NAME Win32::Word::Writer - Create Microsoft Word documents DESCRIPTION Easily create MS Word documents, abstracting away the Word.Application DOM interface and all the required workarounds. The DOM interface is still exposed for doing more fancy stuff. SYNOPSIS use strict; use Win32::Word::Writer; my $oWriter = Win32::Word::Writer->new(); #Adding text and paragraphs with different styles $oWriter->WriteParagraph("Example document", heading => 1); #Heading level 1 $oWriter->WriteParagraph("Usage", style => "Heading 2"); #Style "Heading 2" $oWriter->WriteParagraph("Write sentences to the document using a"); #Normal $oWriter->WriteParagraph("heading level, or Normal if none is specified. "); #\n is new paragraph $oWriter->Write("Add some more text the current paragraph"); $oWriter->NewParagraph(style => "Envelope Return"); #The style must exist $oWriter->Write("Return to sender. "); $oWriter->SetStyle("Envelope Address"); #Change the current style $oWriter->Write("Nope, we changed the style of the entire paragraph"); $oWriter->Write("to a footer style"); #Setting character styles $oWriter->WriteParagraph("Some more normal text. "); $oWriter->SetStyle("Hyperlink"); #A charachter style $oWriter->Write("http://www.DarSerMan.com/Perl/"); $oWriter->ClearCharacterFormatting(); #Clear character style $oWriter->Write(" <-- my "); #Bold/Italics $oWriter->ToggleBold(); #Toggle bold $oWriter->Write("Perl "); $oWriter->SetItalic(1); #Turn on Italic $oWriter->Write("stuff."); $oWriter->ToggleItalic(); #Toggle Italic $oWriter->SetBold(0); #Turn off bold #Bullet point lists $oWriter->ListBegin(); $oWriter->ListItem(); $oWriter->Write("The first bullet item"); $oWriter->ListItem(); $oWriter->Write("The second bullet item"); $oWriter->ListBegin(); #Nested bullet point list $oWriter->ListItem(); $oWriter->Write("The first inner bullet item"); $oWriter->ListItem(); $oWriter->Write("The second inner bullet item"); $oWriter->ListEnd(); $oWriter->ListEnd(); #Do this at regular intervals (say, every couple of 10K of text you add) $oWriter->Checkpoint(); #Tables $oWriter->WriteParagraph("Table example", heading => 1); $oWriter->NewParagraph(); $oWriter->TableBegin(); $oWriter->TableRowBegin(); $oWriter->TableColumnBegin(); $oWriter->SetBold(1); $oWriter->Write("HTML table"); $oWriter->TableColumnBegin(); $oWriter->Write("Win32::Word::Writer"); $oWriter->TableRowBegin(); $oWriter->TableColumnBegin(); $oWriter->SetBold(0); $oWriter->Write(""); $oWriter->TableColumnBegin(); $oWriter->Write("TableBegin()"); $oWriter->TableRowBegin(); $oWriter->TableColumnBegin(); $oWriter->Write(""); $oWriter->TableColumnBegin(); $oWriter->Write("TableRowBegin()"); $oWriter->TableEnd(); #Save the document $oWriter->SaveAs("01example.doc"); CONCEPTS Win32::Word::Writer uses an OLE instance of Word to create Word documents. The documents are constructed in a linear fashion, i.e. you add text to the document and generally don't move around the document a lot. Styles A "style" in Word is a set of properties that can be assigned to a piece of text. There are two types of styles: Paragraph and Character styles. "Normal", and "Heading 1" are example of paragraph styles. When a paragraph gets applied to a piece of text it applies to the entire paragraph, whereas the character style only affects the actual chars. You can see the difference if you open a Word document and look at the available styles. PROPERTIES oWord A Win32::OLE object with a Word Application instance. oDocument A Win32::OLE object with the Application's Document object. Often used shorthand. oSelection A Win32::OLE object with the Application's Selection object. oTable The current Win32::Word::Writer::Table object, if a table is being created, or undef if not. METHODS Note that all methods return 1 or die on errors, unless otherwise stated. new() Create new Word Writer object which can be written to. Return new object, or die on errors. init() Init the object. Called by new. Open($file) Discard the current document and open the Word document in $file. Note that you may want to MoveToEnd() after opening an existing document before adding new text. Note that this object is in an unusable state if the Open fails to load a document. SaveAs($file, %hOpt) Save the document to $file (may be a relative file name). %hOpt is: format => $format -- Save $file as $format (default: Document). Valid values are: Document, DOSText, DOSTextLineBreaks, EncodedText, HTML, RTF, Template, Text, TextLineBreaks, UnicodeText (A common mistake is to inspect the document in another Word instance when re-running a script. The document will be locked by Word and the