Re: Updates of SE-PostgreSQL 8.4devel patches (r1608)

Bruce Momjian wrote:
> KaiGai Kohei wrote:
>> Bruce Momjian wrote:
>>> KaiGai Kohei wrote:
>>>> The series of SE-PostgreSQL patches for v8.4 were updated:
>>>> [1/5] http://sepgsql.googlecode.com/files/sepgsql-core-8.4devel-r1608.patch
>>>> [2/5] http://sepgsql.googlecode.com/files/sepgsql-utils-8.4devel-r1608.patch
>>>> [3/5] http://sepgsql.googlecode.com/files/sepgsql-policy-8.4devel-r1608.patch
>>>> [4/5] http://sepgsql.googlecode.com/files/sepgsql-docs-8.4devel-r1608.patch
>>>> [5/5] http://sepgsql.googlecode.com/files/sepgsql-tests-8.4devel-r1608.patch
>>>>
>>>> - List of updates:
>>>>   * bugfix: sepgsqlCheckProcedureEntrypoint() was invoked twice when
>>>>     security invoker functions are invoked.
>>>>
>>>> Rest of parts are unchanged. Don't mind contracted filename.
>>>> Please comment anything. It will help to improve our code.
>>> I did an analysis of the "core" file:
>>>
>>>     http://sepgsql.googlecode.com/files/sepgsql-core-8.4devel-r1608.patch
>>>
>>>     changed lines      3226
>>>     new files          4075
>>>     syscatalog         9977
>>>     ----
>>>     total             17278
>>>
>>> The good news is that 3226 is the affect on the non-system-catalog main
>>> core code, and is a context diff size, not total changed lines.
>> Hum...? What utility did you use to compute the lines?
>> It seems to me the changed lines except for system catalogs are larger than
>> actual one.
>>
>> The diffstat says:
>>   65 files changed, 4769 insertions(+), 11 deletions(-), 4945 modifications(!)
>>
>> The (4244 + 500) of 4945 modifications come from pg_proc.h and pg_attribute.h
>> due to a new field to store security label of procedures and columns.
>>
>> The new files adds 4014 in total, so rest of (755 + 11 + 201 = 967) lines are
>> estimated changes in the main core code.
>>
>> Anyway, I believe the burden of reviewer became smaller than the prior
>> full-set version.
> 
> I manually created the three files and then got totals;  I have attached
> the files.

OK, I can understand.
It also counts unchanged lines, like:

|  *** base/src/backend/Makefile   Sat Jan  3 13:01:35 2009
|  --- sepgsql/src/backend/Makefile        Fri Feb  6 13:13:08 2009
|  *************** include $(top_builddir)/src/Makefile.glo
|  *** 16,22 ****
|
|    SUBDIRS = access bootstrap catalog parser commands executor foreign lib libpq \
|          main nodes optimizer port postmaster regex rewrite \
|  !       storage tcop tsearch utils $(top_builddir)/src/timezone
|
|    include $(srcdir)/common.mk
|
|  --- 16,22 ----
|
|    SUBDIRS = access bootstrap catalog parser commands executor foreign lib libpq \
|          main nodes optimizer port postmaster regex rewrite \
|  !       security storage tcop tsearch utils $(top_builddir)/src/timezone
|
|    include $(srcdir)/common.mk

But it is not a point to be discussed here.

The significant point is the scale of patches are indeed reduced
as I promised before, by separating a few facilities. :)

I've waited for reviewer's comment.

Thanks,
-- 
OSS Platform Development Division, NEC
KaiGai Kohei <kaigai@ak.jp.nec.com>