On Fri, Aug 27, 2010 at 10:29:16PM +0200, yaffs@piments.com wrote:
>On 08/27/10 19:16, Chuck Coffing wrote:
>>Do not explicitly create a root directory in the image, to match the actual
>>behavior of the filesystem.
>>
>>The yaffs filesystem creates root and several other fake directories in memory
>>(see yaffs_CreateInitialDirectories); they are neither created nor necessary in
>>the NAND.
>>Index: mkyaffs2image/mkyaffs2image.c
>>===================================================================
>>--- mkyaffs2image.orig/mkyaffs2image.c	2010-08-26 17:15:39.000000000 -0600
>>+++ mkyaffs2image/mkyaffs2image.c	2010-08-26 17:16:07.000000000 -0600
>>@@ -569,8 +569,7 @@
>>  	}
>>  	
>>  	printf("Processing directory %s into image file %s\n",argv[1],argv[2]);
>>-	if(write_object_header(1, YAFFS_OBJECT_TYPE_DIRECTORY,&stats, 1,"", -1, NULL) == 0)
>>-		process_directory(YAFFS_OBJECTID_ROOT,argv[1]);
>>+	process_directory(YAFFS_OBJECTID_ROOT,argv[1]);
>>  	
>>  	pad_image();
>>
>>
>>_________
>
>4/4 stands for "fourth out of four" but it's in fact the 5/4 !! While
>true blood hackers always count from zero it does not make much sense
>when listing patches. How will someone looking back at patch 4/4
>realise it is not the last one in the set?

Well #0 of 4 was the intro, not a real patch, so... really are just 4
patches.

>thanks for the effort, those are significant short-comings you've picked up.

no problem.  :-)

--
Chuck