Thread: xml_is_document and selective pg_re_throw
Hi, <br /><br />Consider:<br /><br />SELECT xml '<foo>bar</foo><bar>foo</bar>' IS DOCUMENT;<br /><br/>And I was looking at xml_is_document() source code. It calls xml_parse which throws an error with code set to ERRCODE_INVALID_XML_DOCUMENT.The catch block of xml_parse then rethrows. <br /><br />Now xml_is_document does a selectiverethrow only if the error is not ERRCODE_INVALID_XML_DOCUMENT. I can understand that this function does this toreturn true/false, but doesn't this behavior of not propagating the error up all the way dangerous? InterruptHoldoffCountinconsistencies for instance? <br /><br />A better way would have been to modify xml_parse to take anadditional boolean argument "to_rethrow" and not to rethrow if that is false? Thoughts? <br /><br />Regards,<br />Nikhils<br />
Nikhil Sontakke <nikkhils@gmail.com> writes:
> Consider:
> SELECT xml '<foo>bar</foo><bar>foo</bar>' IS DOCUMENT;
> And I was looking at xml_is_document() source code. It calls xml_parse
> which throws an error with code set to ERRCODE_INVALID_XML_DOCUMENT. The
> catch block of xml_parse then rethrows.
> Now xml_is_document does a selective rethrow only if the error is not
> ERRCODE_INVALID_XML_DOCUMENT. I can understand that this function does this
> to return true/false, but doesn't this behavior of not propagating the
> error up all the way dangerous? InterruptHoldoffCount inconsistencies for
> instance?
No, I don't see any particular risk there. The places that might throw
ERRCODE_INVALID_XML_DOCUMENT are sufficiently few (as in, exactly one,
in this usage) that we can have reasonable confidence we know what the
system state is when we catch that error.
> A better way would have been to modify xml_parse to take an additional
> boolean argument "to_rethrow" and not to rethrow if that is false?
We could do that, but it would greatly complicate xml_parse IMO, since
it still needs its own PG_TRY block to handle other error cases, and
only one of those error cases ought to optionally return failure instead
of re-throwing.
regards, tom lane
No, I don't see any particular risk there. The places that might throw
ERRCODE_INVALID_XML_DOCUMENT are sufficiently few (as in, exactly one,
in this usage) that we can have reasonable confidence we know what the
system state is when we catch that error.
Hmmm, I was writing some code in which I happened to hold a LWLock when this function was called. The first catch/rethrow cleaned up the InterruptHoldoffCount value. A subsequent release of that LWLock tripped up the (Assert(InterruptHoldoffCount > 0);) inside RESUME_INTERRUPTS().
I know holding an lwlock like this might not be a good idea, but this behavior just got me thinking about other probable issues.
Regards,
Nikhils
> A better way would have been to modify xml_parse to take an additionalWe could do that, but it would greatly complicate xml_parse IMO, since
> boolean argument "to_rethrow" and not to rethrow if that is false?
it still needs its own PG_TRY block to handle other error cases, and
only one of those error cases ought to optionally return failure instead
of re-throwing.
regards, tom lane