script can't re-create the file.) Checkpoint() Checkpoint the document, i.e. save it to a temp file. This is necessary to do sometimes because Word seems to keep state until the document is saved, and when using Word automation you tend to exercise the application in ways they haven't tested properly. And after a while you get weird errors, just because Word couldn't deal with all that information. So you should call this after adding, say, 20K of text to the document (this is true for Word 2000, it may be better in later versions). Close() Discard the current document no-questions-asked (i.e. even if it's not saved). Note that this object is in an unusable state until a new document is created or opened. METHODS - ADDING TEXT Write($text) Append $text to the document (using the current style etc). WriteParagraph($text, [heading => $level], [style => $name]) Append $text as a new paragraph of heading $level or style $name. The style overrides heading. The style should be a paragraph style. The default style is "Normal". NewParagraph([heading => $level], [style => $name]) Start a new paragraph of heading $level or with style $name. The style overrides heading. The style should be a paragraph style. The default style is "Normal". SetStyle([$style = "Normal"]) Set the style to $style. If $style is a paragraph style, it will change the style of the current paragraph. If $style is a character style, it will turn on that style. It will be in effect until a new style is set somehow, or until it's cleared with ClearCharacterFormatting(). ClearCharacterFormatting() Clear the characther formatting/set it to default. The paragraph can have a style, and individual characters a separate formatting style. StyleSpec([heading => $level], [style => $name]) Return the final style, given a specification of heading $level or style $name. The style overrides heading. The default style is "Normal". ToggleBold() Toggle the current Bold charachter setting SetBold($enable) Set the Bold status to 1 or 0. Return the new Bold state, or throw OLE exception. ToggleItalic() Toggle the current Italic charachter setting SetItalic($enable) Set the Italic status to 1 or 0. Return the new Italic state, or throw OLE exception. METHODS - BULLET POINT LISTS ListBegin() Begin a new bullet point list. Can be nested to create sub-lists. Use ListItem() to create new bullet points before adding text to the list. ListItem() Start a new bullet point in the list. The first text you Write() after this becomes the new bullet text. You should not WriteParagraph() within a list item. New paragraphs are signals to Word to advance to the next list item, so that will confuse Win32::Word::Writer and/or Word. ListEnd() End an existing bullet point list. If it's the outermost list, go back to normal text. METHODS - TABLES TableBegin() Begin a new table. The table model resembles a HTML table with rows and columns, but you don't have to close columns or rows. Simply start a new one. A row and col must be created with TableRowBegin() and TableColumnBegin() before any text is added. Tables can not be nested. Note that tables are rather fragile so don't expect them to work with very complex layouts, or very wide columns. Prepare for exceptions to be thrown. TableRowBegin() Begin a new row in the current table. Add a column also before adding text to the table. TableColumnBegin() Begin a column in the current table in the current row. Any new text/paragraph added to the document will end up in this table cell until a new row or column is created, or the table is ended. TableEnd() Begin a column in the current table in the current row. Any new text/paragraph added to the document will end up in this table cell until a new row or column is created, or the table is ended. METHODS - MOVEMENT AND SELECTION MoveToEnd() Set the insertion point at the end of the document. SelectAll() Make the selection the entire document. Return 1 on success, else die. METHODS - FIELDS AND TABLES FieldsUpdate() Update the fields in the entire document. Retain the current cursor location. But note this doesn't always work with Table of Contents tables. Return 1 on success, else die. ToCUpdate() Update both entries and page numebers of all the Tables of Contents in the entire document. Retain the current cursor location. Return 1 on success, else die. METHODS - BOOKMARKS BookmarkAdd($name) Add a new bookmark called $name at the current cursor location. Return 1 on success, else die. BookmarkGoto($name) Go to bookmark called $name. The bookmark should exist. Return 1 on success, else die. BookmarkDelete($name) Delete bookmark called $name. The