dbmdz / rosetta-openwayback-vpp Goto Github PK

Hi,
We've been using this vpp plugin for quite a while now with Rosetta, but we've just noticed an issue with the web server (Apache) rewrite when loading web harvests that have a query string section in their own URL.

For instance, if we load an IE from Rosetta with the following primary seed:

http://owb-server.nz/webarchive/wayback/query?is_mobile=false&detail=true&is_rtl=false&@/20190412103841/https://us3.campaign-archive.com/@&dps_dvs=<dps_dvs>&dps_pid=<dps_pid>

The correct parts of the query string get removed using the suggested rewrite commands:

RewriteEngine on
RewriteCond %{QUERY_STRING} ^(([^&]*&)*)@(.*)@(.*)$
RewriteRule ^/webarchive/wayback/query$ /webarchive/wayback%3? [L,PT]

and we get the following, correctly, loading in the browser

http://owb-server.nz/webarchive/wayback/20190412020848/https://us3.campaign-archive.com/

But if the IE has a primary seed that itself contains a query string, then the rewrite rule doesn't appear to work.
For instance:

http://owb-server.nz/webarchive/wayback/query?is_mobile=false&detail=true&is_rtl=false&@/20190412103841/https://us3.campaign-archive.com/?u=57af16fa15f95ed83e0b434a9&id=7b011dfb3a&e=69b9329363@&dps_dvs=<dps_dvs>&dps_pid=<dps_pid>

will not be rewritten, and the exact same URL will be loaded in the browser.

The only place I can think it would be failing is the {QUERY_STRING} parameter in the rewrite rule, that perhaps the second ? is messing with it. Unfortunately I haven't any solutions looking around Apache docs and posts online.

Has anyone else encountered this issue with multiple query string sections?

Thanks,
Ben