Re: [COMMITTERS] pgsql: Clean up the #include mess a little.

Ants Aasma <ants.aasma@eesti.ee>

From: Ants Aasma <ants.aasma@eesti.ee>

To: Alvaro Herrera <alvherre@commandprompt.com>

Cc: Greg Stark <stark@mit.edu>, Bruce Momjian <bruce@momjian.us>, Tom Lane <tgl@sss.pgh.pa.us>, PostgreSQL-development <pgsql-hackers@postgresql.org>

Date: 2011-09-06T22:20:17Z

Lists: pgsql-hackers

Attachments

render_includes.v2.py (text/x-python)
all_with_subgraphs.svgz (image/svg+xml-compressed)
parsenodes.h.svgz (image/svg+xml-compressed)
execnodes.h.svgz (image/svg+xml-compressed)

On Tue, Sep 6, 2011 at 10:18 PM, Alvaro Herrera
<alvherre@commandprompt.com> wrote:
> I wonder what happens if files in the same subdir are grouped in a
> subgraph.  Is that possible?

Possible, and done. Also added possivility to add .c files to the graph,
coloring by subdir and possibility exclude nodes from the graph. I didn't yet
bother to clean up the code - to avoid eye damage, don't look at the source.

Bad news is that it doesn't significantly help readability for the all nodes
case. See all_with_subgraphs.svgz.  It does help for other cases.
For example parsenodes.h.svgz has the result for
render_includes.py --select='nodes/parsenodes.h+*-*' --subgraphs
and execnodes.h.svgz for
--subgraphs --select='nodes/execnodes.h+*-*'

--
Ants Aasma