NAME
cv - a fast gtk+ image viewer loosely modeled after XV
SYNOPSIS
cv
cv directory
cv path...
cv -g <glob expression...>
find .. -print0 | cv -0r
FEATURES
CV is supposed to work similar to the venerable XV image viewer, just
faster. Why faster?
* optimized directory scanning algorithm
The directory scanning in CV uses some tricks that - on most modern
filesystems - makes it possible to detect filetypes faster than
stat()'ing every file. This makes CV suitable for directories with
lots of files (10000+).
This algorithm is quite unprecise - it doesn't make a difference
between files, device nodes, symlinks and the like, and filetype
detection is done using the file extension only.
On the positive side, it is usually many orders of magnitude faster
than traditional scanning techniques (good for directories with
10000 or 100000+ files).
* queuing for all time-consuming background tasks
All tasks, such as unlinking files or generating thumbnails, that
can be done in the background will be done so - no waiting required,
even when changing directories.
* use of asynchronous I/O
CV tries to use asynchronous I/O whereever it makes sense, for
example while scanning directories, waiting for stat data, unlinking
files or generating thumbnails. This usually decreases scanning
times for large directories a bit (especially on RAID devices and
over NFS) and makes CV much more interactive.
* fast image loading
The time span between the user issuing a command and displaying the
new image should be as small as possible. CV uses optimized
(especially for JPEG) loading functions and sacrifices some quality
(e.g no gamma correction, although this might change) to achieve
this speed.
* fast thumbnail creation
Thumbnail creation uses both CPU and Disk-I/O. CV interleaves both,
so on modern CPUs, thumbnailing is usually limited by I/O speed.
Thumbnail creation for JPEGs has been specially optimized and can
even take advantage of multiple CPUs.
* minimum optical clutter
CV has no menus or other user interface elements that take up a lot
of screen space (or are useful for beginning users). The schnauzer
windows can also be somewhat crowded.
The point of an image viewer is viewing images, not a nice GUI. This
is similar to XV's behaviour.
* efficient (and hard to learn) user interface
CV uses key combinations. A lot. If you are an experienced XV user,
you will find most of these keys familiar. If not, CV might be hard
to use at first, but will be an efficient tool later.
* multi-window GUI
CV doesn't force you to use a specific layout, instead it relies on
your window manager, thus enabling you to chose whatever layout that
suits you most.
* i18n'ed filename handling throughout
As long as glib can recognize your filename encoding (either UTF-8
or locale-specific, depending on the setting of G_BROKEN_FILENAMES)
and you have the relevant fonts, CV will display your filenames
correctly.
* extensible through plug-ins
I have weird plug-ins that access remote databases to find a
directory. This is not likely to be of any use to other people.
Likewise, others might have weird requirements I cannot dream of.
* filename clustering
Among the standard plug-ins is a filename clustering plug-in, that
(in case of tens of thousands images in one directory) might be able
to cluster similar names together.
DESCRIPTION
THE IMAGE WINDOW
You can use the following keys in the image window:
q quit the program
< half the image size
> double the image size
, shrink the image by 10%
. enlarge the image by 10%
n reset to normal size
m maximize to screensize
M maximize to screensize, respecting image aspect
ctrl-m toggle maxpect-always mode
ctrl-sift-m toggle using current image size as max image size
u uncrop
r set scaling mode to 'nearest' (fastest)
s set scaling mode to 'bilinear' (default)
shift-s set scaling mode to 'hyper' (slowest)
t rotate clockwise 90°
T rotate counterclockwise°
a apply all rotations loslessly to a jpeg file (using exiftran)
ctrl-shift-t apply current rotation for future image loads
ctrl-v open a new visual schnauzer window for the current dir
ctrl-c clone the current image window
ctrl-e run an editor ($CV_EDITOR or "gimp") on the current image
ctrl-p fire up the print dialog
ctrl-shift-p same as ctrl-p, but automatically selects "ok"
escape cancel a crop action
And when playing movies, these additional keys are active:
left rewind by 10 seconds
right forward by 10 seconds
down rewind by 60 seconds
up forward by 60 seconds
pg_up rewind by 600 seconds
pg_down forward by 600 seconds
o toggle on-screen display
p pause/unpause
escape stop playing
9 turn volume down
0 turn volume up
Any other keys will be sent to the default schnauzer window, which can
be toggled on and off by right-clicking into the image window.
