[Date Prev][Date Next] [Thread Prev][Thread Next] [Date Index] [Thread Index]

Re: Draft announcement for 10.7



On Fri, Dec 04, 2020 at 03:25:18PM +0000, Justin B Rye wrote:
> Roberto C. Sánchez wrote:
> > Jean-Pierre Giraud wrote:
> >> it would be best to change :
> >> s/possible crash with stacking unwinding/possible crash unwinding WITH
> >> stack/
> >> 
> > I disagree.  The text is correct as written. The issue is that stack
> > unwinding proceeds even after a parse error.
> 
> I think at there's at least a surplus "-ing" - shouldn't it be
> 	possible crash with stack unwinding
> ...?
> 
> > The parse error in question involves the SQL keyword WITH.
> 
> I would never have guessed that this was talking about a "WITH-stack";
> it's more natural to read it as a crash "with" particular attributes.
> If it's trying to refer to the keyword WITH, this is probably a second
> thing that needs fixing.
> 
> > The description says
> > "selectExpander in select.c in SQLite 3.30.1 proceeds with WITH stack
> > unwinding even after a parsing error" which can also be read as
> > "selectExpander in select.c in SQLite 3.30.1 can encounter a parse error
> > involving the WITH keyword and proceeds to unwind the stack even after
> > the error".
> 
> In other words, a possible crash involving unwinding of the "WITH"
> stack?  The best way I can see of abbreviating that without making it
> completely unintelligible is:
> 
> 	possible crash (unwinding WITH stack)
> 
> and now I look back I see that's essentially Jean-Pierre Giraud's
> suggestion.

I see.  However, the parentheses make it more clear than both
Jean-Pierre's suggestion and the original.  My vote would be for your
suggestion.

Regards,

-Roberto

-- 
Roberto C. Sánchez


Reply to: