You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
I know this has been previously discussed already but at least I need some clarifications.
First of all, it is not yet clear to me how would deno manage dependencies in a clear, transparent way? So that a developer could open one file and see the dependency tree.
Second, in it's current form, pulling packages from URLs isn't very redundant. If the server goes down you are dead in the water. Why not implement a package management similar to packman on Arch Linux (of course there are other similar package managers). Which pulls the package from a list of repositories provided either by the distribution or by the user.
The text was updated successfully, but these errors were encountered:
First of all, it is not yet clear to me how would deno manage dependencies in a clear, transparent way?
deps.ts is a good convention... but it should be a convention, to allow people to have different conventions.
So that a developer could open one file and see the dependency tree.
You don't get a dependency tree with a package.json, but still the capability is already built into Deno
deno info main.ts
Second, in it's current form, pulling packages from URLs isn't very redundant. If the server goes down you are dead in the water.
If the registry goes down, you're dead in the water. What makes a formal priority package registry special to a web server? Again it is discussed in the manual as well.
Which pulls the package from a list of repositories provided either by the distribution or by the user.
At the end of the day, those are URLs to servers. Not any different from having that exist in code.
I know this has been previously discussed already but at least I need some clarifications.
First of all, it is not yet clear to me how would deno manage dependencies in a clear, transparent way? So that a developer could open one file and see the dependency tree.
Second, in it's current form, pulling packages from URLs isn't very redundant. If the server goes down you are dead in the water. Why not implement a package management similar to packman on Arch Linux (of course there are other similar package managers). Which pulls the package from a list of repositories provided either by the distribution or by the user.
The text was updated successfully, but these errors were encountered: