Community
    • Login

    Save found files?

    Scheduled Pinned Locked Moved Help wanted · · · – – – · · ·
    26 Posts 5 Posters 3.1k Views
    Loading More Posts
    • Oldest to Newest
    • Newest to Oldest
    • Most Votes
    Reply
    • Reply as topic
    Log in to reply
    This topic has been deleted. Only users with topic management privileges can see it.
    • EkopalypseE
      Ekopalypse @Alan Kilborn
      last edited by

      @Alan-Kilborn

      that is strange, I only see the searched content, regardless if I select all
      or only partial results. The first line and every “filename line” gets stripped off.
      I’m testing this on 7.8.1

      Alan KilbornA 1 Reply Last reply Reply Quote 0
      • Alan KilbornA
        Alan Kilborn @Ekopalypse
        last edited by

        @Ekopalypse said in Save found files?:

        I’m testing this on 7.8.1

        I still use 7.8 day-to-day so my previous results were with that. But…switching to 7.8.1 (64 bits) temporarily I do not see any change; meaning that I don’t see it working the way you do!

        1 Reply Last reply Reply Quote 0
        • Alan KilbornA
          Alan Kilborn
          last edited by Alan Kilborn

          So before this thread was started earlier today, I had some Pythonscript code for manipulating text in the Find-result window (that I never really finished or did anything with):

          import ctypes
          from ctypes.wintypes import BOOL, HWND, LPARAM
          import re
          
          class FindResultPanel(object):
          
              def __init__(self):
          
                  WNDENUMPROC = ctypes.WINFUNCTYPE(BOOL, HWND,LPARAM)
                  FindWindow = ctypes.windll.user32.FindWindowW
                  GetWindowText = ctypes.windll.user32.GetWindowTextW
                  GetWindowTextLength = ctypes.windll.user32.GetWindowTextLengthW
                  SendMessage = ctypes.windll.user32.SendMessageW
                  EnumChildWindows = ctypes.windll.user32.EnumChildWindows
                  GetClassName = ctypes.windll.user32.GetClassNameW
                  curr_class = ctypes.create_unicode_buffer(256)
          
                  tag1 = 0
          
                  find_result_panel_scintilla_handle_list = []
          
                  def foreach_window(hwnd, lParam):
                      tag2 = 1
                      if lParam == tag1:
                          cc = curr_class[:GetClassName(hwnd, curr_class, 256)]
                          if cc == u'#32770':
                              length = GetWindowTextLength(hwnd)
                              if length > 0:
                                  buff = ctypes.create_unicode_buffer(length + 1)
                                  GetWindowText(hwnd, buff, length + 1)
                                  if buff.value == u'Find result':
                                      EnumChildWindows(hwnd, WNDENUMPROC(foreach_window), tag2)
                      elif lParam == tag2:
                          cc = curr_class[:GetClassName(hwnd, curr_class, 256)]
                          length = GetWindowTextLength(hwnd)
                          if length > 0:
                              buff = ctypes.create_unicode_buffer(length + 1)
                              GetWindowText(hwnd, buff, length + 1)
                              if cc == u'Scintilla':
                                  find_result_panel_scintilla_handle_list.append(hwnd)
                                  return False  # stop enumeration
                      return True  # continue enumeration
          
                  EnumChildWindows(FindWindow(u'Notepad++', None), WNDENUMPROC(foreach_window), tag1)
          
                  try:
                      hwnd = find_result_panel_scintilla_handle_list[-1]
                  except IndexError:
                      raise RuntimeError('No Find result window found')
          
                  self.hwnd = hwnd
          
                  # from Scintilla.iface ( https://github.com/notepad-plus-plus/notepad-plus-plus/blob/master/scintilla/include/Scintilla.iface )
                  # Retrieve a pointer value to use as the first argument when calling the function returned by GetDirectFunction.
                  #get int GetDirectPointer=2185(,)
                  self.direct_pointer = SendMessage(self.hwnd, 2185, 0, 0)
          
              def GetText(self):
          
                  # https://www.scintilla.org/ScintillaDoc.html#SCI_GETTEXT
                  # Retrieve all the text in the document.
                  # Returns number of characters retrieved.
                  # Result is NUL-terminated.
                  #fun position GetText=2182(position length, stringresult text)
                  # position -> intptr_t position in a document
                  # stringresult -> pointer to character, NULL-> return size of result
                  length_of_text = ctypes.WinDLL('SciLexer.dll', use_last_error=True).Scintilla_DirectFunction(self.direct_pointer, 2182, 0, 0)
                  text_of_document = ctypes.c_char_p('\0' * (length_of_text + 1))  # allocate buffer to hold text
                  ctypes.WinDLL('SciLexer.dll', use_last_error=True).Scintilla_DirectFunction(self.direct_pointer, 2182, length_of_text, text_of_document)
                  return text_of_document.value
          
              def GetFoldLevel(self, line_number):
          
