add path based redirects

up to now, those were managed outside of this repo, which as
unsurprisingly a real hassle to deal with if one wanted to prevent URLs
from breaking when moving pages around. this change removes a large part
of the friction involved in moving content in the Nix manual.

possible next steps for further automation:
- check for content that moved and warn if it's not reachable from
  links that were valid prior to a change
- create redirect rules automatically based on this information

(cherry picked from commit 2b7016cc56)

doc/manual/_redirects

@@ -0,0 +1,30 @@
+# redirect rules for paths (server-side) to prevent link rot.
+# see ./redirects.js for redirects based on URL fragments (client-side)
+#
+# concrete user story this supports:
+# - user finds URL to the manual for Nix x.y
+# - Nix x.z (z > y) is the most recent release
+# - updating the version in the URL will show the right thing
+#
+# format documentation:
+# - https://docs.netlify.com/routing/redirects/#syntax-for-the-redirects-file
+# - https://docs.netlify.com/routing/redirects/redirect-options/
+#
+# conventions:
+# - always force (<CODE>!) since this allows re-using file names
+# - group related paths to ease readability
+# - always append new redirects to the end of the file
+# - redirects that should have been there but are missing can be inserted where they belong
+
+/expressions/expression-language /language/ 301!
+/expressions/language-values /language/values 301!
+/expressions/language-constructs /language/constructs 301!
+/expressions/language-operators /language/operators 301!
+/expressions/* /language/:splat 301!
+
+/package-management/basic-package-mgmt /command-ref/nix-env 301!
+
+/package-management/channels* /command-ref/nix-channel 301!
+
+/package-management/s3-substituter* /command-ref/new-cli/nix3-help-stores#s3-binary-cache-store 301!
+

doc/manual/redirects.js

@@ -1,7 +1,9 @@
-// redirect rules for anchors ensure backwards compatibility of URLs.
+// redirect rules for URL fragments (client-side) to prevent link rot.
-// this must be done on the client side, as web servers do not see the anchor part of the URL.
+// this must be done on the client side, as web servers do not see the fragment part of the URL.
+// it will only work with JavaScript enabled in the browser, but this is the best we can do here.
+// see ./_redirects for path redirects (client-side)
 
-// redirections are declared as follows:
+// redirects are declared as follows:
 // each entry has as its key a path matching the requested URL path, relative to the mdBook document root.
 //
 //     IMPORTANT: it must specify the full path with file name and suffix