MIME-Version: 1.0
References: 
 <CAKAnmmKVwP=yr8t8p8_QtrBFjyENLZyPxscJh4UrkCeqMQuhrw@mail.gmail.com>
 <4042a046-3dc9-4fc6-bfc6-62beb8492f27@eisentraut.org>
 <2228785.1769016115@sss.pgh.pa.us>
 <CAKAnmmK+1Rxv3FSKCku=ePNcwkcd6fqZ5C=y7_3wjyd4+BbwhQ@mail.gmail.com>
 <ab27a660-72b4-41b6-9279-2f007e065fed@eisentraut.org>
In-Reply-To: <ab27a660-72b4-41b6-9279-2f007e065fed@eisentraut.org>
From: Greg Sabino Mullane <htamfids@gmail.com>
Date: Mon, 23 Mar 2026 18:54:28 -0400
Message-ID: 
 <CAKAnmmKWEffLVRRSYkqKMKoK6RoZc1ecT_rPzMkVsuqOgzbjog@mail.gmail.com>
Subject: Re: [PATCH] Provide support for trailing commas
To: Peter Eisentraut <peter@eisentraut.org>
Cc: "pgsql-hackers@lists.postgresql.org" <pgsql-hackers@lists.postgresql.org>,
 Tom Lane <tgl@sss.pgh.pa.us>
Content-Type: multipart/alternative; boundary="000000000000ccf8a1064db8ec05"
Archived-At: 
 <https://www.postgresql.org/message-id/CAKAnmmKWEffLVRRSYkqKMKoK6RoZc1ecT_rPzMkVsuqOgzbjog%40mail.gmail.com>
Precedence: bulk

--000000000000ccf8a1064db8ec05
Content-Type: text/plain; charset="UTF-8"

Alright, I walked through the entire list of every place the original patch
added a comma, and ruthlessly trimmed out the ones that were not really
needed, likely to cause issues, and/or frankly applied on too trivial a
piece of SQL to worry about. Here's my list that people can poke at. Once
we settle on which ones survive, I can make a new patch. The focus is
mostly on the largest potential uses: lists of select items / column names.

Items to keep (i.e. add trailing comma support):

utility_option_list
TableElementList
opt_column_list
optional_period_name
index_params, index_including_params
table_func_column_list
insert_column_list
opt_search_clause (also has some minor refactoring and adds
depth_or_breadth)
sortby_list
group_by_list
vacuum_relation_list
from_list
xmltable_column_list
xml_namespace_list
array_expr_list
target_list
qualified_name_list


Items I am not sure about:

TriggerFuncArgs (seems sensible, low impact)
enum_val_list (since it is kind of a column list)
any_name_list / role_list (weak +1)
type_name_list (trivial but rarely used)
set_clause_list (weak +1)
relation_expr_list (mostly for LOCK)
rowsfrom_list (trivial)
json_table_column_definition_list (another sort column list thing)
expr_list (this is the controversial ROW() one)

Items I can Relucatantly Skip:

alter_table_cmds
reloption_list
hash_partbound
alter_type_cmds
copy_generic_opt_arg_list
opt_without_overlaps
ExclusionConstraintList
part_params
stats_param
generic_option_list
alter_generic_option_list
event_trigger_value_list
def_list
opclass_item_list
opclass_drop_list
privilege_list
parameter_name_list
grant_tole_opt_list
aggr_args_list
aggregate_with_argtypes_list
transform_type_list
operator_with_argtypes_list
operator_def_list
pub_obj_list
pub_all_obj_type_list
transaction_mode_list
drop_option_list (only one option now anyway!)
set_target_list
TableFuncElementList
bxml_attribute_list (code has changed anyway!)
window_definition_list
func_arg_list
type_list
name_list (mostly for permissions)

Cheers,
Greg

--000000000000ccf8a1064db8ec05
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr">Alright, I walked through the entire list of every place t=
he original patch added a comma, and ruthlessly trimmed out the ones that w=
ere not really needed, likely to cause issues, and/or frankly applied on to=
o trivial a piece of SQL to worry about. Here&#39;s my list that people can=
 poke at. Once we settle on which ones survive, I can make a new patch. The=
 focus is mostly on the largest potential uses: lists of select items / col=
umn names.<br><br>Items to keep (i.e. add trailing comma support):<br><br>u=
tility_option_list<br>TableElementList<br>opt_column_list<br>optional_perio=
d_name<br>index_params, index_including_params<br>table_func_column_list<br=
>insert_column_list<br>opt_search_clause (also has some minor refactoring a=
nd adds depth_or_breadth)<br>sortby_list<br>group_by_list<br>vacuum_relatio=
n_list<br>from_list<br>xmltable_column_list<br>xml_namespace_list<br>array_=
expr_list<br>target_list<br>qualified_name_list<br><br><br>Items I am not s=
ure about:<br><br>TriggerFuncArgs (seems sensible, low impact)<br>enum_val_=
list (since it is kind of a column list)<br>any_name_list / role_list (weak=
 +1)<br>type_name_list (trivial but rarely used)<br>set_clause_list (weak +=
1)<br>relation_expr_list (mostly for LOCK)<br>rowsfrom_list (trivial)<br>js=
on_table_column_definition_list (another sort column list thing)<br>expr_li=
st (this is the controversial ROW() one)<br><br>Items I can Relucatantly Sk=
ip:<br><br>alter_table_cmds<br>reloption_list<br>hash_partbound<br>alter_ty=
pe_cmds<br>copy_generic_opt_arg_list<br>opt_without_overlaps<br>ExclusionCo=
nstraintList<br>part_params<br>stats_param<br>generic_option_list<br>alter_=
generic_option_list<br>event_trigger_value_list<br>def_list<br>opclass_item=
_list<br>opclass_drop_list<br>privilege_list<br>parameter_name_list<br>gran=
t_tole_opt_list<br>aggr_args_list<br>aggregate_with_argtypes_list<br>transf=
orm_type_list<br>operator_with_argtypes_list<br>operator_def_list<br>pub_ob=
j_list<br>pub_all_obj_type_list<br>transaction_mode_list<br>drop_option_lis=
t (only one option now anyway!)<br>set_target_list<br>TableFuncElementList<=
br>bxml_attribute_list (code has changed anyway!)<br>window_definition_list=
<br>func_arg_list<br>type_list<br>name_list (mostly for permissions)<br><br=
>Cheers,<div>Greg</div><div><br></div></div>

--000000000000ccf8a1064db8ec05--