August 2026 - Pharo-users - lists.pharo.org

Pharo 14

by Steffen Märcker

Hi! I just noticed that the launcher now offers Pharo 14 stable. Is this already the official release? I haven't seen any announcement yet. Cheers! Steffen

Aug. 20, 2026

UKSTUG Meeting: David Long - Recovering the Design of Existing Software: SecureCell, SageTea and the Road Back to Pharo - 26 August 2026

by UK Smalltalk

Much of today's software is locked inside large, evolving code bases where the original design has gradually been lost. What if we could automatically recover that design, represent it using a language-independent application model, and then allow a modern runtime to render, evolve and maintain the application independently of its original implementation language? This talk presents that vision through SecureCell, a secure communications appliance developed by XTC Mobile for organizations that require privacy, security and local control. David Long will demonstrate how Graphify analyzes existing applications to recover their architectural structure, how this information is transformed into the SageTea Canonical Application Model, and how the runtime can automatically generate user interfaces, navigation and application behavior from that model. Rather than treating software as language-specific source code, the emphasis is on recovering and preserving the application's semantic design. Finally, David will discuss why this architectural approach naturally leads back to Pharo Smalltalk, and why we believe Smalltalk remains an ideal environment for building the next generation of language-independent development tools and self-hosting runtime systems. The session will include a live demonstration of SecureCell and an overview of the current work on the Eclipse SageTea Runtime. David Long is the creator of the Eclipse SageTea project and CEO of XTC Mobile, where he leads the development of secure mobile communications solutions for government, defence, telecommunications carriers and regulated enterprises. A Smalltalk developer for more than thirty years, David is currently developing the next generation of the Eclipse SageTea Runtime—an open-source platform for language-independent application modeling, AI-assisted software modernization and long-term software evolution. His current work focuses on Graphify, canonical application models, runtime architecture and bringing the SageTea Runtime to Pharo Smalltalk. This will be an online meeting. If you'd like to join us, please sign up in advance on the meeting's Meetup page ( https://www.meetup.com/ukstug/events/313860371/ ) to receive the meeting details

Aug. 19, 2026

Re: Benchmarks/Tests for Base Image

by Steffen Märcker

Thanks! I will definitely have a look for the Pharo port of Transducers and a future paper. Ciao, Steffen Guillermo Polito schrieb am Mittwoch, 12. August 2026 16:07:59 (+02:00): Hi Steffen, For Pharo you can take a look at - smark https://github.com/guillep/smark - veritas https://github.com/jordanmontt/PharoVeritasBenchSuite/tree/main Not specifically microbenchmarks of the enumeration protocol though. G Le 10 août 2026 à 7:36 PM, Steffen Märcker via Pharo-users <pharo-users(a)lists.pharo.org> a écrit : Hi! I presented my work on Transducers at IWST recently and I am preparing the final version of the corresponding paper. A week spot of that work is in the evaluation: I replaced the common enumeration methods like select: and collect: with versions using transducers. The image ran stable and I observed no perceptible slowdowns. However, I'd very much like to back this anecdotal evidence with some real data or measurements. Hence: Are you aware of a publicly available benchmark or test suite for the base VW image? Or something covering the common enumeration protocol? Do you have an alternative idea? (Same goes for Pharo at a later date.) Happy coding! Steffen

Aug. 18, 2026

Re: [Esug-list] All Smalltalk Report PDF archive online

by sducasseatwork

Le 09/08/2026 à 07:56, Kasper Osterbye via Pharo-users a écrit : > Very nice idea. > > Unfortunately the pdf files from 91-95 cannot be read on GitHub, but > must be downloaded. Thanks for the feedback. I will see if I can do something about it. Any idea is welcome. S > The pdf's from 96 are readable on GitHub directly. > > I am Mac, and it was not readable in neither Safari, Opera or Google > Chrome. > On the 91_95 the "INDEX.PDF" is readable. > > -- Kasper > >> On 31 Jul 2026, at 17.05, Sean DeNigris via Pharo-users >> <pharo-users(a)lists.pharo.org> wrote: >> >>> I just published all the archive of the old Smalltalk reports I >>> found on my disc: this is nice to be able to browse all such small >>> (yet interesting) articles around Smalltalk (91-96) >> >> Wow, thanks! This is sooo cool :) There were some extra chars in the >> URL; this one works: >> https://github.com/Ducasse/SmalltalkReportPDFArchive >> >

Aug. 14, 2026

Re: Benchmarks/Tests for Base Image

by Guillermo Polito

Hi Steffen, For Pharo you can take a look at - smark https://github.com/guillep/smark - veritas https://github.com/jordanmontt/PharoVeritasBenchSuite/tree/main Not specifically microbenchmarks of the enumeration protocol though. G > Le 10 août 2026 à 7:36 PM, Steffen Märcker via Pharo-users <pharo-users(a)lists.pharo.org> a écrit : > > Hi! > > I presented my work on Transducers at IWST recently and I am preparing the final version of the corresponding paper. A week spot of that work is in the evaluation: I replaced the common enumeration methods like select: and collect: with versions using transducers. The image ran stable and I observed no perceptible slowdowns. However, I'd very much like to back this anecdotal evidence with some real data or measurements. Hence: > > Are you aware of a publicly available benchmark or test suite for the base VW image? Or something covering the common enumeration protocol? Do you have an alternative idea? (Same goes for Pharo at a later date.) > > Happy coding! > Steffen >

