On Wed, Nov 13, 2019 at 04:05:20PM +0900, Michael Paquier wrote: > On Wed, Nov 13, 2019 at 02:29:49PM +0900, Amit Langote wrote: > > On Wed, Nov 13, 2019 at 2:18 PM Michael Paquier <mich...@paquier.xyz> wrote: > >> There could be an argument for keeping > >> GET_STRING_RELOPTION actually which is still useful to get a string > >> value in an option set using the stored offset, and we have > >> the recently-added dummy_index_am in this category. Any thoughts? > > > > Not sure, maybe +0.5 on keeping GET_STRING_RELOPTION. > > Thinking more about it, I would tend to keep this one around. I'll > wait a couple of days before coming back to it.
Committed this one and kept GET_STRING_RELOPTION(). With the removal of those macros, it is possible to actually move a portion of the parsing definitions to reloptions.c for each type, but as we expose the validation function string and the enum element definition that would be more confusing IMO, so I have left that out. -- Michael
signature.asc
Description: PGP signature
