[ale] inode change after vi a file ?

[Dow_Hurst]

This thread is an excellent example of why, not only that I appreciate this list but, the complexity of all the programs we use requires that no assumptions are made about how a newer version might work in the same way as an older version.  I use IRIX with the standard vi and Linux with vim.  The differences are pretty significant now and unified shared home directories don't help at all.  Thanks for the question and answer as I think I was burned by this not long ago but didn't even realize why or what I could've looked for along these lines.  But, it makes perfect sense why logging stopped and I had to restart the daemon!  This wasn't a serious production environment or daemon that users were depending on, so wasn't a big deal.

I don't understand though why people are saying that cycling inodes is a good thing?  I thought one inode per file so it was the number of total files and directories that really mattered.  If you had too many then you had problems.  Is there some reason to reassign inodes?  Thanks,
Dow
 

-----Original Message-----
>From: Joe Steele <joe at madewell.com>
>Sent: Mar 30, 2006 3:40 PM
>To: Atlanta Linux Enthusiasts <ale at ale.org>
>Subject: Re: [ale] inode change after vi a file ?
>
>Jerry Yu wrote:
>> without agressive inode recycling, one'd run out of inodes at some 
>> point too ...
>> it may be a feature for VIM, can it be turned off  by some magic 
>> switch?  I see it as a bug personally. wonder whether/how oowriter and 
>> nano do reliable atomic writes. also wonder how vi on Solaris does it.
>>
>This feature is controlled by the 'backupcopy' or 'bkc' setting.  I 
>suspect your present setting is "auto", so vim tries to rename the 
>original file to a backup name, write the data to a new file, then 
>unlink the original file.  What you are asking for is obtained with the 
>"yes" setting:
>
>    :set bkc=yes
>
>For more details, see ":help backupcopy" in vim. 
>
>--Joe
>_______________________________________________
>Ale mailing list
>Ale at ale.org
>http://www.ale.org/mailman/listinfo/ale


No sig.