                  # https://www.scintilla.org/ScintillaDoc.html#SCI_GETFOLDLEVEL
                  # Retrieve the fold level of a line.
                  #get FoldLevel GetFoldLevel=2223(line line,)
                  SC_FOLDLEVELNUMBERMASK = 0x0FFF
                  return ctypes.WinDLL('SciLexer.dll', use_last_error=True).Scintilla_DirectFunction(self.direct_pointer, 2223, line_number, 0) & SC_FOLDLEVELNUMBERMASK
          
              def GetFoldExpanded(self, line_number):
          
                  # https://www.scintilla.org/ScintillaDoc.html#SCI_GETFOLDEXPANDED
                  # Is a header line expanded?
                  #get bool GetFoldExpanded=2230(line line,)
                  return ctypes.WinDLL('SciLexer.dll', use_last_error=True).Scintilla_DirectFunction(self.direct_pointer, 2230, line_number, 0)
          
          try:
          
              find_result_panel = FindResultPanel()
          
          except RuntimeError:
          
              notepad.messageBox('Could not find Find-result window;\r\nMaybe you have not yet run a search that would create it?', '')
          
          else:
          
              find_result_text = find_result_panel.GetText()
          
              print_all = True if 0 else False
              copy_line_info_that_user_can_see = True if 0 else False
              copy_filepaths_that_user_can_see = True if 1 else False
              unique_filepaths_list = []
          
              recent_start_expanded = True
              recent_filepath_expanded = True
              for (line_number, line_contents) in enumerate(find_result_text.splitlines()):
                  if print_all:
                      print(line_number, find_result_panel.GetFoldLevel(line_number), find_result_panel.GetFoldExpanded(line_number), line_contents)
                  else:
                      if line_contents.startswith('Search'):
                          recent_start_expanded = True if find_result_panel.GetFoldExpanded(line_number) else False
                      elif line_contents.startswith(' '):
                          recent_filepath_expanded = True if find_result_panel.GetFoldExpanded(line_number) else False
                      if copy_line_info_that_user_can_see:
                          if line_contents.startswith('Search'):
                              if waiting_for_start:
                                  pass
                      elif copy_filepaths_that_user_can_see:
                          if line_contents.startswith(' '):
                              if recent_start_expanded:
                                  _ = re.sub(r'^\s+(.+?) \(\d+\shits?\)$', r'\1', line_contents)
                                  if _ not in unique_filepaths_list: unique_filepaths_list.append(_)
          
              if copy_filepaths_that_user_can_see:
                  for _ in unique_filepaths_list: print(_)
          

          In this script, we see in the last line that there is something called unique_filepaths_list.

          We can put this to use for the OP’s task by adding a bit more code at the bottom:

              keep_going = True
              while keep_going:
                  destination_dir = notepad.prompt('Enter destination directory:', 'Copy files with hits', '')
                  if destination_dir == None:
                      keep_going = False  # cancel was pressed
                  elif os.path.isdir(destination_dir): break
          
              if keep_going:
                  from shutil import copy
                  for source_file in unique_filepaths_list:
                      try:
                          copy(source_file, destination_dir)
                      except:
                          notepad.messageBox('Problem copying the file:\r\n"{}"\r\nto the folder:\r\n"{}"'.format(source_file, destination_dir), '')
          

          Running the code now (after doing a search that populates the Find-result window with the desired results) will result in the following popup prompt:

          902c3a28-8c7e-490c-9b8a-c09c9bd58b53-image.png

          Putting a valid folder name in the box and pressing OK will complete the desired copy operation.

          1 Reply Last reply Reply Quote 4
          • Alan KilbornA
            Alan Kilborn
            last edited by Alan Kilborn

            I probably also should add that the files copied by the script are going to be the disk copies of the files. Notepad++ does an in-memory search rather than a disk search when it produces its hit list. So if any of the hit files are dirty/modified/red-iconned when the search is conducted, the file contents that actually get copied are going to be different than the modified contents.

            TL;DR: Save all Notepad++ files (so that none have red icons) before running the script – or when doing Eko’s originally suggested solution!

            1 Reply Last reply Reply Quote 2
            • guy038G
              guy038
              last edited by guy038

              Hi, @ekopalypse, @alan-kilborn and All,

              I did a quick test and I confirm that the Alan’s behavior seems to be the default one !

