srweb breakage 13/12/15

[Jeremy Morse]

Hi,

Today srweb was broken -- on the front page, part of the header was
served, then the rest of the page was a 404. This happened when puppet
was deployed at 0700, thanks to Tylers monitoring scripts I checked out
an old srweb at 0740 (cheers Tyler).

This is due to fe5b95391, aka g:2594 [0], aka 'Add some scripts to rule
them all', which I've reverted. Obviously this can be un-reverted when
there's a diagnosis for what happened, probably something wrong with the
rewrite rules.

While I appreciate that people want to develop srweb things on their own
platform at their own convenience, please remember that there is only
one place where everything has to work, which is on the actual
webserver. A vagrant instance from the puppet repo replicates that
almost perfectly, and the only real environment where integration
testing is feasible.

[0] https://www.studentrobotics.org/gerrit/#/c/2594/

--
Thanks,
Jeremy

[Tom Leese]

Hi,

On Sunday, 13 December 2015 10:12:31 UTC, jmorse wrote:

This is due to fe5b95391, aka g:2594 [0], aka 'Add some scripts to rule
them all', which I've reverted. Obviously this can be un-reverted when
there's a diagnosis for what happened, probably something wrong with the
rewrite rules.

While I appreciate that people want to develop srweb things on their own
platform at their own convenience, please remember that there is only
one place where everything has to work, which is on the actual
webserver. A vagrant instance from the puppet repo replicates that
almost perfectly, and the only real environment where integration
testing is feasible.

I've decided that to cement the idea that running srweb with my scripts is not the way the web server actually runs srweb, I have put them in a separate project: https://github.com/thomasleese/srweb-runner

Tom