Article View: gnu.epoch.misc
Article #74Re: Painful epoch-4.0p0 bugs
From: Kent Williams
Date: Wed, 08 Apr 1992 03:04
Date: Wed, 08 Apr 1992 03:04
45 lines
2517 bytes
2517 bytes
> The first two bugs are easy to duplicate. When running as usual >in minibuffer-per-window mode (with -nm flag) with at least two epoch >screens, do the following: > > - begin a command such as find-file-other-screen, leaving the prompt > in the minibuffer. > > - Move the mouse to another screen, and the minibuffer prompt will > follow it into this screen. > > - Move the mouse again back to the original screen, and click to put > the cursor into it > (the prompt doesn't follow automatically, which is a small bug in itself) > > - Use C-x o to return to the minibuffer > > - Complete the find-file-other-screen command > > The find-file-other-screen works fine, but the focusing is now >stuffed; if you move the mouse to another screen, the cursor simply >moves into the minibuffer of the previous screen. It is possible to >click in screens to move the cursor to them, but the only way to >return to a reasonable state of affairs is by killing epoch and >starting again. This combination of circumstances may sound extremely >rare, but it happens to me about once a day, even when I'm trying to >remember not to cause it. > > A still nastier bug lies in store if you have lots of windows open >and decide to soldier on; eventually in unpredictable circumstances >(quitting a file and closing a screen is one example), epoch will >start to ring the bell continuously, making it near impossible to kill >on the Decstation because X comes to almost a complete halt. > I have seen this same behavior, on both an SGI and a SUN; so I would guess this isn't a platform-specific bug. Hats off to Peter for providing such a concise description; I never had the patience to actually figure out what I had done to make this happen, and then to write up a bug report. And yes the bell ringing stuff is pretty hideous. Our worstations are in a common room, so when epoch loses it's marbles, everyone knows about it. The best thing to do is to login from somewhere else and kill epoch; but there might still be 32K bells left in the buffer waiting to be echoed.
Technical Details
Message-ID:
<9204081304.AA38836@herky.cs.uiowa.edu>Path:
arc1-ca.pugleaf.net!sparky!uunet!cis.ohio-state.edu!herky.cs.uiowa.edu!williamsShow Full Headers
Complete Article Headers
Newsgroups: gnu.epoch.misc Path: arc1-ca.pugleaf.net!sparky!uunet!cis.ohio-state.edu!herky.cs.uiowa.edu!williams From: williams@herky.cs.uiowa.edu (Kent Williams) Subject: Re: Painful epoch-4.0p0 bugs Message-ID: <9204081304.AA38836@herky.cs.uiowa.edu> Sender: daemon@cis.ohio-state.edu Organization: GNUs Not Usenet Distribution: gnu Date: Wed, 8 Apr 1992 03:04:11 GMT Lines: 47