Re: [HACKERS] CREATE CAST WITHOUT FUNCTION should require superuserness?
Tom Lane writes: Okay. Are you intending to work on it? I was thinking of doing some cleanup work in parse_coerce, but will refrain from joggling your elbow if you're going to deal with it. Feel free. -- Peter Eisentraut [EMAIL PROTECTED] ---(end of broadcast)--- TIP 4: Don't 'kill -9' the postmaster
Re: [HACKERS] CREATE CAST WITHOUT FUNCTION should require superuserness?
Tom Lane writes: I think it might be a good idea to require superuser permissions to create a binary-equivalence cast. OK Also, in the present state of the code a WITHOUT FUNCTION cast does not work if AS ASSIGNMENT isn't given: parse_coerce.c won't ever find it. Is this intentional, or just an oversight? It's a transition state that might be inadequately documented. In the near future the code should be restructured to allow for non-implicit binary compatible casts. But currently there are not candidates for that anyway. -- Peter Eisentraut [EMAIL PROTECTED] ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
Re: [HACKERS] CREATE CAST WITHOUT FUNCTION should require superuserness?
Peter Eisentraut [EMAIL PROTECTED] writes: Tom Lane writes: Also, in the present state of the code a WITHOUT FUNCTION cast does not work if AS ASSIGNMENT isn't given: parse_coerce.c won't ever find it. Is this intentional, or just an oversight? It's a transition state that might be inadequately documented. Okay. Are you intending to work on it? I was thinking of doing some cleanup work in parse_coerce, but will refrain from joggling your elbow if you're going to deal with it. regards, tom lane ---(end of broadcast)--- TIP 3: if posting/reading through Usenet, please send an appropriate subscribe-nomail command to [EMAIL PROTECTED] so that your message can get through to the mailing list cleanly
[HACKERS] CREATE CAST WITHOUT FUNCTION should require superuserness?
CREATE CAST WITHOUT FUNCTION is capable of creating binary equivalences that will crash the backend when used (eg, between pass-by-value and pass-by-reference datatypes). The existing restriction that you must own one of the datatypes hardly seems like an adequate permissions check ... especially since any unprivileged user is able to create his own datatypes. I think it might be a good idea to require superuser permissions to create a binary-equivalence cast. This is equivalent to the permissions you'd need to create a dummy C function so as to cause the same failure without using WITHOUT FUNCTION. Comments? Also, in the present state of the code a WITHOUT FUNCTION cast does not work if AS ASSIGNMENT isn't given: parse_coerce.c won't ever find it. Is this intentional, or just an oversight? regards, tom lane ---(end of broadcast)--- TIP 1: subscribe and unsubscribe commands go to [EMAIL PROTECTED]
