Hi Chao, On Mon, Dec 15, 2025 at 7:26 AM Chao Li <[email protected]> wrote: > > Hi Hackers, > > I noticed this while working on a patch with tablecmd: > > ``` > /* > * If the table at either end of the constraint is partitioned, we need to > * recurse and handle every unvalidate constraint that is a child of this > * constraint. > */ > ``` > > Here “unvalidate” looks abnormal to me but not an obvious spelling error. > Then I searched over the source tree, and found two similar occurrences: > > Also in tablecmd.c > ``` > /* an unvalidated constraint is no good */ > ``` > > In pg_dump.c > ``` > /* > * An unvalidated constraint needs to be dumped separately, so > * that potentially-violating existing data is loaded before > * the constraint. > */ > ``` > > Notice the difference: “unvalidate” vs “unvalidated”. > > Then I recalled that Tom referred to the online dictionary > https://www.onelook.com for checking typo in discussion [1], and I checked > with the web site. It shows that “unvalidate” is not found; and “unvalidated” > has 3 matches and means “not confirmed as true yet”. > > So, I believe “unvalidate” is a typo, a “d” should be appended. > > [1] https://postgr.es/m/[email protected] > > Best regards, > -- > Chao Li (Evan) > HighGo Software Co., Ltd. > https://www.highgo.com/ >
Unvalidate does seem grammatically incorrect here. Unvalidated, though uncommon, looks like a valid term for this usage. [1] [2] [1] https://english.stackexchange.com/questions/58584/alternative-of-unvalidated [2] https://books.google.com/ngrams/graph?content=unvalidated&year_start=1930&year_end=2022&corpus=en&smoothing=3&case_insensitive=false -- Best, Xuneng
