Re: node-tty-browserify_0.0.0-1_amd64.changes REJECTED
- Date: Thu, 9 Feb 2017 22:23:49 +0100
- From: Adam Borowski <kilobyte@xxxxxxxxxx>
- Subject: Re: node-tty-browserify_0.0.0-1_amd64.changes REJECTED
(slightly trimmed To:/CC: -- no redundant copies please!)
On Thu, Feb 09, 2017 at 08:55:25PM +0100, Philip Hands wrote:
> Pirate Praveen <praveen@xxxxxxxxxxxxxxxx> writes:
> > "This module is a dependency for browserify." is already present in
> > the description. And short description says "tty module from node core
> > for browsers".
> Then again, they could try looking at the code, which astonishingly does
> nothing other than throw errors about how none of this is implemented,
> except in the case of the one function that does anything ('isatty')
> which unconditionally returns false.
> That being the case, one might have hoped for a description saying
> something like:
> stub version of tty, to (barely) satisfy browserify's dependencies
Now this is a new low.
Let's skip the talk about bogus descriptions for a bit, there's a bigger
problem here: the package doesn't do _anything_ it is supposed to do, just
returns error on every call. What's the point of having it in the archive?
It changes the usual node.js principle from "do one thing and do it wrong"
to "just fail".
I hope you do know you can patch the code, including the import line?
If a dependency is not needed, it's far better to get rid of it than to
package and pull in a non-functional stub.
A stub is useful when it's an alternative the user can replace; here you
have it as a hard-coded dependency. Please patch it out instead.
Autotools hint: to do a zx-spectrum build on a pdp11 host, type:
./configure --host=zx-spectrum --build=pdp11