. • Address https://github.com/seancorfield/honeysql/issues/422 by auto-quoting unusual entity names when :quoted (and :dialect) are not specified, making HoneySQL more secure by default. • Fix https://github.com/seancorfield/honeysql/issues/421 by adding :replace-into for :mysql dialect. • Address https://github.com/seancorfield/honeysql/issues/419 by adding honey.sql.protocols and InlineValue with a sqlize function. • Address https://github.com/seancorfield/honeysql/issues/413 by flagging a lack of WHERE clause for DELETE, DELETE FROM, and UPDATE when :checking :basic (or :checking :strict). • Fix https://github.com/seancorfield/honeysql/issues/392 by adding support for WITH / (`NOT`) MATERIALIZED -- via PR https://github.com/seancorfield/honeysql/issues/420 https://github.com/robhanlon22. Follow-up in #honeysql

❤️ 10

👍 4

🎉 7

djanus18:09:33

And I have another announcement to make! Announcing https://github.com/nathell/cartestian, a small experimental library that helps write more comprehensive tests. From the README: > In the life of a developer who writes tests for their software, there are always two (or more) possibilities. In an A/B test, either your user is assigned to group A, or to group B. Either that API call succeeds, or it fails. The list goes on. > Oftentimes, we are tempted to only test the happy path. But it would be good to test all the combinations – a Cartesian product of the “dimensions”, if you will. Or at least a reasonably-sized subset of them, if the possibility space is large. > This library helps you to do just that. I’d love to hear any feedback on this one. Do people think it’s a good idea? The library is so experimental that I’m not yet building any artifacts nor publishing them on Clojars, but if there’s interest I’ll start doing so.

🎉 1

seancorfield18:09:57

This is interesting. I think one of my biggest problems when trying to test across combinations is how to specify what the expected outcome should be, especially if there's a failure mode that throws an exception. How are you typically mapping from "possibilities" to outcomes in your combination tests?

djanus19:09:13

Great question! It will broadly depend on the particular combination of possibilities (the v in the example), but how exactly depends on the test. For example, in https://github.com/nathell/skyscraper/blob/master/test/skyscraper/cache_test.clj#L59-L75, I’m checking for the :succeeding dimension to determine expected behaviour. Admittedly a more advanced example would be in order, though.

Max13:09:01

This seems similar in spirit to property-based testing (eg test.check). I'm interested in your thoughts on the overlaps and differences with this approach?

djanus16:09:34

I was thinking this too! Think this is a “sub-concept” of property-based testing: more coarse-grained and arguably easier to grasp. I found writing a test.check generator can be tricky. But I guess cartestian could be theoretically rewritten in terms of test.check. 🙂

borkdude21:09:27

A new version of #babashka! Lots of updates, bug fixes and error location improvements. 0.9.162 (2022-09-04) • https://github.com/babashka/babashka/issues/1343: Fix postgres feature • https://github.com/babashka/babashka/issues/1345: add .ssl.SSLException and java.net.SocketTimeoutException classes (https://github.com/lread) • Fix satisfies? with marker protocol (no methods) • Update rewrite-clj • Update deps.clj • Update babashka.cli • Update org.clj-commons/clj-yaml • babashka.fs: fix expand-home on Windows • babashka.fs: expose :win-exts • nREPL: preserve stacktrace on exception • Fix https://github.com/babashka/babashka/issues/1322: improve error location • Fix https://github.com/babashka/babashka/issues/1338: add-watch segfault • Fix https://github.com/babashka/babashka/issues/1339: resolve record name ending with dot. (Also, there's a new project brimming in #babashka-bbin which will be announced later, after more people test it.)

🎉 17

2022-09-04

Channels