You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
The current runner executes all cregit steps sequentially, following the order in the README. However, some of these steps appear to be independent and could run in parallel.
This could reduce total runtime, especially for large repositories.
Current flow
flowchart TD
A[Create view repo] --> B[Create original repo history DB]
B --> C[Create cregit repo history DB]
C --> D[Create persons DB]
D --> E[Create blame of cregit files]
E --> F[Remap commits]
F --> G[Create HTML files]
Loading
Possible parallel flow
flowchart TD
A[Create view repo]
B[Create original repo history DB]
D[Create persons DB]
A --> C[Create cregit repo history DB]
A --> E[Create blame of cregit files]
C --> F[Remap commits]
C --> G[Create HTML files]
D --> G
E --> G
F --> G
Loading
Notes
A possible execution plan would be:
run create view repo, create original repo history DB, and create persons DB in parallel
after create view repo, run create cregit repo history DB and create blame of cregit files in parallel
after create cregit repo history DB, run remap commits
after all required artifacts are ready, run create HTML files
This should be validated against the current implementation, but it looks like a good opportunity to improve performance and make the pipeline dependency-aware instead of strictly sequential.
The current runner executes all cregit steps sequentially, following the order in the README. However, some of these steps appear to be independent and could run in parallel.
This could reduce total runtime, especially for large repositories.
Current flow
Possible parallel flow
Notes
A possible execution plan would be:
create view repo,create original repo history DB, andcreate persons DBin parallelcreate view repo, runcreate cregit repo history DBandcreate blame of cregit filesin parallelcreate cregit repo history DB, runremap commitscreate HTML filesThis should be validated against the current implementation, but it looks like a good opportunity to improve performance and make the pipeline dependency-aware instead of strictly sequential.