On 11/14/18, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Looks like good stuff, pushed with a small additional amount of work.
Thanks.
>> -Does ECPG still need its own functions for mm_alloc() and
>> mm_strdup(), if we have equivalents in src/common?
>
> The error handling isn't the same, so I don't think we can just replace
> them with the src/common functions. It is, however, potentially worth
> our trouble to fix things so that within pgc.l, palloc() and pstrdup()
> are macros for mm_alloc/mm_strdup, thereby further reducing the diffs
> to the core lexer.
>
> I'd be pretty tempted to also change all the hard references to
> base_yylval to go through a pointer variable, so that diffs like
> this go away:
>
> - yylval->str = pstrdup(yytext);
> + base_yylval.str = mm_strdup(yytext);
>
> I believe that that'd result in more compact code, too --- on most
> machines, references to global variables are more expensive in
> code bytes than indirecting through a local pointer variable.
I'll look into that as time permits.
On 11/14/18, Tom Lane <tgl@sss.pgh.pa.us> wrote:
> Alvaro Herrera <alvherre@2ndquadrant.com> writes:
>> On 2018-Nov-13, Tom Lane wrote:
>>> More generally, I wonder if it'd be practical to make pgc.l backup-free
>>> altogether. I'm not sure that the resulting gain in lexing speed would
>>> really be of interest to anyone,
>
>> Maybe time to compile the ecpg test cases during "make check-world"?
>
> I'm dubious that the lexer is a significant part of that, though I could
> be wrong ...
If it were, it'd be much easier to try a Flex flag other than the
default, most compact representation, something like -Cfe. That'd be a
prerequisite for no-backtracking to matter anyway. That's easy enough
to test, so I'll volunteer to do that sometime.
-John Naylor