I have a fairly simple PHP website's code sitting in a Mercurial repository.
Recently, this required a small change, and I tried to set up pipelines to help make the deployment part easier. Unfortunately the web hosting company only supports FTP access to upload website content (long story; TLDR my client insists on using this web host). So naturally I tried using atlassian/ftp-deploy:0.1.0, as that seems to be by far the simplest option.
Unfortunately that seems to be broken, as it uploads the website contents to /publich_html/build, and not to /public_html/ like I specified in `REMOTE_PATH`.
Here is my pipeline configuration:
pipelines: default: - step: name: Deploy to webhost script: - pipe: atlassian/ftp-deploy:0.1.0 variables: USER: $USERNAME PASSWORD: $PASSWORD SERVER: $FTPURL REMOTE_PATH: '/public_html/' # LOCAL_PATH: '<string>' # Optional # DEBUG: '<boolean>' # Optional
Why is it creating and uploading to a subdirectory? Can I make it upload to public_html instead?
I came across this similar question: https://community.atlassian.com/t5/Bitbucket-Pipelines-questions/SFTP-deploy-pipe-creates-build-folder-on-remote-and-fails-after/qaq-p/1053378
Unfortunately that answer (use git ftp) does not work for me, as this is a mercurial repo.
Hi @pretoriusjf . The ftp-deploy pipe is a bit different from the sftp-deploy. Basically, if you add a trailing slash to a REMOTE_PATH, the pipe uploads a base directory as well. If you only want to mirror the contents of the directory, you should remove the trailing slash from the REMOTE_PATH:
REMOTE_PATH: '/public_html'
Thank you! While I find this system quite counterintuitive, your suggestion worked for me.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.
You must be a registered user to add a comment. If you've already registered, sign in. Otherwise, register and sign in.