aboutsummaryrefslogtreecommitdiff
path: root/src/backend/executor/nodeFunctionscan.c
diff options
context:
space:
mode:
authorMichael Paquier <michael@paquier.xyz>2023-01-12 14:23:20 +0900
committerMichael Paquier <michael@paquier.xyz>2023-01-12 14:23:20 +0900
commit8607630d74cd83f636a72eb9f2eb62f67e1fb955 (patch)
tree228eebedb7ac7abcaf0ffeedb98fabbcebf18c57 /src/backend/executor/nodeFunctionscan.c
parentbfd2542001a4a792a82a09607524e9b78a02217c (diff)
downloadpostgresql-8607630d74cd83f636a72eb9f2eb62f67e1fb955.tar.gz
postgresql-8607630d74cd83f636a72eb9f2eb62f67e1fb955.zip
Rename some variables related to ident files in hba.{c,h}
The code that handles authentication for user maps was pretty confusing with its choice of variable names. It involves two types of users: a system user and a Postgres user (well, role), and these were not named consistently throughout the code that processes the user maps loaded from pg_ident.conf at authentication. This commit changes the following things to improve the situation: - Rename "pg_role" to "pg_user" and "token" to "system_user" in IndetLine. These choices are more consistent with the pg_ident.conf example in the docs, as well. "token" has been introduced recently in fc579e1, and it is way worse than the choice before that, "ident_user". - Switch the order of the fields in IdentLine to map with the order of the items in the ident files, as of map name, system user and PG user. - In check_ident_usermap(), rename "regexp_pgrole" to "expanded_pg_user" when processing a regexp for the system user entry in a user map. This variable does not store a regular expression at all: it would be either a string or a substitution to \1 if the Postgres role is specified as such. Author: Jelte Fennema Discussion: https://postgr.es/m/CAGECzQTkwELHUOAKhvdA+m3tWbUQySHHkExJV8GAZ1pwgbEgXg@mail.gmail.com
Diffstat (limited to 'src/backend/executor/nodeFunctionscan.c')
0 files changed, 0 insertions, 0 deletions