Aug. 12, 2026

Benchmarks/Tests for Base Image

by Steffen Märcker

Hi! I presented my work on Transducers at IWST recently and I am preparing the final version of the corresponding paper. A week spot of that work is in the evaluation: I replaced the common enumeration methods like select: and collect: with versions using transducers. The image ran stable and I observed no perceptible slowdowns. However, I'd very much like to back this anecdotal evidence with some real data or measurements. Hence: Are you aware of a publicly available benchmark or test suite for the base VW image? Or something covering the common enumeration protocol? Do you have an alternative idea? (Same goes for Pharo at a later date.) Happy coding! Steffen

Aug. 10, 2026

Re: [Esug-list] All Smalltalk Report PDF archive online

by Kasper Osterbye

Very nice idea. Unfortunately the pdf files from 91-95 cannot be read on GitHub, but must be downloaded. The pdf's from 96 are readable on GitHub directly. I am Mac, and it was not readable in neither Safari, Opera or Google Chrome. On the 91_95 the "INDEX.PDF" is readable. -- Kasper > On 31 Jul 2026, at 17.05, Sean DeNigris via Pharo-users <pharo-users(a)lists.pharo.org> wrote: > >> I just published all the archive of the old Smalltalk reports I found on my disc: this is nice to be able to browse all such small (yet interesting) articles around Smalltalk (91-96) > > Wow, thanks! This is sooo cool :) There were some extra chars in the URL; this one works: https://github.com/Ducasse/SmalltalkReportPDFArchive >

Aug. 9, 2026

Looking for my next Smalltalk opportunity

by Sebastian Jordan

Hi everyone, After almost seven years being part of the Pharo community, I wanted to reach out here because this community has been an important part of my professional journey. I am currently finishing my PhD at INRIA, France, and I am looking for my next opportunity as a software developer. After completing my PhD, I will be moving back to LATAM, Bolivia, and I am looking for a fully remote position starting around mid-October or November. I would love to continue working with Smalltalk. Most of my experience is with Pharo, but I am open to any Smalltalk dialect and different kinds of projects. I am also open to full-time positions, contracts, or consulting work. Over the years, I have contributed to several parts of the Pharo ecosystem. Before my PhD, I worked as an R&D software engineer at INRIA on Pharo, where I developed tools for code analysis and transformation, contributed to Moose and pharo-ai. I have mentored students through Google Summer of Code since 2021. I have continued to work on programming languages and developer tools throughout my PhD, focusing on runtime-related issues, instrumentation, profiling (dynamic analysis), and memory analysis. Solving difficult problems, designing clear object-oriented systems, writing tests, improving existing codebases through refactoring, and building tools that increase developer productivity are all aspects of software development that I enjoy. I also enjoy working on enterprise applications where maintainability and good design are important. If you know of a team looking for someone with my background, or if you would simply like to discuss possible opportunities, I would be very happy to hear from you. I can also send my CV privately to anyone interested. Thanks everyone for your help, and I hope to continue being part of the Smalltalk community. Best regards, Sebastian JORDAN MONTAÑO

Aug. 5, 2026

OpenAPI now validates and tracks JSONSchema master

by Norbert Hartl

Hi all, OpenAPI (https://github.com/zweidenker/OpenAPI) got the same treatment as JSONSchema recently: the baseline dropped its stale 2020-era JSONSchema pin and now tracks JSONSchema's master, and the library actually validates whole documents now instead of just parsing them. What's new: OADocumentValidator checks a full document against the official OAI 3.0 meta-schema (bundled, Apache-2.0), with real error messages, not just true/false. Verified against six real OpenAPI 3.0 documents from the official spec repo (petstore, petstore-expanded, api-with-examples, callback-example, link-example, uspto) - round-tripping them surfaced and fixed several real bugs that hand-written test documents never would have. OpenAPI-Client was exercised against a real spec (Stripe's) end-to-end, which is what it took to find that request bodies were JSON-only when most real APIs use form-urlencoded bodies. That's fixed too. Scope is still explicitly OpenAPI 3.0. 3.1/3.2 use a different schema dialect (JSON Schema 2020-12 + OAS vocabulary - $dynamicRef, $ref-with-siblings, unevaluated*), which needs machinery JSONSchema doesn't have yet. Not pursuing that for now is a deliberate decision, not an oversight. Code and README (installation, quick start, what's not supported) are on GitHub: https://github.com/zweidenker/OpenAPI Norbert

Aug. 1, 2026