On Fri, Aug 21, 2015 at 1:02 PM, Simon Slavin <slavins at bigfraud.org> wrote:
> On 21 Aug 2015, at 8:13pm, Scott Hess <shess at google.com> wrote: > > Since renameTriggerFunc() > > follows renameParentFunc(), my guess is that triggers are also handled. > > The documentation says that statements within TRIGGERs are not changed > correctly: > > <https://www.sqlite.org/lang_altertable.html> > > Perhaps renameParentFunc() suggests that you can rename a TRIGGER, not > that the commands within TRIGGERs are changed. > > I do not know whether the current version of SQLite correctly changes > TRIGGERs or not. Since it would require complete parsing of the action > text, I suspect not. Hmm. My impression of both triggers and views is that they don't really have independent existence, instead they are inlined into relevant statements. So that would make sense as an omission. That said, writing code to do this manually has potential to be error prone. It might make sense under an ENABLE flag. It feels like an obvious thing to have, not something out-of-scope like embedding a new language or support for a new storage methodology. -scott
