New issue
Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.
By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.
Already on GitHub? Sign in to your account
Html link parser #316
Html link parser #316
Conversation
…f is an image Simple bug - if(image) was always true. Needed if(image.length)
Wow, this looks fantastic, People will be blown away when they can get a visual indication of what they're clicking on sites without even opening a browser. And all the text manipulation power as well. Can I tweet this page? |
This is so much better. It fixes most of the things I’ve always hated and reminds me of what’s left to fix.
IMO, the image URL should be a fallback for the name. It should prefer things in this order
The good news is that Quicksilver will already do this as is, we just have to get the script to populate the title field correctly. I tried and tried and failed, which is why I went to Python. I have a working script, but like I said, couldn’t get QS to run it. I know a lot more now than I did, so I’ll try again.
There are things I wish it parsed. For instance, http://projects.skurfer.com/QuicksilverPlug-inReference.mdown, but we could spend weeks arguing over what should be included, so I don’t know if it’s worth it. I suppose ideally, it would allow arrowing into any link that pointed to something that had a content-type of text/html, but that would require it to request each child from the web server which would probably be waaaaay too slow. |
I think that's probably the right order. Unfortunately I don't know perl enough (at all) to be able to alter the script so that it sets the As I said, I think the perl script works alright for the time being; we should probably try improving that first, but if the python script is better (got the email Rob) then it should be implemented - I don't feel this warrants the extra work at the time being (for me at least)
exactly. The list of 'allowed file types' could go on for ever and ever. I suggest I implement a method similar to that for checking tlds (an array) and then if we decide we should add a certain file type we just add it to an array. I've also noticed one more bug (which I don't plan to fix here, since I don't know enough about it) |
I got it to pass HTML through my Python script and it handles the image name precedence as described above, but I guess I need to wait until this gets merged before I submit my changes. This looks good to me. I’ll run with it for a few hours and merge it if I don’t see any problems. |
Still love it. Merged. |
I looked into it and that Perl module doesn’t even make those attributes available, so there’s nothing you could do. I tried other modules, but either they didn’t work, or weren’t available by default on Mac OS X. |
Nice :) I'm close to getting the web search icon to load reliably for web searches, I'll post again in the dev discussion incase somebody would like to take a On 13 May 2011 12:10, skurfer <
|
Lots of small bug fixes
<a href="link">some text</a>
was an image being linked
if(image.length)
, since the Link Parser was always returning an image string, just most of the time if was empty@""
http://google.com/services/
If I may say so myself, this is freakin awesome!
There are still a few bugs:
I think what we'd need to do here is compile a list of web page extensions, then do something similar to the check with .tlds. I couldn't be bothered compiling a list now.
P.S. @HenningJ - I've just seen loads of comments IconLoader.m:58 about 'a good place for setting the web search icon'. Was this you? I don't think it's the best place for it as it says in the comment.
What I know:
Search URLs that are added to a 'web search list' catalog entry are imported by the web search plugin. It should deal with the icons.
Search URLs that are added with the default catalog entry (from a remote URL) use the HTMLLinkParser. I've already started implementing a setIcon method there, you can see it commented out.