|
urlshot is a Perl hack. It reads in a list of
filename-URL pairs, loads the URLs up in a web browser,
takes screenshots of the browser, post-processes them and
saves them to the corresponding filenames.
To do this, like any self-respecting Perl hack, it uses a
lot of other facilities:
Xvfb, the Virtual Framebuffer X Server hosts the
display and web browser without need for interactivity or
framebuffer hardware (any running X server can be contacted
by urlshot, however).
Although any web browser can be used by, this version of
urlshot was designed with skipstone in mind: a
small, gecko-based browser with a minimal but
workable feature set.
All the image processing is done with PerlMagick
(Image::Magick), which obviously depends on the ImageMagick
package. As a side effect, urlshot can save files in
virtually any raster format you could think of.
The base X clients are also needed, as well as
lsof and the GET script from libwww-perl.
urlshot reads its input from the specified
FILE, or from the standard input if no files are
given on the command line, or if the -
‘file’ is specified.
In reading from standard input, if standard input is a
teletype (e.g. if urlshot is run with no file
arguments), the program enters interactive mode, where it
presents brief instructions and a prompt before each
filename-URL pair.
The download process first checks to see if the page
exists and what content type it returns. If the page exists,
the browser is instructed to load it, and urlshot
waits for this to complete.
Once the browser has loaded the URL, a screenshot is
generated using xwd(1), cropped, and if required,
sharpened and resized.
|
