- Pharo-dev - lists.pharo.org

Re: Smalltalk & clawdbot

by Tim Mackinnon

Hi John - it sounds like we should sign you up for a future UKST session as there seems to be enough different to warrant a session in its own right. Certainly the implications of LLM's on all programming environments are not going away and we all need to understand how to leverage the best bits. Tim On Sun, 25 Jan 2026, at 9:38 AM, John M McIntosh wrote: > Runs the Squeak 6.0 environment on Linux. > Access from your telegram account > > https://clawdhub.com/johnmci/smalltalk > > Clawdbot did all the work, now an approved contributor to https://github.com/CorporateSmalltalkConsultingLtd/ClaudeSmalltalk > > > .... > John M. McIntosh. Corporate Smalltalk Consulting Ltd https://www.linkedin.com/in/smalltalk > > > Sent from Proton Mail <https://proton.me/mail/home> for iOS.

Jan. 26, 2026

Smalltalk & clawdbot

by John M McIntosh

Runs the Squeak 6.0 environment on Linux. Access from your telegram account https://clawdhub.com/johnmci/smalltalk Clawdbot did all the work, now an approved contributor to https://github.com/CorporateSmalltalkConsultingLtd/ClaudeSmalltalk .... John M. McIntosh. Corporate Smalltalk Consulting Ltd https://www.linkedin.com/in/smalltalk Sent from [Proton Mail](https://proton.me/mail/home) for iOS.

Jan. 25, 2026

Re: AI systems & Smalltalk

by John M McIntosh

Well seems there are other offerings https://github.com/mumez/pharo-smalltalk-interop-mcp-server A shift in how developers think has to occur, Write me a class that I can use to dump all the stacks that arenât housekeeping ones based on a file being touched. Or examine how random numbers are made and the which algorithm is being used. Also identify non cryptography safe usages in the system and why. Look at the network class, identify any race conditions, suggest how to ensure it is green thread safe, give me a new class. -> it suggested worker queues, read/writer processâ¦ Iâm using stdin/stdout each LLM session spins up a squeak/Cuis VM change file /dev/null to avoid conflicts but image shouldnât be altered on disk anyway. Oh and take this MQTT v3.1.1 Smalltalk implementation give me a V5. That took 45 minutes, will release in a few weeks. Ya SIGUSR1 .... John M. McIntosh. Corporate Smalltalk Consulting Ltd https://www.linkedin.com/in/smalltalk Sent from [Proton Mail](https://proton.me/mail/home) for iOS. -------- Original Message -------- On Friday, 01/23/26 at 05:47 Tim Mackinnon <tim(a)testit.works> wrote: > Hi John - this sounds very interesting and the next UKSTUG meeting is on the 28th cover AI in Smalltalk, so I will mention this and see how comparable/complementary that work is - ideally leading to some folks making your solution run in Pharo as well. (of course - might be interesting for you to join us on the 28th too if you have time) > > Tim > > On Fri, 23 Jan 2026, at 3:24 AM, John M McIntosh wrote: > >> This evening I added ChatGPT support to our work on AI systems talking to Squeak and Cuis systems. >> Likely 10 minutes to enable Pharo to join the party, any takers? >> https://github.com/CorporateSmalltalkConsultingLtd/ClaudeSmalltalk >> >> .... >> John M. McIntosh. Corporate Smalltalk Consulting Ltd https://www.linkedin.com/in/smalltalk

Jan. 24, 2026

Re: AI systems & Smalltalk

by Kasper Osterbye

I have been trying my stuff on Ollama, Mistral, Gemini, OpenAI, Claude and Grok, and some others. And for each of these I have tried four of their models. All are different, and all comparable. I hope to get some questions I cannot answer - such are mostly the most interesting. Best, Kasper > On 23 Jan 2026, at 15.05, omar abedelkader <omar.abedelkader(a)inria.fr> wrote: > > Hi > > Iâm working on ChatPharo, which uses a different Agent in the backend, Ollama, Gemini, and Mistral > > https://github.com/omarabedelkader/ChatPharo > > -- > OA > >> Le 23 janv. 2026 Ã 14:46, Tim Mackinnon <tim(a)testit.works> a Ã©crit : >> >> Hi John - this sounds very interesting and the next UKSTUG meeting is on the 28th cover AI in Smalltalk, so I will mention this and see how comparable/complementary that work is - ideally leading to some folks making your solution run in Pharo as well. (of course - might be interesting for you to join us on the 28th too if you have time) >> >> Tim >> >> On Fri, 23 Jan 2026, at 3:24 AM, John M McIntosh wrote: >>> This evening I added ChatGPT support to our work on AI systems talking to Squeak and Cuis systems. >>> Likely 10 minutes to enable Pharo to join the party, any takers? >>> https://github.com/CorporateSmalltalkConsultingLtd/ClaudeSmalltalk >>> >>> .... >>> John M. McIntosh. Corporate Smalltalk Consulting Ltd https://www.linkedin.com/in/smalltalk >>> >> >

