From mboxrd@z Thu Jan 1 00:00:00 1970 Path: news.gmane.io!.POSTED.ciao.gmane.io!not-for-mail From: Eli Zaretskii Newsgroups: gmane.emacs.bugs Subject: bug#41239: GTK builds crashing in XTread_socket after deleting a frame Date: Thu, 14 May 2020 17:22:32 +0300 Message-ID: <837dxe61g7.fsf@gnu.org> References: <83v9kz679v.fsf@gnu.org> <0af1b68c-93d3-2e8b-7810-41d60ef6a2a8@gmx.at> Injection-Info: ciao.gmane.io; posting-host="ciao.gmane.io:159.69.161.202"; logging-data="108866"; mail-complaints-to="usenet@ciao.gmane.io" Cc: 41239@debbugs.gnu.org To: martin rudalics Original-X-From: bug-gnu-emacs-bounces+geb-bug-gnu-emacs=m.gmane-mx.org@gnu.org Thu May 14 16:48:31 2020 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 1jZF9w-000S9E-9J for geb-bug-gnu-emacs@m.gmane-mx.org; Thu, 14 May 2020 16:48:28 +0200 Original-Received: from localhost ([::1]:35412 helo=lists1p.gnu.org) by lists.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1jZF9v-0001Au-A1 for geb-bug-gnu-emacs@m.gmane-mx.org; Thu, 14 May 2020 10:48:27 -0400 Original-Received: from eggs.gnu.org ([2001:470:142:3::10]:39848) by lists.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_256_GCM_SHA384:256) (Exim 4.90_1) (envelope-from ) id 1jZElK-0004z4-9u for bug-gnu-emacs@gnu.org; Thu, 14 May 2020 10:23:02 -0400 Original-Received: from debbugs.gnu.org ([209.51.188.43]:51011) by eggs.gnu.org with esmtps (TLS1.2:ECDHE_RSA_AES_128_GCM_SHA256:128) (Exim 4.90_1) (envelope-from ) id 1jZElJ-0006C5-V1 for bug-gnu-emacs@gnu.org; Thu, 14 May 2020 10:23:01 -0400 Original-Received: from Debian-debbugs by debbugs.gnu.org with local (Exim 4.84_2) (envelope-from ) id 1jZElJ-0005EU-RZ for bug-gnu-emacs@gnu.org; Thu, 14 May 2020 10:23:01 -0400 X-Loop: help-debbugs@gnu.org Resent-From: Eli Zaretskii Original-Sender: "Debbugs-submit" Resent-CC: bug-gnu-emacs@gnu.org Resent-Date: Thu, 14 May 2020 14:23:01 +0000 Resent-Message-ID: Resent-Sender: help-debbugs@gnu.org X-GNU-PR-Message: followup 41239 X-GNU-PR-Package: emacs Original-Received: via spool by 41239-submit@debbugs.gnu.org id=B41239.158946617520085 (code B ref 41239); Thu, 14 May 2020 14:23:01 +0000 Original-Received: (at 41239) by debbugs.gnu.org; 14 May 2020 14:22:55 +0000 Original-Received: from localhost ([127.0.0.1]:34324 helo=debbugs.gnu.org) by debbugs.gnu.org with esmtp (Exim 4.84_2) (envelope-from ) id 1jZElD-0005Ds-7X for submit@debbugs.gnu.org; Thu, 14 May 2020 10:22:55 -0400 Original-Received: from eggs.gnu.org ([209.51.188.92]:36810) by debbugs.gnu.org with esmtp (Exim 4.84_2) (envelope-from ) id 1jZElB-0005Dg-IH for 41239@debbugs.gnu.org; Thu, 14 May 2020 10:22:54 -0400 Original-Received: from fencepost.gnu.org ([2001:470:142:3::e]:33862) by eggs.gnu.org with esmtp (Exim 4.90_1) (envelope-from ) id 1jZEl6-00063J-3j; Thu, 14 May 2020 10:22:48 -0400 Original-Received: from [176.228.60.248] (port=4131 helo=home-c4e4a596f7) by fencepost.gnu.org with esmtpsa (TLS1.2:RSA_AES_256_CBC_SHA1:256) (Exim 4.82) (envelope-from ) id 1jZEl4-0002cJ-8I; Thu, 14 May 2020 10:22:46 -0400 In-Reply-To: <0af1b68c-93d3-2e8b-7810-41d60ef6a2a8@gmx.at> (message from martin rudalics on Thu, 14 May 2020 09:54:02 +0200) 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:180210 Archived-At: > Cc: 41239@debbugs.gnu.org > From: martin rudalics > Date: Thu, 14 May 2020 09:54:02 +0200 > > >> This looks like yet another GTK specific crash where we apparently can't > >> do much. But maybe someone can spot a pattern we could employ to avoid > >> the XTread_socket calls when deleting a frame. > > > > Isn't block_input inside Fdelete_frame what you want? > > Do you mean the already existing one in delete_frame or adding a new > one? The existing ones cover only small parts of delete_frame, and the safe_call2 call is outside both of them, AFAICT. So yes, I mean blocking input around the safe_call2 call. > > However, the first backtrace doesn't show Fdelete_frame in the > > backtrace, so I'm not sure what's going on there. > > Neither do I. Note some of the prerequisites: > > - Tooltips must be GTK+ native ones - no crash with our tooltips. > > - There must be two frames - no crash with just one frame. > > - Deleting the frame must be via Alt-F4 - no crash with C-x 5 0. What about the frame whose menu bar is being updated -- is that the frame we deleted, by any chance? Is it a live frame? The crashes are all in memory allocation routines, so another possibility is that we have some memory corruption.