When combining and pipelining blocks, a special field called
subblocks is updated. This field indicates how the final block was
built.
When combining blocks, the subblocks are set as a list. For example,
(block-subblocks (combine-blocks* a b)) ;=> `(,a ,b)
Combining more blocks will append to the list.
(block-subblocks (combine-blocks* (combine-blocks* a b) c)) ;=> `(,a ,b ,c)
When running a pipeline, the subblocks of each pipeline element are inserted into a list. Elements without subblocks are inserted directly.
(block-subblocks (run-blocks*
(combine-blocks* a b)
c d)) ;=> `((,a ,b) ,c ,d)
A side effect of this is it is impossible to tell by looking at the
subblocks field alone whether a block created by running a
pipeline or by combining blocks. A block created by running a pipeline
of blocks that were not previously combined would look identical to a
block created by combining blocks.
The elements of the subblocks field are the original blocks.
When running a pipeline, the blocks are not executed directly.
Instead, the code field is updated by replacing chain, and
that new block is the one that executes. However, that is not
reflected in the subblocks. The subblocks code field will not be
replaced.