Why does this Stack Overflow API query return binary data?

Question

The Stack Overflow API query

http://api.stackoverflow.com/1.1/search?tagged=javascript&intitle=%3Cdiv%3E

returns what appears to be binary data:

���I�%&/m�{J�J��t��$ؐ@�������iG#)�*��eVe]f@�흼��{����{����;�N'���?\fdl��J�ɞ!���?~|?"��Ey�')=��y6�����h����z��<=���~|W>�Z5�u�A:�f��/>����y�(��G�d^ϲe��!z���Ѳ�Y)5��G��A�8�VeU?������/I=����Yd�E��n�գt����/�q� �IU�n �ۻ�/No�������N�|� ��&�}�0��euU���?_O�Y��T˦*s�3�0_�/� �ۼV�@ʄJ��TH�����Y>���-���tY-���e^7�U���ŋ:�[���+�'U�V�G����]�ݻU��e��F�v{^/gy]�{�#����.���u��su�S��]���Uq�?��'賏��E�d�� B����w�^�5�[�*-���w?N�W���2����:�*f���ݝ��=�,|�+h~�%�{��.���0�;J]�-�������E��l���q]d�(�vN����(�<_�����z�:[6ۯ��u�գ��_M^�id����i�L�b%3�����x�fn��Ll5%�����v�� �f������m�nh���&o�by��D��F�@��/� @��6+�&����~��5A��-�,�+B��U�����9wE�5}�5$w��aS���(�WW�b���'�uuECh�z�\����Y��K�tN�8�Ҍ�����S�#P1}S�����7H"F�ʧ�y1�!.���.�ie��7 +������FID����Y�KR Ly���^��i�]kROK��E��=��'�i�7��dMA��:Ĺº���:|cy|���%���n� �i�|��!"Q�yA�m���,����Z�Y n��fKH���r�(g�����c�Oq~�w��B�Y~����U.<�c����?~���S��'v]UŒ�6��W��������Uܗ��Zg�3O�lq-@�1���_�ybq�=�O�����?���8<��

Why?

Answer 1

ASP.NET was freaking out about the < in your search term.

Fixed it*.

More generally, the API always returns compressed content; so everything is "binary," this case just showed it more obviously due to the unusual error.

_{*It's worth noting that the actual < and > are not indexed by our Lucene search provider, so you can't actually search for <div>: you'll be searching for div instead. It doesn't explode anymore if you try though.}

Answer 2

I guess there are two problems:

1. Leaving out the opening %3C does show (different?) results, so why does that opening bracket have this effect?

2. With the opening angle bracket, this yields an ASP error page, but gzipped. However, the response headers don't say it's compressed, leaving the browser (or whatever application is handling this) oblivious about it:

   curl -o output.txt -v --compressed \
   http://api.stackoverflow.com/1.1/search?tagged=javascript\&intitle=%3Cdiv%3E
   
   > GET /1.1/search?tagged=javascript&intitle=%3Cdiv%3E HTTP/1.1
   > User-Agent: curl/7.19.7 (universal-apple-darwin10.0) libcurl/7.19.7 [...]
   > Host: api.stackoverflow.com
   > Accept: */*
   > Accept-Encoding: deflate, gzip
   
   < HTTP/1.1 500 Internal Server Error
   < Cache-Control: private
   < Content-Type: text/html; charset=utf-8
   < X-AspNet-Version: 4.0.30319
   < Date: Fri, 01 Apr 2011 09:22:43 GMT
   < Content-Length: 1194
   

   However, manually uncompressing does result in an "Server Error in '/' Application" error page. Also, a normal response does include the proper header:

   < HTTP/1.1 200 OK
   < Cache-Control: private
   < Content-Type: application/json; charset=utf-8
   < Content-Encoding: gzip
   < [...]