Gump generates data about the state of interactions of OSS projects, this data is the real work product of Gump (not the output artifacts stored in it's repositories).

Gump interacts with DBM databases (except on M$ where Python does not support them) and MySQL databases.

Simple Statistics

Gump stores simple 'running counter' type statistics for each GumpEntity. These records have the following form:

  • Name
  • Successes
  • Failures
  • Prereqs (not done due to pre-requisite failures)
  • datetime of first (success)
  • datetime of last (success) .. i.e. latest.
  • current state (e.g. success)
  • previous state
  • datetime of start of this state
  • sequence (in runs) in this state

If no MySQL database is available Gump uses DBM for these

Historical Information

If a MySQL database is available Gump uses it with the following GumpDatabase Schema.

On a per GumpEntity basis Gump stores run information:

these two (above) constitute the primary key

  • The state
  • The reason code
  • The cause (which other GumpEntity caised grief)
  • Start datetime
  • End datetime

Possible extras

  • Number of outputs
  • Total size of outputs
  • Numbers (including totals) of dependees/dependencies
  • No labels