Open sylvain-morin opened 2 years ago
Thanks indeed @sylmorin-gbif.
I think that this issue is something weird that happens because we are using a docker environment (and BTW didn't happened to me before, only to other users).
The deploy script is this:
https://github.com/living-atlases/generator-living-atlas/blob/master/generators/app/templates/deploy-branding.sh
and yarn install
should install the deps.
Some related discussion: https://atlaslivingaustralia.slack.com/archives/CCTFGEU1G/p1637074047119100?thread_ts=1635761393.047800&cid=CCTFGEU1G
I thought this was fixed already with this commit: https://github.com/living-atlases/la-toolkit/commit/dc655e95aa6bf8f68b29e21c9999e1a6977431ec but as this is something that I cannot reproduce, I'm trying to fix reading about similar issues in the net, ...
Did you exec npm install
from the console?
I totally used the slack discussion you mentioned to fix my issue :) Executing the npm install from the console (so, inside the docker container)
By the way (since I prepared my demo, I rehearsed serveral times), once, I did the npm install from my laptop, in the branding config directory - but not from the docker container
What a mistake !!! since my local npm version was not the same than the npm in the docker container, everything was messy after that...
That's why it would be great if the deploy script could execute it, from the docker container, don't you think?
I still cannot reproduce, I tried several deployments and all worked for me.
Probably related:
I do npm install
(for things not related with the branding) and yarn install
for the branding inside the Dockerfile
so the deps should be there.
Branding deployment is failing because some npm libs are missing.
Executing:
npm install npm update
is fixing the issue.
It would be great to have it executed by default