#994 Rewrite of the thumbnail system

[Patrick Kellum]

Could we re-write the thumbnail/media support so that we aren't relying on the media filename to find a thumbnail? Currently I can't use thumbnails for text files and pdf's. Perhaps a custom _THUM tag, containing the path to the thumbnail?

[KosherJava]

Logged In: YES
user_id=634811
Originator: NO

To create a thumbnail for a PDF (or any other), create a thumbnail image in any image format (gif, jpg, png etc) and rename it .pdf to match the exact name of the media and add it to the thumbs dir. A browser does not care what file extension is used in in <img> tag, and if it is a valid image it will display. I have been doing this for years. Just remember that you might have an issue loading the .pdf file in many programs, since they expect a pdf and not an image. IMHO this is a lot better than a custom GEDCOM tag.

[Patrick Kellum]

Logged In: YES
user_id=157163
Originator: YES

This seems like a rather dirty hack IMO, but I suppose if it works :|

instead of a custom tag, why not a field in the database?

[Greg Roach]

Logged In: YES
user_id=1466942
Originator: NO

While we're looking at this, I'd like the facility to have completely different media types for thumb/main images.

e.g. flickr.com is happy for you to "deep-link" to the thumbnails, but asks you to link to the html-page of the main image.

So, the thumbnail is flickr.com/...../xxx.jpg
and the main image is flickr.com/...../xxx.html

[Patrick Kellum]

Logged In: YES
user_id=157163
Originator: YES

This goes hack to my idea of a thumbnail filename not being forced to be the same as the source filename. I intend to look at the thumbnails more after I finish up the stats extension.