OurBigBook stores the timestamp of the last successful:
for each file.
For ID extraction, we always skip the extraction if the filesystem timestamp of a source file is older than the last successful extraction.
For render:
  • we also skip rendering by default when you invoke ourbigbook on a directory, e.g. ourbigbook ., as this greatly speeds up the interactive error fixing turnaround time
  • we always re-render fully when you specify a single file, e.g. ourbigbook path/to/README.bigb
However, note that skipping renders, unlike for ID extraction, can lead to some outdated pages.
This option disables the timestamp skip for rendering, so ensure that you will get a fully clean updated render.
E.g. consider if you had two files:
file1.bigb
= File 1

== File 1 1
file2.bigb
= File 2

== File 2 1

\x[file-1-1]
We then do the initial conversion:
ourbigbook .
we see output like:
extract_ids file1.bigb
extract_ids file1.bigb finished in 45.61287499964237 ms
extract_ids file2.bigb
extract_ids file2.bigb finished in 15.163879998028278 ms
render file1.bigb
render file1.bigb finished in 23.21016100049019 ms
render file2.bigb
render file2.bigb finished in 25.92908499762416 ms
indicating full conversion without skips.
But then if we just modify fil1.bigb as:
= File 1

== File 1 1 hacked
{id=file-1-1}
the following conversion with ourbigbook . would look like:
extract_ids file1.bigb
extract_ids file1.bigb finished in 45.61287499964237 ms
extract_ids file2.bigb
extract_ids file2.bigb skipped by timestamp
render file1.bigb
render file1.bigb finished in 41.026930000633 ms
render file2.bigb
render file2.bigb skipped by timestamp
and because we skipped file2.bigb render, it will still have the outdated "File 1 1" instead of "File 1 1 hacked".
We could in principle solve this problem by figuring out exactly which files need to be changed when a given ID changes, and we already have to solve a similar problem due to query bundling. Also, this will need to be done sonner or later for the OurBigBook Web. But lazy now: github.com/cirosantilli/ourbigbook/issues/207, this is hard stuff.

Ancestors