Re: [PATCH] Fixed malformed error message on malformed SCRAM message.

On Tue, May 30, 2017 at 04:39:55PM -0700, Michael Paquier wrote:
> On Mon, May 29, 2017 at 2:40 PM, Alvaro Herrera
> <alvherre(at)2ndquadrant(dot)com> wrote:
> > Michael Paquier wrote:
> >> On Mon, May 29, 2017 at 2:06 PM, Alvaro Herrera
> >> <alvherre(at)2ndquadrant(dot)com> wrote:
> >> > These messages look all wrong to me.
> >>
> >> So your complain would be to do the following for each error message
> >> that uses parenthesis to include details? Like that I suppose:
> >> --- a/src/backend/libpq/auth-scram.c
> >> +++ b/src/backend/libpq/auth-scram.c
> >> @@ -283,11 +283,13 @@ pg_be_scram_exchange(void *opaq, char *input,
> >> int inputlen,
> >> if (inputlen == 0)
> >> ereport(ERROR,
> >> (errcode(ERRCODE_PROTOCOL_VIOLATION),
> >> - (errmsg("malformed SCRAM message (empty message)"))));
> >> + errmsg("malformed SCRAM message"),
> >> + errdetail("Empty message.")));
> >
> > Yeah, but along the lines of errdetail("The message is empty.")
>
> Okay. What do you think about the attached patch then? Does it address
> your concerns about the format of those error messages?
> --
> Michael

> diff --git a/src/backend/libpq/auth-scram.c b/src/backend/libpq/auth-scram.c
> index 99feb0ce94..366a11feb8 100644
> --- a/src/backend/libpq/auth-scram.c
> +++ b/src/backend/libpq/auth-scram.c
> @@ -283,11 +283,13 @@ pg_be_scram_exchange(void *opaq, char *input, int inputlen,
> if (inputlen == 0)
> ereport(ERROR,
> (errcode(ERRCODE_PROTOCOL_VIOLATION),
> - (errmsg("malformed SCRAM message (empty message)"))));
> + errmsg("malformed SCRAM message"),
> + errdetail("The message is empty.")));
> if (inputlen != strlen(input))
> ereport(ERROR,
> (errcode(ERRCODE_PROTOCOL_VIOLATION),
> - (errmsg("malformed SCRAM message (length mismatch)"))));
> + errmsg("malformed SCRAM message"),
> + errdetail("Input length does not match.")));

auth.c uses COMMERROR for this sort of thing; why does auth-scram.c use ERROR?