Left-clicking into the image window will let you crop the image (usually
to zoom into large images that CV scales down).
THE VISUAL SCHNAUZER
Any image-loading action in a schnauzer window acts on the
"last-recently-activated" imagewindow, which currently is simply the
last image window that received a keypress.
You can use the following keys in the schnauzer window:
ctrl-space,
space move to and display next image
ctrl-backspace,
backspace move to and display previous image
ctrl-return,
return display selected picture, or enter directory
cursor keys move selection
page-up move one page up
page-down move one page down
home move to first file
end move to last file
ctrl-a select all files
ctrl-shift-a select all files currently displayed in the schnauzer window
ctrl-d delete selected files WITHOUT ASKING AGAIN
ctrl-g force generation of thumbnails for the selected files
ctrl-shift-g remove thumbnails for the selected files
ctrl-s rescan current direcory or files updates/deletes etc.
ctrl-u update selected (or all) icons if neccessary
ctrl-- unselected thumbnailed images
ctrl-+ keep only thumbnailed images, deselect others
^ go to parent directory (caret).
0-9,
a-z find the first filename beginning with this letter
Right-clicking into the schnauzer window displays a pop-up menu with
additional actions.
SELECTION
You can select entries in the Schnauzer in a variety of ways:
Keyboard
Moving the cursor with the keyboard will first deselect all files
and then select the file you moved to.
Clicking
Clicking on an entry will select the one you clicked and deselect
all others.
Shift-Clicking
Shift-clicking will toggle the selection on the entry under the
mouse.
Dragging
Dragging will select all entries between the one selected when
pushing the button and the one selected when releasing the button.
If you move above or below the schnauzer area while drag-selecting,
the schnauzer will move up/down one row twice per second. In
addition, horizontal mouse movement acts as a kind of invisible
horizontal scrollbar.
Hint: double-click works while click-selecting
You can double-click any image while click-selecting to display it
without stopping the selection process. This will act as if you
normally double-clicked the image to display it, and will toggle the
selection twice, resulting in no change.
FILES
When starting, CV runs the .cvrc file in your $HOME directory as if it
were a perl script. in that, you will mostly load plug-ins.
Example:
system "fping -q -t 10 ether"
or require "/fs/cv/cvplugin.pl";
This will load a plug-in, but only if the machine *ether* is reachable
(supposedly the plug-in is networked in some way :).
ENVIRONMENT VARIABLES
CV_EDITOR
The program that gets executed when the user presses "CTRL-e" in the
Schnauzer or image window. The default is "gimp".
CV_AUDIO_PLAYER
Program used to play all sorts of audio (wav, aif, mp3, ogg...),
default "play". Will be called like "$CV_AUDIO_PLAYER -- <path>".
CV_PRINT_DESTINATION
The default (perl-style) destination to use in the print dialog.
CV_TRASHCAN
When set, must point to a directory where all files that are deleted
by the "Delete Physically" (ctrl-d) action are moved to (other
deletion actions still delete!). If unset, files that are deleted
are really being deleted.
SIGNALS
Sending CV a SIGUSR1 signal will cause all image viewers to reload the
currently loaded image. This is useful if you use CV as a viewer for
changing data - just run it in the background with some path and each
time the image changes, send it a SIGUSR1.
SECURITY CONSIDERATIONS
CV uses Pixbuf to load non-JPEG images. Pixbuf is not considered safe
for this purpose, though (from the gtk-2.2 release notes):
"While efforts have been made to make gdk-pixbuf robust against invalid
images, using gdk-pixbuf to load untrusted data is not recommended, due
to the likelyhood that there are additional problems where an invalid
image could cause gdk-pixbuf to crash or worse."
BUGS/TODO
Lots of functionality is missing.
Pixbuf doesn't always honor G_BROKEN_FILENAMES, so accessing files with
names incompatible with utf-8 might fail.
rotate on disk
lots of ui issues
save(?)
preferences
AUTHOR
Marc Lehmann <cv@plan9.de>.