Jan. 23, 2026

Re: AI systems & Smalltalk

by omar abedelkader

Hi Iâm working on ChatPharo, which uses a different Agent in the backend, Ollama, Gemini, and Mistral https://github.com/omarabedelkader/ChatPharo -- OA > Le 23 janv. 2026 Ã 14:46, Tim Mackinnon <tim(a)testit.works> a Ã©crit : > > Hi John - this sounds very interesting and the next UKSTUG meeting is on the 28th cover AI in Smalltalk, so I will mention this and see how comparable/complementary that work is - ideally leading to some folks making your solution run in Pharo as well. (of course - might be interesting for you to join us on the 28th too if you have time) > > Tim > > On Fri, 23 Jan 2026, at 3:24 AM, John M McIntosh wrote: >> This evening I added ChatGPT support to our work on AI systems talking to Squeak and Cuis systems. >> Likely 10 minutes to enable Pharo to join the party, any takers? >> https://github.com/CorporateSmalltalkConsultingLtd/ClaudeSmalltalk >> >> .... >> John M. McIntosh. Corporate Smalltalk Consulting Ltd https://www.linkedin.com/in/smalltalk >> >

Jan. 23, 2026

Re: AI systems & Smalltalk

by Tim Mackinnon

Hi John - this sounds very interesting and the next UKSTUG meeting is on the 28th cover AI in Smalltalk, so I will mention this and see how comparable/complementary that work is - ideally leading to some folks making your solution run in Pharo as well. (of course - might be interesting for you to join us on the 28th too if you have time) Tim On Fri, 23 Jan 2026, at 3:24 AM, John M McIntosh wrote: > This evening I added ChatGPT support to our work on AI systems talking to Squeak and Cuis systems. > Likely 10 minutes to enable Pharo to join the party, any takers? > https://github.com/CorporateSmalltalkConsultingLtd/ClaudeSmalltalk > > .... > John M. McIntosh. Corporate Smalltalk Consulting Ltd https://www.linkedin.com/in/smalltalk >

Jan. 23, 2026

AI systems & Smalltalk

by John M McIntosh

This evening I added ChatGPT support to our work on AI systems talking to Squeak and Cuis systems. Likely 10 minutes to enable Pharo to join the party, any takers?https://github.com/CorporateSmalltalkConsultingLtd/ClaudeSmalltalk .... John M. McIntosh. Corporate Smalltalk Consulting Ltd https://www.linkedin.com/in/smalltalk

Jan. 23, 2026

Claude & Smalltalk

by John M McIntosh

Hi, if you've been watch Claude Code taking over the world, missing interacting with Smalltalk, then I've build out an interface so you can talk to Cuis Smalltalk. However the Python MCP bridge interfaces to a Smalltalk solution via MQTT so a bit of work could be done to enable one to talk to a Pharo image. https://github.com/CorporateSmalltalkConsultingLtd/SmalltalkCuisMCP .... John M. McIntosh. Corporate Smalltalk Consulting Ltd https://www.linkedin.com/in/smalltalk

Jan. 21, 2026

Re: Looking for Metacello beta testers

by stephane ducasse

Hi Gabriel Happy new year > Nice, Metacello needs some love. Indeed :) And we are making progress. > I will try with our repos and report back if any problems arise. Excellent. > > Are there any docs or details on the strict mode? Not that I remember. Guille can tell us. > Is it enabled by default, No because nothing would work :) > or do we need to enable it per repo? I can give it a try too. I will ask Guille to tell us. > > Regards, > Gabriel > > On Mon, Dec 29, 2025 at 5:55â¯AM stephane ducasse <stephane.ducasse(a)inria.fr <mailto:stephane.ducasse@inria.fr>> wrote: >> Hello >> >> Under the radar, we are actively continiously cleaning Pharo deep core so that we can work >> on new features. >> >> For metacello, in Pharo 13 we introduced a strict mode (so that we can detect hidden dependencies), >> and other loading features to support module experimentations. It will make Pharo code more robust >> in the long term. >> >> Since a couple of months we are actively simplifying/cleaning Metacello (as the version available in Pharo) because we want to introduce new features such as the possibility to force the reload of baselines and others. >> >> Now the unit test coverage of metacello is quite low and we do not use/know all the features >> that are available and that you may rely on. >> >> So we would love to have beta testers that can just try to load their code in Pharo 14 alpha >> and report potential bugs or features drop we would have introduced. >> >> So do not hesitate and this is a nice way to help pharo >> >> S.

Jan. 4, 2026

Re: Looking for Metacello beta testers

by Gabriel Cotelli

Nice, Metacello needs some love. I will try with our repos and report back if any problems arise. Are there any docs or details on the strict mode? Is it enabled by default, or do we need to enable it per repo? I can give it a try too. Regards, Gabriel On Mon, Dec 29, 2025 at 5:55â¯AM stephane ducasse <stephane.ducasse(a)inria.fr> wrote: > Hello > > Under the radar, we are actively continiously cleaning Pharo deep core so > that we can work > on new features. > > For metacello, in Pharo 13 we introduced a strict mode (so that we can > detect hidden dependencies), > and other loading features to support module experimentations. It will > make Pharo code more robust > in the long term. > > Since a couple of months we are actively simplifying/cleaning Metacello > (as the version available in Pharo) because we want to introduce new > features such as the possibility to force the reload of baselines and > others. > > Now the unit test coverage of metacello is quite low and we do not > use/know all the features > that are available and that you may rely on. > > So we would love to have beta testers that can just try to load their code > in Pharo 14 alpha > and report potential bugs or features drop we would have introduced. > > So do not hesitate and this is a nice way to help pharo > > S.

Jan. 2, 2026