Your first build
Build a target, see what it produced, and get the files out.
Build it
bldr build site --name "first build"You get a live tree of what the build is doing. --name is worth the habit: it
is what you see in bldr builds afterwards, and "first build" reads better
than a UUID.
The first run pulls the base image and runs your pod. The second run does neither:
bldr build site --name "again"Nothing changed, so nothing ran. That is not a heuristic. Your member's content hashed the same, so the pod's inputs hashed the same, so bldr already had the answer.
Now change the HTML in .bldr.ts and build again. That pod runs, because its
input is different. Nothing else does.
Get the files
A build produces content in the node's store. To put it on your disk, ask for an export:
bldr build --export site/site:./out --name "export the site"
cat out/index.html--export <filter>:<path> writes the tree of the one target the filter
matches to a path. The filter takes a target's full name, the member and then
the target, so the site output of the site member is site/site. Naming an
action also selects its target, so you do not have to ask for both.
Other things you can attach to a build:
--export site/site:./out | write the tree to a path |
--ref release | point a named ref at the output |
--mount site:/mnt/site | FUSE-mount a named output, no copy |
See what else you could build
bldr build --no-targets --name "what is here"That resolves your build code and reports every target the workspace declares without producing anything. It is the fastest way to ask "does my build code still work, and what does it think it can build".
Next: run it.