Project Upgrade Approach for Modified Theia 0.12.0-Based Project

[zazzzf]

I’m upgrading a project based on Theia 0.12.0 with significant customizations, including:

Removed Node modules and dependencies for certain features, replacing them with static assets for distribution.
Modified the filesystem to fetch data from a custom server source.
Replaced the terminal module to use server-based data sources.
Customized the core module, including layout, styling, and dimensions.
I need advice on choosing the right upgrade strategy:

Option 1: Use theia-extension to create a base project. Can this approach allow deep customization of core and terminal modules to meet my needs?

Option 2: Directly modify theia-ide code. Would this lock me out of future Theia version updates (e.g., due to compatibility issues)?

[msujew]

Hey @zazzzf,

the general recommendation is to consume the @theia/* packages as dependencies and use the dependency injection mechanism offered to adjust the behavior of the services. Both of the options that you've outlined follow this recommendation. The theia-extension yeoman generator will setup a very basic project, while a fork of theia-ide will simply give you a bit more out of the box. However, both are simply pulling the main @theia/* packages from npm and building an IDE product on top of that. Neither of them will lock you out of future Theia versions.

Can this approach allow deep customization of core and terminal modules to meet my needs?

Generally yes. At least I don't see any issues with being able to implement the outlined customizations using the DI mechanism.

[zazzzf]

Thank you, I understand now. Regardless of the approach, we should treat the @theia/* modules as modules and reference them properly, adjusting their functionality via dependency injection to meet our business requirements. The old approach of directly modifying the module's code in the past project was an improper implementation.