Escaped fragment bug

CupidCatz · 2014-05-13T20:50:58.000Z

Hi i’m found so hashbang support work a little incorrect.
So for url http://example.com/#!id=1 all works fine
But for url http://example.com/?uid=1#!id=1 it not works correct.
Problem so it always replace “#!” to “?escaped_fragment=” without check if “?” already presented

Here is string from log:
99.16.156.124 - - [13/May/2014:22:09:43 +0200] “GET /photoTiles.php?aid=7748?escaped_fragment=3314 HTTP/1.1” 200 4006 “-” “Twitterbot/1.0” “-”

Could you please confirm and fix?

Thank you

andypiper · 2014-05-14T15:31:47.000Z

Hi, can you explain a little more of what you are trying to do?

As I understand it, you’re trying to create a photo card which points to a URL containing a hashbang - is that correct?

CupidCatz · 2014-05-14T16:35:42.000Z

Hi,

I’m have url like http://loveme.eu/video.php?1=5#!102
With summary card inside.
So twitter when read it replace “#!” to escaped fragment
and i’m get url like http://loveme.eu/video.php?1=5?escaped_fragment=102

So from one side it replaced, but seems like Twitterbot not checking if in url already exist "?"
so more correct url be like http://loveme.eu/video.php?1=5&escaped_fragment=102

Here is log from this url from nginx

199.16.156.124 - - [14/May/2014:09:02:18 +0200] “GET /video.php?1=5?escaped_fragment=102 HTTP/1.1” 302 5 “-” “Twitterbot/1.0” “-”

Sorry if i’m unclear

andypiper · 2014-05-14T23:22:50.000Z

That makes it more clear to me, thank you! let me take a look at this.

CupidCatz · 2014-05-14T23:34:36.000Z

Sure thank you.

I’m miss code tag before, so log looks like not full.
So here is logs examples in normal format
199.16.156.126 - - [14/May/2014:09:02:18 +0200] “GET /video.php?1=5?escaped_fragment=102 HTTP/1.1” 302 5 “-” “Twitterbot/1.0” “-”

199.16.156.124 - - [15/May/2014:01:30:55 +0200] “GET /photoTiles.php?aid=7748?escaped_fragment=3313 HTTP/1.1” 302 5 “-” “Twitterbot/1.0” “-”

I’m non sure why, but here if i’m use code or not use it cuts underscore character “_”, so assume so it in log

jbulava · 2014-07-02T19:49:10.000Z

@CupidCatz, the engineering team has pushed a fix for this issue to production. Apologies for the inconvenience. Please run some tests and if you’re still seeing issues, I’ll pass your feedback along to the team!