              My configuration :

              Notepad++ v7.8.1   (32-bit)
              Build time : Oct 27 2019 - 22:46:07
              Path : D:\@@\781\notepad++.exe
              Admin mode : OFF
              Local Conf mode : ON
              OS Name : Microsoft Windows XP (32-bit) 
              OS Build : 2600.0
              Plugins : DSpellCheck.dll ExtSettings.dll mimeTools.dll NppConverter.dll NppExport.dll 
              

              • A Ctrl + A action ( or right-click on the Select All option ), followed with an Ctrl + C action and, finally, a Ctrl + V action, in a new tab, do copy all Find result contents. So your regex works fine, extracting the absolute pathnames of all the files involved in the search !

              • A Ctrl + A action ( or right-click on the Select All option ), followed with a right-click on the Copy option and, finally, a Ctrl + V action, in a new tab, only copy the lines, containing the matched string, from all the files scanned, in the Find result window

              May be a plugin issue ?

              Best Regards

              guy038

              Alan KilbornA PeterJonesP 2 Replies Last reply Reply Quote 2
              • Alan KilbornA
                Alan Kilborn @guy038
                last edited by

                @guy038 said :

                May be a plugin issue ?

                This is a good point. @Ekopalypse can you try testing this copy/paste behavior again with a very-clean 7.8.1?

                1 Reply Last reply Reply Quote 0
                • PeterJonesP
                  PeterJones @guy038
                  last edited by

                  @guy038 said in Save found files?:

                  • A Ctrl + A action ( or right-click on the Select All option ), followed with an Ctrl + C action and, finally, a Ctrl + V action, in a new tab, do copy all Find result contents. So your regex works fine, extracting the absolute pathnames of all the files involved in the search !

                  • A Ctrl + A action ( or right-click on the Select All option ), followed with a right-click on the Copy option and, finally, a Ctrl + V action, in a new tab, only copy the lines, containing the matched string, from all the files scanned, in the Find result window

                  I thought I would emphasize this to @Ekopalypse : the behavior is different depending on whether you use keystrokes or menu selection for the copy. I just replicated this on portable 7.8.1-64bit. If you use the Ctrl+C keystroke, it copies the filename; if you use RClick > Copy, it only copies the lines, not the filenames. (I also confirmed that you cannot use File > Copy, because that’s using the active editor window, even when Find result is the foreground pane.) I pasted not only into Notepad++, but into other apps as well, confirming that it’s actually the copy-operation that’s different depending on Ctrl+C vs RClick > Copy.

                  I went back to my 7.7.1-64bit, and it behaves the same way, with Ctrl+C vs RClick > Copy having separate behavior.

                  Alan KilbornA 1 Reply Last reply Reply Quote 2
                  • Alan KilbornA
                    Alan Kilborn @PeterJones
                    last edited by

                    @PeterJones said:

                    (I also confirmed that you cannot use File > Copy, because that’s using the active editor window, even when Find result is the foreground pane.)

                    Presume you meant:

                    (I also confirmed that you cannot use Edit > Copy, because that’s using the active editor window, even when Find result is the foreground pane.)

                    1 Reply Last reply Reply Quote 2
                    • EkopalypseE
                      Ekopalypse
                      last edited by

                      I’m still a little bit confused as it seems to work most of the time like Peter described
                      but sometimes it doesn’t, means I don’t get the filename lines copied into clipboard.
                      Using a fresh portable 7.8.1 x64 version. Need to do further tests to see if I can find
                      a reproducible way.

                      1 Reply Last reply Reply Quote 1
                      • Patricia MayerP
                        Patricia Mayer
                        last edited by Patricia Mayer

                        @Alan-Kilborn FANTASTIC! Thank you so much, it works perfectly. My colleague and I are delighted. :))

                        Thank you other guys as well.

                        Awesome forum.

                        1 Reply Last reply Reply Quote 2
                        • Alan KilbornA
                          Alan Kilborn
                          last edited by

                          My script code has a .getText function on the FindResultPanel object.
                          With it I can, obviously, get the text of the Find-result window.
                          However, I’d like to be able to get the matching text, but I don’t know how to accomplish this.
                          Visually, I can see the hit results in the window as red text on a yellow background.
                          Anyone know how I can pull this information?

                          EkopalypseE 1 Reply Last reply Reply Quote 0
                          • EkopalypseE
                            Ekopalypse @Alan Kilborn
                            last edited by

                            @Alan-Kilborn

                            Why not using the styling information to identify the matched text?

                            Alan KilbornA 2 Replies Last reply Reply Quote 2
                            • Alan KilbornA
                              Alan Kilborn @Ekopalypse
                              last edited by

                              @Ekopalypse said in Save found files?:

                              Why not using the styling information to identify the matched text?

                              Well, TBH when I earlier went searching in the N++ source code for SCE_SEARCHRESULT_*, I saw no occurrences of SCE_SEARCHRESULT_WORD2SEARCH occurring at all, so I was confused as to how it worked and was thinking it might not be possible at all to get this info via PS, even with ctypes usage.

