[gcmd-dev] Reliable basic file operations

[Micha]

I.  F5 is hilarious. Two examples:

(1) F5 asks for different filename even when i copy multiple files,
leading to some funny weirdness if i enter a new one.
F5 in Krusader copies without asking for different filename.
That's at least avoiding any problems, and it's a valid solution too.
(Most time one does not need to change filename, and you always can rename)

(2) copy 'foo' from /home/bar to /tmp (or /tmp/ ), entering a replacement for 
the default, still always copies to the opposite pane, creating 'bar' (not foo).


II.  File list display is flakey

Ctrl+r moves the 'current file' highlight always to the top '..' entry.
And anhilates any marked selection. Which is quite annyoing, if you need to 
locate and track a file in a directory with high network traffic. 


III.  Updating the file lists is wicked

For example, my internet proxy server catches all downloaded files as hardlinkes
in several index directories. But in gcmd, only the initial file creation is 
reflected, 
while the 'filling' of the file with Kilo- or Megabytes is not, independent of 
the file
size. Browser downloads are shown correctly, however. 
It does not depend on fam or gamin installed or not. 
(We should clarify that - gcmd invokes gnomevfs monitors which use libfam 
directly, right ?) Now you can say, special exception, but so what ? krusader
gets it correctly. And it's not clear if this bug depends on (admittedly rare) 
hardlinking 
at all. So what ? Using Ctrl+r more often ? But then II. snaps me.
Essentially i need to use another FM for that kind of work.

I need all these features every day. For example, I use to copy or move or just 
track 
files between different NFS mounts, every day, and i'm also picking or tracking 
stuff 
catched by a local web proxy server (which works more as kind of content 
manager). 
I need accurate monitoring of creation time, and the recent size even of 
unclosed files,
in all these cases.

I know i should file feature requests, or bug reports, but then, i rather like 
to ask,
instead of collecting spurious bug reports, piling one bug fix onto the other, 
shouldn't 
we audit the basic file operations systematically ? - And what resources do we 
have, 
especially if it would be necessary to rewrite part of the code from scratch ? 
But shouldn't basic operations have high priority, for a file manager ? 

(maybe after python, to keep nagging beta testers busy in other places ;)

greets,

niggles the naggler