Re: [gentoo-user] Dolphin confusing different run instances.

[Frank Steinmetzger <Warp_7@gmx.de>]

[-- Attachment #1: Type: text/plain, Size: 3370 bytes --]

Am Fri, Sep 13, 2024 at 01:19:00AM -0500 schrieb Dale:

> >> P. S.  Planning to try that checksum script soon.  It's a large number
> >> of files so it will take a long time to run.  I think you mentioned that
> >> if stopped, it will resume where it left off.
> > Only if it creates checksums, because it knows by the existence of checksums 
> > where to resume. But if you want to read checksums and verify them, you need 
> > to use arguments to tell it how many directories to process and how many to 
> > skip at the beginning.
> >
> > Perhaps try it first with a few small directories to get a feel for its 
> > behaviour. The normal way to go is:
> >
> > dh -u [DIR] to create the checksum files
> > dh [DIR] do read it back
> > Use the --skip option to skip the given number of dirs at the beginning.
> >
> > Remember that by default it will not create checksums in directories that 
> > have subdirectories. I know this sounds a little strange, but for a 
> > hierarchy of music albums, this seemed sensible 10 years ago.
> >
> 
> Well, I read through the help page and settled on this.  I might have
> did this wrong.  ;-)
> 
> /root/dh -c -F 1Checksums.md5 -v
> 
> Right now I have the command in /root.  I just did a cd to the parent
> directory I wanted it to work on and then ran that command.  Right now,
> it is working on this bit.
> 
> 
> (dir 141 of 631)
> 
> and
> 
> (file  8079 of 34061)

I am thinking about adding filesize information, but that would require 
updating the status line during the processing of a file instead of only 
between files. That’s not trivial, as it involves timers and threads.

> I was wondering tho, is there a way to make it put all the checksum
> files in one place, like a directory call checksums, and they just all
> go in there?

Hm … from an algorithmic point of view, it would actually not be that 
complicated by creating a shortened filename from the source directory, but 
the real-world use seems a bit far-fetched. Checksums should be close to 
their data. If you have read errors for either, then the other is useless 
anyways. :D

> Or just a single file in the parent directory?  That way
> the files aren't in each directory.

That’s what the -s option is for. This will create only one checksum file at 
the root level for each directory argument. So if you run `dh -us foo/ bar/`, 
then it will go into foo/, create one checksum file there and put all lines 
into that one file, even for subdirectories, and do the same in bar/.

However, at the moment automatically detecting and properly verifying those 
files is still in the works. So I think you have to use the -s option or -F 
all to read them.

> Thing is, can I still just run it
> on one directory if I have a suspected bad one?

Not with one checksum file at the root level for an entire tree. The way I 
would handle this case: run dh -u on the directory of interest and then 
compare the checksums in the root-level file and the newly created file with 
a diff tool. Or copy the lines from the existing checksum file, create a new 
file in the directory of interest, remove the directory part of the paths 
and then run dh on just that directory.

-- 
Grüße | Greetings | Salut | Qapla’
I don’t have a problem with alcohol, just without!

[-- Attachment #2: signature.asc --]
[-- Type: application/pgp-signature, Size: 833 bytes --]