npmjs is awesome, thanks!
Was wondering what the chances would be of your adding the ability to query arbitrary package.json properties (to find packages with their existence or with specific values, and ideally returning the packages and/or those values) across a registry (and do you support querying other registries)?
I'd love to be able to add or utilize generic service discovery mechanisms to package.json which worked with the fixed versioning nature of npm.
One use case I have is allowing packages to designate themselves as, say, being string manipulation functions, and then any tool, whether an extensible Atom context menu package or browser context menu tool, email program menu, etc., could provide discovery of any simple functions or suites of functions that will manipulate text (whether upper-casing, stripping entities, inserting the current date, etc.) without individual add-ons needing to reinvent the wheel for each environment.
(I specifically like the idea of leveraging the benefits of versioning in npm with the security features of deno so that tools could query npm registries to find deno-specific code which could be run locally without fear of escalating to network or file system privileges; without actually installing from npm, they could download the package.json designated deno file for string manipulation (downloading and extracting out of the tgz if the API couldn't itself just return a specific file) to cache for local execution by the tool.)
npmjs is awesome, thanks!
Was wondering what the chances would be of your adding the ability to query arbitrary
package.jsonproperties (to find packages with their existence or with specific values, and ideally returning the packages and/or those values) across a registry (and do you support querying other registries)?I'd love to be able to add or utilize generic service discovery mechanisms to
package.jsonwhich worked with the fixed versioning nature of npm.One use case I have is allowing packages to designate themselves as, say, being string manipulation functions, and then any tool, whether an extensible Atom context menu package or browser context menu tool, email program menu, etc., could provide discovery of any simple functions or suites of functions that will manipulate text (whether upper-casing, stripping entities, inserting the current date, etc.) without individual add-ons needing to reinvent the wheel for each environment.
(I specifically like the idea of leveraging the benefits of versioning in npm with the security features of deno so that tools could query npm registries to find deno-specific code which could be run locally without fear of escalating to network or file system privileges; without actually installing from npm, they could download the
package.jsondesignated deno file for string manipulation (downloading and extracting out of the tgz if the API couldn't itself just return a specific file) to cache for local execution by the tool.)