WP: How to change WPs default domain

This WordPress Codex page fully covers the topic.

WP: Redirect back to custom page outside of wp directory after comment

Finally added product reviews for A-Rada.com. Used existing WP installation and it's comment system to show reviews on product page (will cover the topic on showing WP content on custom pages outside of it's default directory in following posts).
The problem rose with the default WP "after comment" redirect. By default after user posts it's valuable review he is redirected to default page view (say a-rada.com/blog/?page_id=32), but what we wanted is to redirect him back to appropriate product page.
With an info from wonderful Wordpress.org Support pages and with the simple code in functions.php the desired behavior was achieved.

add_filter( 'comment_post_redirect', 'my_comment_post_redirect' );
function my_comment_post_redirect( $location ){
 if (preg_match("/page_id=([0-9]*)/", $location, $match)) {
  $pageID = $match[1] + 0;
  $page_title = get_the_title( $pageID ); 

  return "http://" . $_SERVER['HTTP_HOST'] . "/product.php?" . $page_title . "#reviews";
 }
 return $location;
}

As you may see, all pages are bound to concrete products, while all posts are actually blog posts, but as I said that's another story (short one) :)

Batch files: easy way of iterating over arguments

Here's the ease way of iterating over argument list passed to batch files (got it from stackoverlow topic). Used it to invoke ImageMagic's commands that involve their own elaborate interpretation of % operator.

rem im.bat
@echo off
setlocal enableDelayedExpansion
set argCount=0
for %%x in (%*) do set /A argCount+=1
set /a arg=0
for /l %%x in (1, 1, %argCount%) do (
 set /a arg=!arg!+1
 call im_impl.bat %%!arg!
)
endlocal

rem im_impl.bat
rem ImageMagic commands
convert %1_%%d.jpg[1-3] -resize 170x130 -set filename:f %%t_also.jpg %%[filename:f]
convert %1_%%d.jpg[1-4] -resize x109 -crop 124x109+3+0 +append %1_4small.jpg

Now it's possible to run im_impl.bat multiple times with different arguments just calling im.bat:

>im.bat clippy_small chandler_mojito

HTML5: Useful posts on responsive design viewport meta tag particularry

Here will be the bunch of useful posts and tutorials on developing responsive web pages.
Is the viewport meta tag really necessary?
HTML5 Boilerplate Doc

UX idea: Advanced highlighting and formatting for code editors and IDEs

It's 2014 but we still lack intellectual tools that could help us to produce high quality code.

As Steve McConnell mentions in his book 'Code Complete' naming standards are vital for any software project.

It's good for you if you have established one and if you adhere to a good naming scheme, even better if whole of your company does, but everyday developer's tasks include an amount of drilling down through code written by others. In the best case that includes looking through code libraries and legacy systems' code, in the worst case it's looking at the code written by your team member (and swearing through teeth cause he refuses to adhere to naming and formatting standards).

When you touch other's code you are on a foe's territory. Foreign naming standards rarely resemble yours so you first have to understand it and second you have to psychologically admit it (that's not always as easy as it seems).

That's where our overpowered CPUs can become handy. IDEs and code editors could know what semantic element is represented with each construct and so they could highlight it with typeface, font weight and color. An extreme example: ordinal type variables could be highlighted differently from class instances and so on. Such advanced highlighting could greatly help in understanding others' code.

There's some advance in the desired direction (see some themes for Sublime, that make use of advanced on the fly parsing), but there's still much to implement!

Another idea is to use CPU power to implement online formatting of code constructs according to your defined rules. This could be done for properly structured and compiled (or checked if we deal with interpreted language) pieces of code, so every team could look at the library code through their own "formatting lens". It seems not so convenient from original authors' point of view, as they will get bug reports with unfamiliar formatting, but they could use their own lens to reformat code back to their likes.

That's not new ideas, they are old ones (see Jeff Raskin's 'Interface', for example). Still such intellectual features are implemented at such a slow pace, that it's worth repeating ideas just not to forget them.

PL/SQL: Generic function for converting arbitrary cursor to clob (delimited text). Take two

Addressing large dataset issue that was mentioned for previously developed cursor to clob conversion functions (see previous post) a new version of a function that converts any query into delimited text was born.
Not so elegant nevertheless it handles nulls (thanks to this StackOverflow topic) and large datasets (that still needs approval).
If you fill able, you can try to convert this procedure into highly optimized pipelined function making use of bulk operations, though I failed due to some bug producing ORA-24374 while working with xml and multisets (maybe I'll cover this bug in the following posts).
So by now the following procedure just inserts clob values into given table

Here are prerequisites:

/* ty_int_to_VCHAR2_row ty_int_to_VCHAR2_tbl type declarations ...*/
create or replace
type TY_INT_TO_VCHAR2_ROW force as object (
 int ingeger
 , vchar varchar2( 4000 char )
 , CONSTRUCTOR FUNCTION TY_INT_TO_VCHAR2_ROW( SELF IN OUT NOCOPY TY_INT_TO_VCHAR2_ROW, a_int in pls_integer, a_vchar in varchar2 ) 
  RETURN SELF AS RESULT
);
/

create or replace
type body TY_INT_TO_VCHAR2_ROW as

constructor function TY_INT_TO_VCHAR2_ROW ( SELF IN OUT NOCOPY TY_INT_TO_VCHAR2_ROW , a_int in pls_integer, a_vchar in varchar2 ) 
 return self as result
as
begin
 int := a_int;
 vchar := a_vchar;
   return;
end TY_INT_TO_VCHAR2_ROW ;

end;
/

create or replace type TY_INT_TO_VCHAR2_TBL force as table of TY_INT_TO_VCHAR2_ROW;
/

create global temporary table tmp_pk_utils_xml_result_set( 
  row_num integer, name_ varchar2( 30 char ), value_ varchar2( 4000 char )
) on commit preserve rows;
/

