The infer-upstream package

[Tags:github, mit, program]

Find the repository from where a given repo was forked

[Skip to Readme]


Dependencies ansi-wl-pprint (>=0.5 && <0.7), base (>=4.0 && <5.0), github (==0.8.*), optparse-applicative (>=0.5), parsec (>=3.0 && <3.2), process (>=1.1 && <1.3), text (>=0.10 && <0.12) [details]
License MIT
Author Noon Silk
Stability Unknown
Category Development
Home page
Uploaded Fri Aug 15 03:28:28 UTC 2014 by NoonSilk
Distributions NixOS:
Downloads 415 total (8 in the last 30 days)
0 []
Status Docs not available [build log]
Successful builds reported [all 6 reports]


Maintainer's Corner

For package maintainers and hackage trustees

Readme for infer-upstream

Readme for infer-upstream-


Takes a repo name, and looks up the upstream repository. If there is an upstream repository, writes it to standard out. Otherwise writes nothing.


noon@dev> infer-upstream -r scirate3 -u silky

Another usage:

noon@dev> cd scirate3
noon@dev> git remote add upstream `infer-upstream --using-cwd`

A more interesting usage (and the reason I wrote this) is to use the script. It performs the following task.

For all folders in a given directory:

  • go into each one,
  • if it is a github repo,
  • look up the upstream repo,
  • if we find it,
  • set it as a new remote.


noon@~> cd dev
noon@dev> git clone
noon@dev> infer-upstream/

With upstream set on your repos, you can then run, if you like, which will bring down any incoming changes and give print out a short summary.


fetching upstream for Javascript-Voronoi ...
 1 file changed, 2 insertions(+), 2 deletions(-)


  • This script assumes every folder in the ~/dev directory is the clone of a Github repository.
  • Github currently limits you to 60 unauthenticated API requests per hour; this approach uses one API request per folder.