From owner-freebsd-hackers Sat Feb 11 18:20:49 1995 Return-Path: hackers-owner Received: (from root@localhost) by freefall.cdrom.com (8.6.9/8.6.6) id SAA00243 for hackers-outgoing; Sat, 11 Feb 1995 18:20:49 -0800 Received: from vmbb.cts.com (vmbb.cts.com [192.188.72.18]) by freefall.cdrom.com (8.6.9/8.6.6) with SMTP id SAA00237 for ; Sat, 11 Feb 1995 18:20:47 -0800 Received: from io.cts.com by vmbb.cts.com with smtp (Smail3.1.28.1 #9) id m0rdO6Z-0000RDC; Sat, 11 Feb 95 12:07 PST Received: (from root@localhost) by io.cts.com (8.6.9/8.6.9) id LAA18292 for hackers@freebsd.org; Sat, 11 Feb 1995 11:57:41 -0800 From: Morgan Davis Message-Id: <199502111957.LAA18292@io.cts.com> Subject: MSDOS FS panic, resuming make world To: hackers@FreeBSD.org Date: Sat, 11 Feb 1995 11:57:41 -0800 (PST) X-Mailer: ELM [version 2.4 PL23] Content-Type: text Content-Length: 1437 Sender: hackers-owner@FreeBSD.org Precedence: bulk This message has two parts: a bug report and a question. First, the bug report. About half-way through a 'make world', I was working in /mnt upon which was mounted my MS-DOS C: drive. Okay, the make is going on in the background, and I did this: io# cd /mnt/win95 io# mv retail/* bld324 After pressing Enter, everything stops for about three seconds (there was no disk activity that I could see), the screen clears and my bios boot display comes up. The first thing I did was to boot into DOS and check the filesystem. All was fine. I rebooted FreeBSD, and, of course, it cleaned up some minor filesystem glitches. All is well. Now, my question. What's the best way to resume an interrupted 'make world' so that I don't have to start from scratch, and that it completes pretty much as if 'make world' had never stopped? I see no obvious make target for this. I assume I can do this: cd /usr/src make make install cd /usr/share/man make makedb If that's the safest solution, could we add a 'resume' or 'continue' target to the top Makefile? I'm sure this would come in handy. Perhaps some brains could be added to keep some state information on how far 'make world' actually gets so that 'make resume' knows which incomplete level to jump back to in order to continue onward. (This suggestion is made after too many 'make world' attempts that bomb out because someone botched a Makefile further down in the tree.)