Hi Stewart, On Mon, Mar 29, 2010 at 9:41 AM, Stewart Smith <[email protected]>wrote:
> On Sun, 28 Mar 2010 12:04:46 -0700, Brian Aker <[email protected]> wrote: > > On Mar 28, 2010, at 2:28 AM, Zimin wrote: > > > > > As I see from the MySQL side, there is a connect string for this > > > 'create table' statement. But I don't find it's in Drizzle. Am I > > > right? And what's the usual practice to do such task? > > > > The interface in the parser needs to be changes to key/pair that can > > then be passed down and verified in the Storage Engine. > > Have you managed to look at my parser bits for this yet? IIRC it was > adding another shift/reduce conflict, and we don't like those... so was > hoping you had some magic solution. > > I've been looking into this branch tree these days and managed to solve the shift/reduce conflict newly introduced. originally, create_table_option ::= ENGINE_SYM equal ident_or_text | indent_or_text equal ident_or_text | etc... but IMHO, these above two sub-statements can be merged into one, so the new rule for create_table_options is reduced to indent_or_text equal indent_or_text | etc... And the word ENGINE can still appear as the key part. This can solve the shift/reduce conflict. I hope I'm not missing some corner cases with this change. The change can be tracked from my branch lp:~ziminq/drizzle/fs-plugin revno:1422 However, the overall code still needs some polishing to work, I'll work on that. Besides, I would like to hear from you what the custom table options finally should shape into. passing key-value pair from parser down to storage engine and other requirements? Thanks. --ziminq (then there's the rest of the patch, but the parser bit is a blocker no > matter what) > > -- > Stewart Smith >
_______________________________________________ Mailing list: https://launchpad.net/~drizzle-discuss Post to : [email protected] Unsubscribe : https://launchpad.net/~drizzle-discuss More help : https://help.launchpad.net/ListHelp