                              But with some experimentation with SCI_GETSTYLEAT it seems that I can recall the info; for example, in my “hit” text I get a style result of 4, which is indeed SCE_SEARCHRESULT_WORD2SEARCH.

                              Is N++ source using hardcoded magic numbers for this instead of the “tags”, or some other mechanism that is strange or at least isn’t clear to me?

                              Hmm, had a thought to search the Scintilla part of N++, where I do see SCE_SEARCHRESULT_WORD2SEARCH used. That must be how it is done.

                              1 Reply Last reply Reply Quote 2
                              • Alan KilbornA
                                Alan Kilborn @Ekopalypse
                                last edited by

                                @Ekopalypse

                                BTW, I’m still confused after looking at N++ source on how N++ identifies the hit text, so that the “search result” lexer knows where it is, so if you can shed any light on that… if you’re sufficiently interested in doing so… :-)

                                EkopalypseE 2 Replies Last reply Reply Quote 1
                                • EkopalypseE
                                  Ekopalypse @Alan Kilborn
                                  last edited by

                                  @Alan-Kilborn

                                  Is N++ source using hardcoded magic numbers for this instead of

                                  Not sure I understand the question correctly.
                                  Every lexer has hard coded style ids, which then get mapped
                                  with a color via stylers.xml. See searchResult.

                                  how N++ identifies the hit text

                                  It uses an internal struct MarkingsStruct which seems to be filled and provided
                                  as a property to the document. Maybe something you can use to your advantage?

                                  1 Reply Last reply Reply Quote 1
                                  • EkopalypseE
                                    Ekopalypse
                                    last edited by

                                    Sorry, Forgot to include the structure reference.

                                    1 Reply Last reply Reply Quote 1
                                    • EkopalypseE
                                      Ekopalypse @Alan Kilborn
                                      last edited by Ekopalypse

                                      @Alan-Kilborn

                                      Made me a bit of a headache because I assumed that the pointer was an integer data type
                                      but the npp code actually said that it was a const char pointer.
                                      In short, here’s a python script that can determine the positions of the matches.

                                      I leave the changes from py3 to py2 to you :-)

                                      class SearchResultMarking(ctypes.Structure):
                                          _fields_ = [('_start', ctypes.c_long),
                                                      ('_end', ctypes.c_long)]
                                      
                                      
                                      class SearchResultMarkings(ctypes.Structure):
                                          _fields_ = [('_length', ctypes.c_long),
                                                      ('_markings', ctypes.POINTER(SearchResultMarking))]
                                      
                                      
                                      # const char *addrMarkingsStruct = (styler.pprops)->Get("@MarkingsStruct");
                                      addrMarkingsStruct = f_editor.getProperty("@MarkingsStruct").encode()
                                      
                                      # SearchResultMarkings* pMarkings = NULL;
                                      pMarkings = ctypes.pointer(SearchResultMarkings())
                                      
                                      # sscanf(addrMarkingsStruct, "%p", (void**)&pMarkings);
                                      ctypes.cdll.msvcrt.sscanf(addrMarkingsStruct, b"%p", ctypes.byref(pMarkings))
                                      
                                      for i in range(pMarkings.contents._length):
                                          print(f'line {i} start:{pMarkings.contents._markings[i]._start} - end:{pMarkings.contents._markings[i]._end}')
                                      
                                      Alan KilbornA 1 Reply Last reply Reply Quote 2
                                      • Alan KilbornA
                                        Alan Kilborn @Ekopalypse
                                        last edited by

                                        @Ekopalypse said in Save found files?:

                                        I leave the changes from py3 to py2 to you

                                        Not a big burden, only the “print” line for that…

                                        But…What’s your magic behind f_editor? Instead of that I would have expected something like what I did in the earlier code I posted in this thread; something much messier, like:

                                        ctypes.WinDLL('SciLexer.dll', use_last_error=True).Scintilla_DirectFunction(self.direct_pointer, 2182, length_of_text, text_of_document)```
                                        
                                        but using a different number than `2182` and different parameters, of course...
                                        EkopalypseE 1 Reply Last reply Reply Quote 1
                                        • EkopalypseE
                                          Ekopalypse @Alan Kilborn
                                          last edited by

                                          @Alan-Kilborn

                                          Not a big burden, only the “print” line for that…

                                          and the encode is not needed as py2 returns bytes
                                          as well as the b before the “%p”.

                                          But…What’s your magic behind f_editor?

                                          Nothing, this is just my object which represents the editor in the find in files window.
                                          As you said, what you would have expected must be done on your side.

                                          1 Reply Last reply Reply Quote 0
                                          • First post
                                            Last post
                                          The Community of users of the Notepad++ text editor.
                                          Powered by NodeBB | Contributors