GitLab

Previously resource_ids requires manual curation. These files contain
an implicit "structure" that imposes grouping, dependency (avoids
overlap), and parallel streams (allows overlap) of item lists.

It turns out that the implicit structure of start IDs can be modeled
as a Series-Parallel (SP) Graph. By introducing "META": {"join": #}
fields to resource_ids, the graph structure can be made concrete.

This CL adds a new GRIT tool that updates resource_ids file while
preserving structure. The inputs are:
* Existing start ID assignment, whose numerical sequence (along with
  newly added "join" meta data) stores the SP Graph.
* GRD files linked to by resource_ids, which are parsed to find per-tag
  ID usages.
  * For GRD files that may be unavailable (e.g., generated ones, or
    src-internal ones), "META": {"size": {...}} fields is needed to
    specify bounds for per-tag ID usage.

The main output is the resource_ids file with renumbered start IDs.
Debug flags are available to access intermediate data.

Chrome's main resource_ids file is updated to include the new "META"
flags. However, as a precaution, the start IDs are not yet reassigned
using the tool -- this will be done as a follow-up.

Bug: 979886

Change-Id: If7c0a43f57a1d27950784da7eb8405a6a0567f03
Reviewed-on: https://chromium-review.googlesource.com/c/chromium/src/+/1895736
Commit-Queue: Samuel Huang <huangs@chromium.org>
Reviewed-by: Andrew Grieve <agrieve@chromium.org>
Cr-Commit-Position: refs/heads/master@{#715838}