Quote:
Originally Posted by JimmXinu
I'm not seeing failures with those. The convert to jpeg and resize went fine for me for all images in that story.
Do you have other image options sets? Like grayscale_images? You could try no_image_processing.
Another possibility is the site changing behavior after some arbitrary X downloads. I had to set user_agent: to something browser-like to download from that site at all.
Also, I think force_img_self_referer_regexp was added for scribblehub. You'd likely see that behavior if the image site was serving an HTML file instead.
Code:
## If set, and the img tag's URL matches the regular expression, FFF
## will use the img tag's own src URL as the HTTP Referer instead of
## the page URL. This is useful for some image hosting sites that
## watermark or return HTML instead of image data without it.
#force_img_self_referer_regexp:gelbooru.com|photobucket.com
|
I do not have grayscale_images and I didn't have no_image_processing. All I had was include_images:true and keep_summary_html:true at the [epub] level.
I didn't notice it originally, but after reading this I noticed that the first time I opened an image in a new tab it redirected me to a web page with the image instead of directly to the image. The second time I opened it, it would directly open the image.
However, none of the settings above seemed to work for me, although no_image_processing resulted in it generating a .jpg file for each image that it says is "Not a valid image".
When I just tried looking at the images again, it's always giving me the web page instead of just the image, so I probably triggered something by trying to many times. I'll do more testing later and spread it out more to see if your suggestions do fix it.
Out of curiosity, what do you use for your user_agent?
Also, would you mind sharing your actual force_img_self_referer_regexp setting? If I use one that works for you, it takes one more variable out of my equation.
- Andrew