A bit of recognition

[mageaustralia]

Hi Fabrizio,

I just wanted to say what you're pumping out is simply amazing. Just doing an upgrade to 25.11.0. now (site still in development). Obviously, there was a lot of breaking changes but I could rip through it pretty quickly with help from Claude Code. It feels good to modernise the codebase. I really don't know how you have the time to get so much done.

[fballiano]

Thank you so much for this message @trabulium!

Please let me know when you publish a store with Maho, I'd love to publish it on Maho's website.

Can I ask you what breaking changes you did find?

[mageaustralia]

Hi Fabrizio - here's a basic summary.

Summary Statistics

• Total files modified: 300+
• Total commits: 15+
• Zend_Db_Expr replacements: 100+
• Zend_Date replacements: 13
• PDF class updates: 775
• JavaScript files modernized: 4
• Validation replacements: 26
• Modules fixed: 12+

Zend_Mail/Zend_Mime → Symfony Mailer
Zend_Currency
Zend_Http_Client → Symfony HttpClient
Zend_Pdf → Laminas PDF
Zend_Db_Expr → Maho\Db\Expr
Zend_Serializer → Symfony Serializer
Zend_Date → DateTime
Zend_Mail/Zend_Mime → Symfony Mailer
Zend_Oauth → League OAuth2 Client (12 instances)
Zend_Currency (full migration) → Brick\Money

I'm. also working on a few things that I plan to open source for Maho and / or Submit as PR's . I've worked heavily with Claude Code for a long time now but when I was trying to use it with Maho, it would often still try and use Magento1 methods: ie: Zend / Prototype etc. So I've been working on this "Claude Skill" to help produce Maho modules from text prompt, convert Mage1 -> Maho modules quickly and easily with quite a low token count - The module maker uses claude to create a json/yaml file that it can then produce all files from - it adds fields to the grid and handles date, file upload, wysiwyg editor, crons, frontend and backend routes, side blocks , many to many relationships etc. - it allows "Prompt to basic CRUD module creation" in minutes.

https://github.com/trabulium/mahocommerce-extension-builder/tree/master

An example json can be found here:

https://github.com/trabulium/mahocommerce-extension-builder/blob/master/test-recipe-full.json

The other modules I am working on right now:

Point of Sale system built on Vue - Current features:

• Barcode + Fuzzy Search using Meilisearch - It supports (or will)

• Customer Group Pricing

• Split Payments

• Gift Certificates

• Coupon Codes

• Customisable Receipt Printouts

• Guest or Customer

• Switch to Different Store,

• Ability to find Disabled / Out of Stock Products etc.

• Here's some screenshots - colour scheme will improve. It's an Admin Maho Module at the moment. It's probably ~60-70% complete right now.

Additionally, to create the above - I've built out the following GraphQL APIs to support the POS (so it could be run as a mobile app later / offline sync etc)

./packages/maho/api-checkout/schema/checkout.graphqls
./packages/maho/api-customer/schema/customer.graphqls
./packages/maho/api-core/schema/base.graphqls
./packages/maho/api-cart/schema/cart.graphqls
./packages/maho/api-catalog/schema/catalog.graphqls <-- this uses Meilisearch with MySQL fallback if not available)

And also the Gift Card module with both QR Code / Barcode support, email, scheduling etc.

I will discuss with you if you want to integrate any of these into core Maho - I do think it would have numerous advantages to have them available in the core but obviously, it has to fit with your vision for Maho. In any case, they will be open sourced.

[fballiano]

wow, it's a loong post (thank you for writing it), let me try to address everything :-)

Zend_Oauth → League OAuth2 Client (12 instances)

I was wondering if we should migrate our oauth1 to oauth2 with that library, what do you think? should we keep both oauth1 and oauth2? is it important to have oauth2 as soon as possible?

Zend_Currency (full migration) → Brick\Money

When I removed Zend_Currency I came across the Money library but I made without it in order not to add another dependency. What did you like about it that made it necessary to your project?

https://github.com/trabulium/mahocommerce-extension-builder

🥹 this project is truly fantastic. Can I ask you why you choose python for it? Do you see it as part of the official toolchain, inside the mahocommerce github organization? How do you install it (both as a Claude Code skill or standalone)? Simply cloning the repo?

Point of Sale system

i was also thinking about something like this, I was wondering if it could be another frontend theme or if it should be another backend theme instead, you took a completely different route. As I don't love javascript frameworks I'd prefer for something like the POS-view to be integrated in the rest of the framework without it being something completely different, but your screenshots look amazing, prop to your work!

Additionally, to create the above - I've built out the following GraphQL APIs to support the POS (so it could be run as a mobile app later / offline sync etc)

A few months ago I tried to add GraphQL support to the core but my knowledge is too limited on it so I couldn't really understand if I was doing it the right way or not, and ended up throwing everything away. I was reading that GraphQL is no longer the go-to for this kind of thing, is it true? Is there something new? Would it be a good addition to the core in your point of view?
My test was an attempt of making every Maho model available (in an automatic generalized way) via GraphQL but does it even make sense? Would it even be possible?

And also the Gift Card module with both QR Code / Barcode support, email, scheduling etc.

I'd love for this to be part of the core, it would be an amazing improvement!

Thanks again for this very interesting infos and ideas!

[mageaustralia]

Zend_Oauth → League OAuth2 Client (12 instances)

Oauth1(a) seems to be deprecated everywhere, so I think Oauth2 is the obvious choice if Maho is going to be truly modern.

Zend_Currency (full migration) → Brick\Money

This was more Claude Code's choice than mine - but that said, Mage1 was notorious for Discount / Tax rounding issues off by 1c - this issue plagued me with Paypal - I didn't investigate the Brick/Money library that deeply but I typically prefer to offload to more specialised libraries that I feel have already solved the problems I'm likely to encounter

Can I ask you why you choose python for it?

