Hi, On Mon, Oct 08, 2007 at 03:18:10PM +0000, place wrote:
BTW, are you still opposed to the idea of click-to-load-images? I'm on dialup, so one of the first things I had to do was port my old hacks from my gtk version.
I am/was not opposed. I thought a good solution was to have a small tools icon (near bug meter) where you could toggle images on/off, my_colors on/off, wide_rendering on/off etc. Now, being in no-images mode and being able to request a single one instead of all at once looks like a good idea. Of course you'd also have the option of turning images on, and have all requested for that page. As most of the images don't specify a size, a place holder must be provided (as the broken image link in other browser). From there, a right click menu could offer to download that image only.
Also, do you have any thoughts on the proper solution for pages using character sets that aren't utf-8 or a subset thereof? I made something that reads the charset from Content-Type in the header or meta tags and then, in Html->write, runs the buffers through iconv. (Pages are mostly readable, but I think I'm either losing partial characters at buffer boundaries or losing partial tags between Html->write_raw() calls or something...)
Basically that. It's not simple though. For instance an http page may specify the charset in Content-Type, but when saved not give any clue in its meta tags. Or have only meta but not Content-Type as served, or give no clue at all, etc. The main idea is to keep the page as is in the cache (so it can be saved verbatim) and to pass it trough a filter to make the conversion to utf-8 inside Dillo. If the iconv backpack adds too much to Dillo's size, the task can be delegated to a dpi, and a small subset with the main conversions be kept inside Dillo. -- Cheers Jorge.-