| From: | "Hitoshi Harada" <umi(dot)tanuki(at)gmail(dot)com> |
|---|---|
| To: | "Tom Lane" <tgl(at)sss(dot)pgh(dot)pa(dot)us> |
| Cc: | pgsql-hackers(at)postgresql(dot)org |
| Subject: | Re: Location for window function API definitions |
| Date: | 2008-12-22 04:55:27 |
| Message-ID: | e08cc0400812212055lf058b81p4e5d28dd2a3cf998@mail.gmail.com |
| Views: | Whole Thread | Raw Message | Download mbox | Resend email |
| Thread: | |
| Lists: | pgsql-hackers |
2008/12/22 Tom Lane <tgl(at)sss(dot)pgh(dot)pa(dot)us>:
> The window functions patch defines some API functions and macros for
> window functions. These are currently defined in executor/nodeWindow.h,
> but that seems to me like a fairly inappropriate location. The executor
> main body, which would be the normal consumer of that include file, has
> no interest in the API functions; and conversely the window-plan-node
> functions that are defined for the executor are not things for window
> functions to call. So I think the API definitions need to go someplace
> else.
>
> The only existing file that seems appropriate is include/funcapi.h, but
> that also contains a lot of stuff that wouldn't be useful to a typical
> window function.
>
> If we invent a new include file for the purpose, possible names might be
> include/windowapi.h
+1
--
Hitoshi Harada
| From | Date | Subject | |
|---|---|---|---|
| Next Message | Hitoshi Harada | 2008-12-22 04:57:42 | Re: Is "Window" really safe as a global typedef name? |
| Previous Message | David Fetter | 2008-12-22 04:36:02 | Re: about truncate |