unofficial mirror of bug-guix@gnu.org 
 help / color / mirror / code / Atom feed
From: Maxime Devos <maximedevos@telenet.be>
To: 58198@debbugs.gnu.org
Subject: bug#58198: topological-sort does not sort topologically in case of diamonds
Date: Wed, 12 Oct 2022 14:34:19 +0200	[thread overview]
Message-ID: <77b74931-cf0a-580b-b558-461ec8d5c2ef@telenet.be> (raw)
In-Reply-To: <26f63a49-ad17-865b-8ad3-43bf6e76389b@telenet.be>


[-- Attachment #1.1.1: Type: text/plain, Size: 239 bytes --]



On 08-10-2022 20:13, Maxime Devos wrote:
> I found a solution: [...]

It's buggy, it doesn't handle situations like

	    libnewsboat
	  /   |
	 |  regex-rs
          |    |
         strprintf.

Revised module is attached.

[-- Attachment #1.1.2: topological-sort.scm --]
[-- Type: text/x-scheme, Size: 4339 bytes --]

;;; GNU Guix --- Functional package management for GNU
;;; Copyright © 2019 Ludovic Courtès <ludo@gnu.org>
;;; Copyright © 2022 Maxime Devos <maximedevos@telenet.be>
;;;
;;; This file is part of GNU Guix.
;;;
;;; GNU Guix is free software; you can redistribute it and/or modify it
;;; under the terms of the GNU General Public License as published by
;;; the Free Software Foundation; either version 3 of the License, or (at
;;; your option) any later version.
;;;
;;; GNU Guix is distributed in the hope that it will be useful, but
;;; WITHOUT ANY WARRANTY; without even the implied warranty of
;;; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
;;; GNU General Public License for more details.
;;;
;;; You should have received a copy of the GNU General Public License
;;; along with GNU Guix.  If not, see <http://www.gnu.org/licenses/>.

;; To be used by the implementation of workspaces.
;; Extracted from (guix import utils), and changed from (guix sets)
;; to a guile-pfds equivalent.
(define-module (topological-sort)
  #:export (topological-sort topological-sort*)
  #:use-module (srfi srfi-1)
  #:use-module ((srfi srfi-69) #:select (hash))
  #:use-module ((ice-9 match) #:select (match))
  ;; XXX: Cuirass compiles even build-side only modules.
  #:autoload (pfds hamts) (make-hamt hamt-ref hamt-set))

(define (topological-sort nodes
                          node-dependencies
                          node-name)
  "Perform a breadth-first traversal of the graph rooted at NODES, a list of
nodes, and return the list of nodes sorted in topological order.  Call
NODE-DEPENDENCIES to obtain the dependencies of a node, and NODE-NAME to
obtain a node's uniquely identifying \"key\"."
  ;; It is important to do a breadth-first traversal instead of a depth-first
  ;; traversal -- a simpler depth-first traversal has caused failures in the
  ;; past.
  (define (is-dependency? potential-dependency potential-dependents)
    (member (node-name potential-dependency)
	    (map node-name
		 (append-map node-dependencies potential-dependents))))
  (let loop ((unexpanded-nodes nodes)
	     (result '()) ; in reverse topological order
	     ;; Identical to 'result', except for using a different data
	     ;; structure.
	     (visited (make-hamt hash equal?)))
    (if (null? unexpanded-nodes)
	(reverse result) ; done!
	(let inner-loop ((current-unexpanded-nodes unexpanded-nodes)
			 (later-unexpanded-nodes '())
			 (result result)
			 (visited visited)
			 (progress? #false))
	  (match current-unexpanded-nodes
	    ((first . current-unexpanded-nodes)
	     (cond ((hamt-ref visited (node-name first) #false)
		    ;; Already visisted, nothing to do!
		    (inner-loop current-unexpanded-nodes
				later-unexpanded-nodes result visited
				#true))
		   ;; XXX: would be nice to not recompute
		   ;; 'node-dependencies'.
		   ((is-dependency? first current-unexpanded-nodes)
		    ;; The node was a dependency of something on the previous
		    ;; level, but also of something of the current level.
		    ;; Delay it for later.
		    (inner-loop current-unexpanded-nodes
				(cons first later-unexpanded-nodes)
				result
				visited
				progress?))
		   (#true
		    ;; Expand 'first', putting dependencies in
		    ;; 'later-unexpanded-nodes'.
		    (inner-loop current-unexpanded-nodes
				(append (node-dependencies first)
					later-unexpanded-nodes)
				(cons first result)
				(hamt-set visited (node-name first) #true)
				#true))))
	    (()
	     ;; All nodes on the current level are expanded, descend!
	     ;; But first check for a cycle.
	     (if progress?
		 (loop later-unexpanded-nodes result visited)
		 (error "cycle"))))))))

(define (topological-sort* nodes node-dependencies node-name)
  "Like TOPOLOGICAL-SORT, but don't assume that NODES are roots.  Instead,
consider all nodes in the closure of NODES."
  (define artificial-root (make-symbol "root")) ; uninterned, fresh symbol
  (define nodes* (list artificial-root))
  (define (node-dependencies* node*)
    (if (eq? node* artificial-root)
	nodes
	(node-dependencies node*)))
  (define (node-name* node*)
    (if (eq? node* artificial-root)
	artificial-root
	(node-name node*)))
  (define (proper-node? node*)
    (not (eq? node* artificial-root)))
  (filter proper-node?
	  (topological-sort nodes* node-dependencies* node-name*)))

[-- Attachment #1.1.3: OpenPGP public key --]
[-- Type: application/pgp-keys, Size: 929 bytes --]

[-- Attachment #2: OpenPGP digital signature --]
[-- Type: application/pgp-signature, Size: 236 bytes --]

  reply	other threads:[~2022-10-12 12:39 UTC|newest]

Thread overview: 5+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2022-09-30 18:10 bug#58198: topological-sort does not sort topologically in case of diamonds Maxime Devos
2022-10-05  8:42 ` Maxime Devos
2022-10-08 18:13 ` Maxime Devos
2022-10-12 12:34   ` Maxime Devos [this message]
2022-10-17 19:01 ` Maxime Devos

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

  List information: https://guix.gnu.org/

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=77b74931-cf0a-580b-b558-461ec8d5c2ef@telenet.be \
    --to=maximedevos@telenet.be \
    --cc=58198@debbugs.gnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
Code repositories for project(s) associated with this public inbox

	https://git.savannah.gnu.org/cgit/guix.git

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).