Is it possible to somehow view the source in a raw format of what the Card Validator gets?
Unfortunately that isn’t possible, but it’s a good idea. Will add it to our list of possible enhancements.
In general what I do to test this is to use
curl -A Twitterbot <URL> to determine what the validator is seeing.
Is it at all possible that the curl command above won’t match what the actual Twitterbot will get?
When I use the curl command I get what I expect. There are 5 twitter meta tags. However with the card validator it says it can’t render and that there are no meta tags found. I don’t think it is a caching issue because if I prefix ‘?’ with any random numbers the result doesn’t change.
My site is a dynamically generated one, so to make Twitterbot work I have to pass requests through a CGI script to get the generated site. I noticed that if I actually just reroute Twitterbot to a hard static file, it works fine. But if I have the CGI script print the file then it doesn’t…Maybe there is some weird stray character in there? Headers missing? I don’t know.
FYI here is a link that should work in the card validator: http://johnwissler.com/portfolio?103821 (You can pick any random numbers for the end, it gets redirected to the same page)
With curl -A Twitterbot I see the right thing, but the card validator says there are no meta tags.
I was able to figure out my problem. My CGI script wasn’t outputting the ‘Content-Type’ header which the Twitterbot apparently requires. The requirement may have been specified somewhere in the documentation, however it would be nice if the card validator gave an error telling you that you were missing the Content-Type header.
Thanks for the update and glad you figured it out. We should certainly make sure that is somewhere in the docs, and I’ll take that on board!