I've noticed a few quirks, don't know how easy/hard it would be to fix.
When adding a new movie/editing an old one that doesn't have a
cover-url (I have a couple of greek movies that don't have images in
imdb and have very generic names, one is called "e-mail", so I can't
get a url for an image of the movie. I barely found something remotely
relevant, and I know that the link won't stay there for long so I
downloaded the image on my pc). The 'upload' box however doesn't work.
I.e. it doesn't even copy the image to the cache folder (i checked).
Also regarding the cache folder, I'm assuming on linux machines it
needs to have world read/write privileges? A website I'm running that
used nucleus had a similar directory for users to upload images
through the web, and we got hacked. Someone (I'm assuming located on
the same server as us) managed to copy a php script using a web client
to that directory that gave him a directory listing and the ability to
execute commands (always using 'world' privileges). The security on
the server was pretty secure so whoever did it couldn't do much, but
all files that were "world" readable the script could access and read,
that included all php code the website was using. My solution (for an
apache server, always) was adding an .htaccess file in all the
world-write enabled directories first disabling indexing in all
directories, and then adding the line:
which basically disassociates any files with the .php extension from
execution. I'm assuming you could be extra careful and also add the
html and htm extensions to that, and just have it ge load a text file
if someone browses to that directory saying that "there's nothing to
see here" or something like that...
Also, adding a new movie from Amazon XML (or "overwriting" the data of
an old entry with an Amazon XML link) results in a Database error.
There's a "WHERE id = [Amazon XML id]" clause at the end of the query,
which shouldn't be there. On the "overwrite" it should be the id of
the item. (problem doesn't occur when "add missing" is selected).
Finally, are the "Filename/Filesize/codecs/dimension" fields supposed
to be automatically filled when entering a filename and "add missing"?
Because for me they're not, and there's no way in hell i'm gonna be
looking up all that info for all of my movies ;)
In FireFox (I'm assuming all gecko based browsers) the PDF Export and
RSS 2.0 images are not one above the other, they're next to each
other. As a result, if the user name is long (like mine,
'constantinos'), the text runs over the images.
That's it for the bugs, now to the suggestions:
Maybe add *.php in the .cvsignore file for cache/smarty/ ? It's
annoying to see all those compiled templates appear as unknown files
while doing a cvs update. (files are stored as .tpl.php, is that a
bug, or is that how they're supposed to be stored?)
How hard would it be to add another custom field? As an extension to
that, how hard would it be to add a _customizable_ number of custom
fields? If I'm guessing right (haven't looked at the code or database
structure for this) the custom fields are hard-coded in the main
table? Could everything be moved in a separate 4-column table (index,
itemid, customid, value) and just have as many fields as necessary? I
might be able to take a look into implementing this after I'm done
with this big project at work (sometime early next week) if its
something that would be considered useful...
As a side note, I have the updated greek language file, and I asked
some international friends from college to translate the few words
that were missing for polish, french, dutch bulgarian and serbian..
should be getting them by next week :)