Our environment variable practice is slightly error-prone. Variables like SERVICE or DC_USER are sufficiently ambiguous that it wouldn't be surprising if they conflicted with variables set in someone's personal shell environment.
Others, like DISABLE_SENTRY in the frontend, are slightly more "one-off" that it would be surprising if they conflicted and we set those in scripts. It's not unfounded that someone could have DISABLE_SENTRY in their shell environment though, so disambiguating those might not be a bad idea either.
Suggested Improvement
Rename the following environment variables:
SERVICE -> OPENVERSE_CATALOG_EXEC_SERVICE
DC_USER -> OPENVERSE_DC_USER
DISABLE_SENTRY -> OPENVERSE_DISABLE_SENTRY
TEST_COMMAND -> OPENVERSE_PLAYWRIGHT_TEST_COMMAND
Benefit
Disambiguates these variables. The suggested names above are sometimes more descriptive and self-documenting than the current ones, especially in the monorepo context.
Current Situation
Our environment variable practice is slightly error-prone. Variables like
SERVICE
orDC_USER
are sufficiently ambiguous that it wouldn't be surprising if they conflicted with variables set in someone's personal shell environment.Others, like
DISABLE_SENTRY
in the frontend, are slightly more "one-off" that it would be surprising if they conflicted and we set those in scripts. It's not unfounded that someone could haveDISABLE_SENTRY
in their shell environment though, so disambiguating those might not be a bad idea either.Suggested Improvement
Rename the following environment variables:
SERVICE
->OPENVERSE_CATALOG_EXEC_SERVICE
DC_USER
->OPENVERSE_DC_USER
DISABLE_SENTRY
->OPENVERSE_DISABLE_SENTRY
TEST_COMMAND
->OPENVERSE_PLAYWRIGHT_TEST_COMMAND
Benefit
Disambiguates these variables. The suggested names above are sometimes more descriptive and self-documenting than the current ones, especially in the monorepo context.