Skip to contentSkip to conversation

2.6.14-mm2: no .config.old any more?

PATCH Christian Kujau 2 participants 3 messages Nov 14, 2005 – Nov 15, 2005 Kconfig

lore
Share & tools
lore
#0 lore raw

hi,

i noticed that 2.6.14-mm2 does not generate a .config.old anymore, so that i can undo changes. i see that the Kconfig system is probably in flux again ("Why did oldconfig's behavior change in 2.6.15-rc1?"), but i have not seen this issue being reported:

% cp .config .config.really-old
% make menuconfig
[...]
scripts/kconfig/mconf arch/x86_64/Kconfig
#

# using defaults found in .config
#

*** End of Linux kernel configuration.
*** Execute 'make' to build the kernel or try 'make help'.

% diff .config.really-old .config
4c4
< # Tue Nov 15 00:23:53 2005
---

Quoted 1 line

# Tue Nov 15 00:24:41 2005

1454c1454
< CONFIG_PRINTK_TIME=y
---

Quoted 1 line

# CONFIG_PRINTK_TIME is not set

% ls .config.old
ls: .config.old: No such file or directory

attached patch reverts one change introduced in 2.6.14-mm1 and fixes it for me, but i doubt that it is the right thing to do....

thanks,
Christian.

Signature · 3 lines
BOFH excuse #404:

Sysadmin accidentally destroyed pager with a large hammer.
scripts/kconfig/confdata.c renamed from scripts/kconfig/confdata.c.2.6.14-mm2 +4 -14
@@ -518,23 +518,13 @@ int conf_write(const char *name)
 		if (!name)
 			name = conf_def_filename;
 		sprintf(tmpname, "%s.old", name);
-//		printf("rename1(%s, %s)\n", name, tmpname);
-//		rename(name, tmpname);
+		rename(name, tmpname);
 	}
 	sprintf(tmpname, "%s%s", dirname, basename);
-//	printf("rename2(%s, %s)\n", newname, tmpname);
-#if 0
 	if (rename(newname, tmpname))
 		return 1;
-#else
-	{
-		char buf[256];
-		sprintf(buf, "cp %s %s", newname, tmpname);
-		system(buf);
-		unlink(newname);
-	}
-#endif
- 	sym_change_count = 0;
- 
+
+	sym_change_count = 0;
+
 	return 0;
 }
#1 lore raw
Quoted 8 lines · Christian Kujau

hi,

i noticed that 2.6.14-mm2 does not generate a .config.old anymore, so that i can undo changes. i see that the Kconfig system is probably in flux again ("Why did oldconfig's behavior change in 2.6.15-rc1?"), but i have not seen this issue being reported:

Yeah, sorry, the diff you noticed is a horrid seven-second-hack I've carried in my tree since the new Kconfig stuff went in (2.5.early) because I want my .config to be a symlink to a revision-controlled file and the Kconfig system (brokely) insists on blowing away the symlink each time you run it.

A proper patch (which maybe does an lstat+special-stuff) would be nice.

But that particular diff only appears in the -mm rollup when I'm carrying other patches against confdata.c, which rarely happens. So it'll go away again.

#2 lore raw
Quoted 2 lines · Andrew Morton

A proper patch (which maybe does an lstat+special-stuff) would be nice.

hm, sorry, but i don't think i can help out here.

Quoted 3 lines

But that particular diff only appears in the -mm rollup when I'm carrying other patches against confdata.c, which rarely happens. So it'll go away again.

thank you.

Christian.

Signature · 3 lines
BOFH excuse #74:

You're out of memory