And here's the code:

procedure query_to_flat( 
  a_query in clob, a_delimiter in varchar2, a_line_break in varchar2, a_target_table in varchar2 
)
as
  v_prev_rn pls_integer := 0;
  v_clob clob;
  v_cur_desc sys_refcursor;
  v_desc_tab dbms_sql.desc_tab;
  v_count number;
  v_cur_handle NUMBER;
  
  v_cur sys_refcursor;
  
  v_column_names_tbl TY_INT_TO_VCHAR2_TBL := TY_INT_TO_VCHAR2_TBL();
  v_truncate_result pls_integer;
  v_truncate_message varchar2( 32767 char );
begin
$if ( pk_utils_cc.dbg = 1 ) $then
  dbms_output.enable( null );
$end  
  -- read column description into v_desc_tab
  open v_cur_desc for 'select * from ( ' || a_query || ' ) where 1=0';
  
  v_cur_handle := dbms_sql.to_cursor_number( v_cur_desc );
  -- we'll get all columns and their names
  dbms_sql.describe_columns( v_cur_handle, v_count, v_desc_tab );
  
  if v_cur_desc%ISOPEN then
    close v_cur_desc;
  end if;
-- move column descripition data into sql-level table type, so we can use it in queries
  v_column_names_tbl.extend( v_desc_tab.count );
  for v_column_idx in 1 .. v_desc_tab.count loop
    v_column_names_tbl( v_column_idx ) := 
      ty_int_to_vchar2_row( v_column_idx, v_desc_tab( v_column_idx ).col_name )
    ;
  end loop;
  
  open v_cur for a_query;
  
  -- just a truncate that retries if resource is busy, not necessary here, but it's an attitude
  pk_safe_ddl.safe_truncate( 'MY_SCHEMA', 'TMP_PK_UTILS_XML_RESULT_SET' );
  
  -- insert resultset into temporary table, here we'll got only not null columns
  insert into tmp_pk_utils_xml_result_set( row_num, name_, value_ )
  SELECT t.rn /*+ NO_XML_QUERY_REWRITE */
    , t2.COLUMN_VALUE.getrootelement() NAME
    , EXTRACTVALUE ( t2.COLUMN_VALUE, 'node()' ) VALUE
  FROM ( select /*+ NO_XML_QUERY_REWRITE */ rownum rn, column_value from TABLE ( XMLSEQUENCE ( v_cur ) ) ) t
    , TABLE ( XMLSEQUENCE ( EXTRACT ( t.COLUMN_VALUE, '/ROW/node()' ) ) ) t2
  ;
  -- first cross join to get all column names for every row in resulting dataset
  -- then join with temporary table that stores actual values
  for cur_val in (
    select a.row_num
      , a.name_
      , q.value_
    from ( 
      select r.row_num
        , c.int as col_idx
        , c.vchar as name_
      from ( 
        select row_num
        from tmp_pk_utils_xml_result_set
        group by row_num
      ) r
      cross join table( cast( v_column_names_tbl as TY_INT_TO_VCHAR2_TBL ) ) c
    ) a
    left outer join ( 
      select row_num
        , name_
        , value_
      from tmp_pk_utils_xml_result_set
      where name_ is not null
    ) q
      on a.row_num = q.row_num
      and a.name_ = q.name_
    order by a.row_num
      , a.col_idx
  ) loop
    -- some kind of listagg functionality but for potentially long lines (clobs)
    if ( 0 = v_prev_rn ) then
      v_clob := cur_val.value_; 
    else
      if ( v_prev_rn != cur_val.row_num ) then
        v_clob := v_clob || a_line_break;
        execute immediate 'insert into ' || a_target_table || ' values( :v_clob )'
        using v_clob
        ;
$if ( pk_utils_cc.dbg = 1 ) $then
        dbms_output.put_line( v_clob );
$end
        v_clob := cur_val.value_;
      else
        v_clob := v_clob || a_delimiter || cur_val.value_;
      end if;
    end if;
    v_prev_rn := cur_val.row_num;
  END LOOP;
  
  execute immediate 'insert into ' || a_target_table || ' values( :v_clob )'
  using v_clob
  ;
$if ( pk_utils_cc.dbg = 1 ) $then
  dbms_output.put_line( v_clob );
$end
  if v_cur%ISOPEN then
    close v_cur;
  end if;
  commit;
end;

Perl: perserve case while replacing

According to this stackoverflow topic it's not quite easy to make a replacement preserving case with simple pattern unless we have perl6's 'ii' modifier, but for practical purposes it mentions great solution like the following one:

$ perl -pi -e 's/(S|s)licing/{$1 eq 'S' ? 'Excel' : 'excel'}/ge' *.pas