The Bizarro Feed, or, When 404 Doesn't Mean 'Not Found'
(This post was migrated from the syn_training community, originally made by 
seaofdestiny.)
This was a strange request.
As the user mentioned, the profile page for
zakhad_blog mentioned that the last fetch resulted in an error '404 Not Found' -- but if you clicked on the 'XML' button on the profile, you could see content in your browser.
After a bit of digging, including a command-line utility (I used lwp-request, which comes with Perl's LWP package), I found something weird: when I fetched the feed, I got data, yes. But when I added some debugging flags to show me the HTTP status, it told me "404 Not Found"! A bit more digging with netcat and speaking raw HTTP to the server (always fun ^^~) corroborated that.
So it appears that the server, for whatever reason, was delivering data, but with a status code indicating that there was no data there! My browser, apparently, figured that if the content was not text/html (likely an error page) but text/xml, then there was indeed content despite the status code; but LiveJournal, apparently, chose to trust the status code (as I think it should have).
Fortunately, the user was able to provide an alternate feed location which behaved properly (returning '200 OK' as its HTTP status), and I used syn_editurl to change the feed URL to the new URL.
In related news, if anyone is interested in a quick tutorial in using lwp-request and/or netcat to diagnose syn-related problems, let me know. If there's interest, I'll try to make time to write something down about what I do.
(I use a Unix shell for this, but Perl runs on Windows, too, and I wouldn't be surprised if netcat had been ported to Windows as well, so the information might be applicable even if you don't have access to a Unix shell but 'only' a Windows command line.)
This was a strange request.
As the user mentioned, the profile page for
zakhad_blog mentioned that the last fetch resulted in an error '404 Not Found' -- but if you clicked on the 'XML' button on the profile, you could see content in your browser.After a bit of digging, including a command-line utility (I used lwp-request, which comes with Perl's LWP package), I found something weird: when I fetched the feed, I got data, yes. But when I added some debugging flags to show me the HTTP status, it told me "404 Not Found"! A bit more digging with netcat and speaking raw HTTP to the server (always fun ^^~) corroborated that.
So it appears that the server, for whatever reason, was delivering data, but with a status code indicating that there was no data there! My browser, apparently, figured that if the content was not text/html (likely an error page) but text/xml, then there was indeed content despite the status code; but LiveJournal, apparently, chose to trust the status code (as I think it should have).
Fortunately, the user was able to provide an alternate feed location which behaved properly (returning '200 OK' as its HTTP status), and I used syn_editurl to change the feed URL to the new URL.
In related news, if anyone is interested in a quick tutorial in using lwp-request and/or netcat to diagnose syn-related problems, let me know. If there's interest, I'll try to make time to write something down about what I do.
(I use a Unix shell for this, but Perl runs on Windows, too, and I wouldn't be surprised if netcat had been ported to Windows as well, so the information might be applicable even if you don't have access to a Unix shell but 'only' a Windows command line.)