bookmark should exist. Return 1 on success, else die. METHODS - UTILITY MarkDocumentAsSaved() Mark the Word document as "saved". This is in effect until the document is changed again. Being saved e.g. means it can be abandoned without questions. Return 1 on success, else die. GetFileTemp() Return a temporary file name in fileTemp(). DESTROY Release objects including the OLE Word object. KNOWN BUGS Supressing dialog boxes The most serious problem I have with Word is that the documented way of supressing interactive dialog boxes... doesn't work! This is worked around in a few cases (see below), but mostly it's broken. I don't know if this only goes for my Office 2000 Word, but it may affect you too. It's a very bad thing anyhow, since it can cause your program to just freeze, waiting for user interaction. To boot, the dialog boxes are usually displayed below other applications. I blame Bill. OLE errors during global destruction If you are in the middle of a table and something goes wrong, there will be strange OLE warnings during global destruction. I haven't found out why this happens. Layout too complex I have run into this problem where, despite the no-don't- show-dialogs, Word pops up an error dialog below all other windows (so you can't see it, great!). After clicking Ok in this dialog a number of times, the OLE call finally fails properly and dies in the Perl application layer. http://support.microsoft.com/kb/292174 The only way to not run into this problem seems to be to save the document to disk after adding some text. The Checkpoint() method does this for you. Rouge WINWORD.EXE processes Sometimes it seems like Win32::OLE has some problems with closing the Word instance during global destruction. This happens mostly when things die(). TODO Tests for Tables of Contents etc Tests for Bookmarks APPLICATION DOM INFORMATION So what does the Word DOM look like? Actually, the documentation is available when installing Office. Start Word and press Alt-F11 to bring up the VBA window. There is an Object Browser in the toolbar. Select an object, method or property and press F1 to bring up the help. A good way to figure out how to do something is to record a Macro and then bring up the VBA window and inspect the code written by the Macro Recorder. DESIGN ISSUES Software versions This is tested and developed using w2k and Office 2000. Things may be different with other versions. Please let me know. Supressing the "Save as..." dialog box The problem with this is that it doesn't work to follow the manual and advice found on the Net. The usual answer is to set DisplayAlerts to False, or wdAlertsNone. That doesn't work for me. What works is to set the Document.Saved property to False before quitting (the MarkDocumentAsSaved() method). That's why the ActiveX object is Quit from the DESTROY method, and not using the exit handler in CreateObject which is the normal course of action. GOOD IDEAS Keep an eye on the Task Manager When you fiddle around with this program, it's useful to keep the Task Manager window open to keep track of any WINWORD.EXE processes that may be stuck in memory if you e.g. C-Break out of the script (don't do that, Win32::OLE won't have a chance of cleaning up the Word instance it created). Kill abandoned Word processes (but make sure you don't kill any documents you may be editing :) EXTENDING THE MODULE The interface of this module is spotty in an opportunistic way; I have added utility methods as I needed them. If you need to add your own methods, I suggest you simply inject them in this namespace to get your application working and send me a patch. PRIVATE PROPERTIES These are considered implementation details, but you may need to fiddle with them if you extend the module. hasWrittenParagraph Whether the writer has written a paragraph yet. hasWrittenText Whether the writer has written any text or paragraph yet. levelIndent The indentation level for bullet point lists. Default: 0 hasWrittenInIndent Whether the writer has written anything after changing indentation level. rhConst Ref to hash with imported Word constant symbold. styleOld The previous style. fileTemp The name of a temporary file. AUTHOR Johan Lindström, "" BUGS Please report any bugs or feature requests to "bug-win32-word-writer@rt.cpan.org", or through the web interface at . I will be notified, and then you'll automatically be notified of progress on your bug as I make changes. ACKNOWLEDGEMENTS COPYRIGHT & LICENSE Copyright 2005 Johan Lindström, All Rights Reserved. This program is free software; you can redistribute it and/or modify it under the same terms as Perl itself.