No, you are using /g in list context which swallows the whole file now and returns a flat list.

You need an external loop to process the 3 groups

update

With a post while you can have the loop in the same line

 push @res, [ $1,$2,$3 ] while m/.../g

In the OP you were originally looping with $_ , that's confusing.

you should rather be concerned about maintainability than shortness.

you should rather be concerned about maintainability than shortness.

Agreed.

Still, I am fascinated by idiomatic Perl and its myriad other ways of doing something (the Schwartzian Transform comes to mind).

Searched for donut and crumpit. Found donate and stumbit instead.

Here's what I came up with to "post-process" the array to get what I wanted:

#! /usr/bin/perl -w

use strict;
use Data::Dump qw(dump);

my @src = (
  "/projects/2138137193/iblab?ref=discovery",
  2138137193,
  "iBlab",
  "/projects/2060538158/nz-hosting?ref=discovery",
  2060538158,
  "NZ Hosting",
  "/projects/1956727289/how-to-build-a-spaceship?ref=discovery",
  1956727289,
  "How To Build A Spaceship"
);
my @dst = ();
my $by  = 3;
my $len = @src / $by;

die "Not an integral multiple\n" unless ( 0 == @src % $by );

for (my $i = 0; $i < $len; $i++) {
  push  @dst, [ splice @src, 0, 3 ];
}
dump \@dst;
[download]

The output:

[
  [
    "/projects/2138137193/iblab?ref=discovery",
    2138137193,
    "iBlab",
  ],
  [
    "/projects/2060538158/nz-hosting?ref=discovery",
    2060538158,
    "NZ Hosting",
  ],
  [
    "/projects/1956727289/how-to-build-a-spaceship?ref=discovery",
    1956727289,
    "How To Build A Spaceship",
  ],
]
[download]

I don't know why but lately I've been in a mood to "do it in one statement" or "do it in place". So I have to ask: "Can this be done in-place with just one array?"