This is / was a tough decision - Python is Claude's strongest language, I'd say and I wanted to get through it pretty quickly - actually I had an issue where I'd lost ~10 hours of work and had to rebuild it again and I just wanted to get it done. What I'm thinking of doing though is taking Marius' UMC - https://github.com/tzyganu/UMC1.9, porting it to Maho and getting something working where Claude can define everything and then UMC generates all the code.. UMC itself is a Magento 1.9 module so should be able to be used via the admin also. I forgot about Marius' module when I started building it.

i was also thinking about something like this, I was wondering if it could be another frontend theme or if it should be another backend theme instead, you took a completely different route. As I don't love javascript frameworks I'd prefer for something like the POS-view to be integrated in the rest of the framework without it being something completely different, but your screenshots look amazing, prop to your work!

This was a hard decision, tbh. We have used a POS from BoostMyShop for a long time and I extended it quite a lot, improving the slow searching, adding Gift card support etc. I also had a customer say it was difficult for his older staff to learn it (compared to Shopify's). BoostMyShop was reliant on jQuery - I also am not a fan of most modern JS frameworks myself but the rest of the world is, unfortunately - I wanted it to be dynamic, manage state etc. It's built as an admin module. Doing split payments, marking items / orders as partially shipped or fully shipped and numerous other stuff, it makes more sense being authenticated via adminhtml - Also, we associate staff login with the POS and the receipt shows who served the user. I've been working with Flutter on mobile / tablet a lot the last two years, so I also wanted an interface that I could potentially build mobile apps (frontend / pos / product manager) at some stage, so that's why I build the pos around a graphQL interface.

Also, so many people are sold on headless - for Magento2, Shopify or even Netsuite. People having the option, if they choose, to run headless on Maho I think again, presents Maho is a real modern alternative to Mage2 / Shopify. I'm relatively fresh to graphQL - I always found it a little confusing and not very approachable. The main advantage of graphQL is its ability to do 'subsets/filtering' of data requests similar to SQL.

The main takeaways:

• Over-fetching. REST forces you to fetch entire product objects when you only need price + inventory.
• GraphQL lets you ask for exactly what you need.
• N+1 queries. REST often requires multiple calls; GraphQL lets you batch efficiently in one request.
• Shopify aggressively promotes GraphQL as their primary API.
• Magento 2 PWA. M2's headless strategy (with GraphQL) became the official modern path.

I'm not really a big fan of Headless / PWA but others are and GraphQL potentially paves a way forward and once more, makes Maho lose it's Mage1 roots and stand out as a modern Ecommerce framework, imo.

Re: Gift Card module with both QR Code / Barcode support, email, scheduling etc.

I'd love for this to be part of the core, it would be an amazing improvement! Awesome!

• I wanted the POS to be integrated with Gift Cards so I felt like the GC module would have to be part of core so yes, I'm happy if you'd accept it. You may need to review and ensure it's up to your. standards. It's almost there. I just need to do a bit more testing and then I'll publish it. I'm just trying to improve, finish up the Meilisearch Module to work with the default theme / composer packaging etc first.

[fballiano]

Zend_Oauth → League OAuth2 Client (12 instances)

Oauth1(a) seems to be deprecated everywhere, so I think Oauth2 is the obvious choice if Maho is going to be truly modern.

yeah my idea was to add oauth2 next to 1.1 not to break BC but it's true that everybody already migrated.

Zend_Currency (full migration) → Brick\Money

If we find somewhere where we can integrate it, why not :-)

This is / was a tough decision - Python is Claude's strongest language, I'd say and I wanted to get through it pretty quickly - actually I had an issue where I'd lost ~10 hours of work and had to rebuild it again and I just wanted to get it done. What I'm thinking of doing though is taking Marius' UMC - https://github.com/tzyganu/UMC1.9, porting it to Maho and getting something working where Claude can define everything and then UMC generates all the code.. UMC itself is a Magento 1.9 module so should be able to be used via the admin also. I forgot about Marius' module when I started building it.

UMC would be amazing, I saw it many years ago but forgot about it too

Also, so many people are sold on headless

A lot of people are sold on the new shiny thing which disappears in a couple of years, I'm proud that we kept our compass straight so headless could be a nice addition but I wouldn't go full headless for everybody because of complexity

I'm not really a big fan of Headless / PWA but others are and GraphQL potentially paves a way forward and once more, makes Maho lose it's Mage1 roots and stand out as a modern Ecommerce framework, imo.

yeah for sure, I'd love to be able to expose all models as graphql endpoints in one shot, instead of maintaining every single entity (like SOAP/REST), which is a lot of work especially for maintenance. but i'm not sure it can be done :-\

• I wanted the POS to be integrated with Gift Cards so I felt like the GC module would have to be part of core so yes, I'm happy if you'd accept it. You may need to review and ensure it's up to your. standards. It's almost there. I just need to do a bit more testing and then I'll publish it. I'm just trying to improve, finish up the Meilisearch Module to work with the default theme / composer packaging etc first.

whenever you want :-) I'll be a bit busy in the next few weeks so I'll be a slower but still present :-)

[mageaustralia]

@fballiano -off topic but can you double check that mass updates on grids work for you after last update? For example, product grid, select two or more and try and edit attributes? It seems to send a comma separated string of numbers and not an array and says please select items. Same with other grids like cache, indexing grids etc.

I've got a fix for it but wanted to make sure it's not just me before I push the PR

[fballiano]

@trabulium seems you're right, thanks for telling me!