From mboxrd@z Thu Jan 1 00:00:00 1970 Path: news.gmane.io!.POSTED.blaine.gmane.org!not-for-mail From: Stefan Monnier via "Bug reports for GNU Emacs, the Swiss army knife of text editors" Newsgroups: gmane.emacs.bugs Subject: bug#58361: 29.0.50; noverlay branch is O(N) for important calls Date: Fri, 07 Oct 2022 14:38:31 -0400 Message-ID: References: <878rlrfyje.fsf@rfc20.org> Reply-To: Stefan Monnier Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: quoted-printable Injection-Info: ciao.gmane.io; posting-host="blaine.gmane.org:116.202.254.214"; logging-data="24647"; mail-complaints-to="usenet@ciao.gmane.io" User-Agent: Gnus/5.13 (Gnus v5.13) Emacs/29.0.50 (gnu/linux) Cc: gerd.moellmann@gmail.com, matt@rfc20.org, 58361@debbugs.gnu.org, eliz@gnu.org To: Andreas Politz Original-X-From: bug-gnu-emacs-bounces+geb-bug-gnu-emacs=m.gmane-mx.org@gnu.org Fri Oct 07 20:40:26 2022 Return-path: Envelope-to: geb-bug-gnu-emacs@m.gmane-mx.org Original-Received: from lists.gnu.org ([209.51.188.17]) by ciao.gmane.io with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.92) (envelope-from ) id 1ogsGn-0006CI-VS for geb-bug-gnu-emacs@m.gmane-mx.org; Fri, 07 Oct 2022 20:40:26 +0200 Original-Received: from localhost ([::1]:35906 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1ogsGm-0000SN-Gl for geb-bug-gnu-emacs@m.gmane-mx.org; Fri, 07 Oct 2022 14:40:24 -0400 Original-Received: from eggs.gnu.org ([2001:470:142:3::10]:56240) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1ogsGR-0000R7-M5 for bug-gnu-emacs@gnu.org; Fri, 07 Oct 2022 14:40:04 -0400 Original-Received: from debbugs.gnu.org ([209.51.188.43]:38280) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_128_GCM_SHA256:128) (Exim 4.90_1) (envelope-from ) id 1ogsGQ-0001yj-Nb for bug-gnu-emacs@gnu.org; Fri, 07 Oct 2022 14:40:03 -0400 Original-Received: from Debian-debbugs by debbugs.gnu.org with local (Exim 4.84_2) (envelope-from ) id 1ogsGQ-0002KK-8E for bug-gnu-emacs@gnu.org; Fri, 07 Oct 2022 14:40:02 -0400 X-Loop: help-debbugs@gnu.org Resent-From: Stefan Monnier Original-Sender: "Debbugs-submit" Resent-CC: bug-gnu-emacs@gnu.org Resent-Date: Fri, 07 Oct 2022 18:40:02 +0000 Resent-Message-ID: Resent-Sender: help-debbugs@gnu.org X-GNU-PR-Message: followup 58361 X-GNU-PR-Package: emacs X-Debbugs-Original-Cc: Gerd =?UTF-8?Q?M=C3=B6llmann?= , Matt Armstrong , bug-gnu-emacs@gnu.org, Eli Zaretskii Original-Received: via spool by submit@debbugs.gnu.org id=B.16651679568867 (code B ref -1); Fri, 07 Oct 2022 18:40:02 +0000 Original-Received: (at submit) by debbugs.gnu.org; 7 Oct 2022 18:39:16 +0000 Original-Received: from localhost ([127.0.0.1]:37358 helo=debbugs.gnu.org) by debbugs.gnu.org with esmtp (Exim 4.84_2) (envelope-from ) id 1ogsFg-0002Iv-9S for submit@debbugs.gnu.org; Fri, 07 Oct 2022 14:39:16 -0400 Original-Received: from lists.gnu.org ([209.51.188.17]:39630) by debbugs.gnu.org with esmtp (Exim 4.84_2) (envelope-from ) id 1ogsFd-0002Il-Sr for submit@debbugs.gnu.org; Fri, 07 Oct 2022 14:39:15 -0400 Original-Received: from eggs.gnu.org ([2001:470:142:3::10]:51884) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1ogsFZ-000086-7B for bug-gnu-emacs@gnu.org; Fri, 07 Oct 2022 14:39:12 -0400 Original-Received: from mailscanner.iro.umontreal.ca ([132.204.25.50]:9898) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1ogsF4-0001su-D8; Fri, 07 Oct 2022 14:38:59 -0400 Original-Received: from pmg2.iro.umontreal.ca (localhost.localdomain [127.0.0.1]) by pmg2.iro.umontreal.ca (Proxmox) with ESMTP id 5F1E18004C; Fri, 7 Oct 2022 14:38:35 -0400 (EDT) Original-Received: from mail01.iro.umontreal.ca (unknown [172.31.2.1]) by pmg2.iro.umontreal.ca (Proxmox) with ESMTP id 50FA380401; Fri, 7 Oct 2022 14:38:33 -0400 (EDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=iro.umontreal.ca; s=mail; t=1665167913; bh=TfRJTRm0zM57+oiReAoWVUmuMZkOqz0SJtRY/HWhT6Q=; h=From:To:Cc:Subject:In-Reply-To:References:Date:From; b=d9Iht9JQBzpoMQq3sIPI9glDmwxLa0fStR7KLPd9WiNSo/s3UgRNnE27jAG5RMdhQ m3R+Tk1TRSvkZYrMot/4GOi9nnOH8skkrnpokGe0HqiRlMjmfs7ZiHzwQbFdYhktrR Ouc7M2dZK5pFgTTdRPKTpDlrXUb0uafv+YK3raaPUTm1cRB8aA7Lp/qRg+e3ui3dTI dN3/Ke5cNtYNcyK98tjU91TOjniFp8vgZARwovDqewBZrUYHO62sELDShX5avzReIZ /JkIaNv7OWrOlWWX4L+zh8yHdsOOkL/QNSVZwEdg3nhq7QGqML+WlFSyFNtdYhRMZN u3Q0ojOb+Q2yw== Original-Received: from pastel (65-110-220-202.cpe.pppoe.ca [65.110.220.202]) by mail01.iro.umontreal.ca (Postfix) with ESMTPSA id D1BA212047D; Fri, 7 Oct 2022 14:38:32 -0400 (EDT) In-Reply-To: (Andreas Politz's message of "Fri, 7 Oct 2022 18:51:14 +0200") Received-SPF: pass client-ip=132.204.25.50; envelope-from=monnier@iro.umontreal.ca; helo=mailscanner.iro.umontreal.ca X-Spam_score_int: -42 X-Spam_score: -4.3 X-Spam_bar: ---- X-Spam_report: (-4.3 / 5.0 