I think it's the http headers that are being sent by Bitbucket. I have this in a Dockerfile:
ADD https://bitbucket.org/teros/datavana/raw/"$commit"/moodwire/mwapi/requirements.txt '/teros/temp/requirements.txt'
ADD https://bitbucket.org/teros/datavana/raw/"$commit"/moodwire/mwapi/constraints.txt '/teros/temp/constraints.txt'
but the problem is those files haven't changed, but Docker doesn't cache them - it re-runs them:
Step 32/63 : ARG commit
---> Using cache
---> a60e2b6be500
Step 33/63 : ADD https://bitbucket.org/teros/datavana/raw/"$commit"/moodwire/mwapi/requirements.txt '/teros/temp/requirements.txt'
Downloading 1.545kB
---> 2e3b379a1ac4
Step 34/63 : ADD https://bitbucket.org/teros/datavana/raw/"$commit"/moodwire/mwapi/constraints.txt '/teros/temp/constraints.txt'
Downloading 1.545kB
is because Bitbucket is not responding with an HTTP 300 code? Maybe there is a query param I can use something like:
https://bitbucket.org/teros/datavana/raw/"$commit"/moodwire/mwapi/constraints.txt?use_cache=yes
or something like that?