requ) BAYES_00=-1.9, DKIM_SIGNED=0.1, DKIM_VALID=-0.1, DKIM_VALID_AU=-0.1, RCVD_IN_DNSWL_MED=-2.3, SPF_HELO_NONE=0.001, SPF_PASS=-0.001 autolearn=ham autolearn_force=no X-Spam_action: no action X-BeenThere: debbugs-submit@debbugs.gnu.org X-Mailman-Version: 2.1.18 Precedence: list X-BeenThere: bug-gnu-emacs@gnu.org List-Id: "Bug reports for GNU Emacs, the Swiss army knife of text editors" List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , Errors-To: bug-gnu-emacs-bounces+geb-bug-gnu-emacs=m.gmane-mx.org@gnu.org Original-Sender: "bug-gnu-emacs" Xref: news.gmane.io gmane.emacs.bugs:244838 Archived-At: > I think, a straightforward way to use 2 trees, one for begin and one for > end, could be to create another abstraction above those trees, while for = the > most part duplicating the existing interface. This abstraction would t= hen > either delegate to one or both trees, depending on the operation. The tri= ck > would be to kinda multiplying the end-tree by -1, i.e. reverse begin and > end and multiply with -1 all inputs and outputs of this tree. > > Would that work ? Could be but I'm not sure we want to pay this memory&cpu price to try and fix a performance bug that's still hypothetical. For all we know, in those cases where this performance problem could bite, other performance problems bite us harder anyway. Stefan >> Am 07.10.2022 um 17:23 schrieb Matt Armstrong : >>=20 >> =EF=BB=BFTo start, I don't think this issue should delay a merge to mast= er. I >> don't think it is clear we need to fix anything here. >>=20 >> I would like a note or FIXME in code noting the potentially slow >> algorithm (patch sent), because it is currently well hidden behind a >> generator loop. >>=20 >>=20 >> Stefan Monnier writes: >>=20 >>>> Here we traverse overlays in ASCENDING order of BEG positions. The be= st >>>> we can say is that this loop executes in O(K*log(N)) time, where K is >>>> the MIN of number of overlays that overlap POS and the number of valid >>>=20 >>> The core operation in itree.c is the equivalent of `overlays-in/at`. >>=20 >> [...] >>=20 >> Yes, and for this O(K*log(N)) performance is a good result. The key >> insight is that previous and next overlay changes require examining a >> large K (in worst case, extending all the way to the beginning or end of >> the buffer) because there is no ordering by END positions. >>=20 >>> Realistic benchmarks would be most welcome. >>=20 >> I am working on polishing off >> https://git.sr.ht/~matta/emacs-overlay-perftests. Good news is that >> redisplay is faster on the noverlay branch for the "realistic" case of >> overlaping not overlapping eachother in pathalogical ways. >>=20 >>=20 >>> [ Site note: `previous-overlay-change` is probably not very important in >>> practice, but `next-overlay-change` OTOH is indeed important because >>> it's used during redisplay. So if someone comes up with a trick to >>> speed up only one direction, it should be good enough. ] >>>=20 >>> Maybe one way to improve the behavior is to accept the worst-case >>> bound but to try and avoid paying it over-and-over each time the >>> redisplay needs the "next change". IOW instead of a >>> `next_overlay_change` function which takes a POS and returns the next >>> change after that, the xdisp.c might benefit from having a >>> `next_overlay_changes` *generator* which takes a starting POS and >>> returns an iterator which will return (each time it's called) the >>> successive positions where there's an overlay change. >>>=20 >>> Hopefully this way we'd pay the O(N) cost once per redisplayed window >>> rather than once per "small step in the rendering engine" (i.e. per >>> next_overlay_change). >>=20 >> At the moment I can't think of a reasonable way to implement such a >> generator efficiently without, effectively, computing a temporary >> ordered collection over overlay END positions. >>=20 >> This is why I keep coming back to the idea of storing both BEG and END >> positions in ordered collections at all times. >>=20 >>=20 >>> Another way to do basically the same is to let next_overlay_change >>> fill up a cache of change-positions which would be flushed whenever >>> some overlay is modified/added/removed (or the current_buffer is >>> different from last time). That might be easier to use with the >>> current code since xdisp.c wouldn't need to pass around this iterator >>> (which could require significant reworks). >>=20 >> ...possibly, but the problem with caching is the time spent filling the >> cache back up. I like the idea of storing both BEG and END positions in >> an ordered collection because in that case the (potentially slow) >> recomputation need not occur with every key press. If we're not worried >> about that kind per-key-press of delay, then I argue there is no need >> for a cache either.