This is the history file for the last revisions for Mif2Go package 33. Rev 53, February 24, 2009 ---------------------------------------------------------------------- This version adds Frame 9 support. The new m2rbook.dll is REQUIRED for use with Frame 9. Earlier versions will result in failure of Frame 9 to write any MIF files. We've replaced dcl.exe with build 36, which gives the correct error message when there is no .ini in place. We've tested it extensively, but if you get a crash when running *any* conversion, please get the older dcl.exe from dcl34.zip under Components. We've also changed the context-sensitive Help for Mif2Go itself from WinHelp (HLP) to HTML Help (CHM). This requires copying ugmif2go.chm from the Mif2Go directory where you unzip the distro, to \windows\help (or, for some of us, \winnt\help). See the User's Guide for full installation instructions. **For HTML (build 286), Seraphim Larsen found some more issues with the export of conditional text properties to DITA as attributes. Then Aryeh Sanders found a few more, where the condition-bearing elems did not nest properly with typographic elements like and . We have fixed all these issues. K'Chebe Grace used the system variable Chapter Number in text (rather than in an autonumber) and was surprised to see that it didn't work. It always has in autonumbers, but we didn't expect it to be used in other contexts. It works correctly now, as does the Volume Number var. We also added support for the Frame 9 <$sectionnum> and <$subsectionnum> building blocks in autonumbers and in other system variables. K'Chebe also wanted to use the Frame variables in Mif2Go macros. The problem is that Frame variable names may not be compatible with the macro variable naming rules. We solved this for DITA a while back by making a second version of the name, in which: 1. Punctuation other than underscore is removed. 2. Spaces are converted to underscores. 3. If the first character is a digit, it is prefixed with "x". So we've now turned this on for non-DITA outputs too, and hooked it up to the macro variable mechanism. We've also added these predefined macro variables: <$$_volnum> <$$_chapnum> <$$_sectionnum> <$$_subsectionnum> These are references to the corresponding Frame system variable building blocks, so if Frame's "Chapter Number" variable is defined as "<$chapnum>" in Frame, you could use its contents in a Mif2Go macro with either <$$Chapter_Number> or <$$_chapnum>. Becca Heins had problems in Java Help because of an error when we derived the .lst file name; we fixed that bug, and a related one that affected GraphSubdir and a few others. Then she had difficulty with the PixelSpacerImage, where we failed to apply the GraphSubdir path to references to 1p.gif; we do now. Finally, she got a crash during index generation for See/SeeAlso when SortSeeAlsoFirst=No (the default). Brad Currier reported a similar problem at the same time; we fixed both. Yves Barbion found that we were writing a space at the end of text lines that were broken for length inside DITA elements, before the actual line-end. When his XML editor rewrapped lines, it wound up with to spaces where the line breaks had been. We removed the extra space. (If anyone really needed it, let us know!) Siravij Boonsue had both a graphic and a table anchored in the same para, and they started coming out in HTML in reverse order from the one in Frame. That was because of an enhancement in the last release, which sorted both into ones that come before a FrameAfter, and ones that come after it. However, when there is no Frame After, this made the order different. We've corrected it so that the sorting only happens when there really is a Frame After for that para. Bjarne Melin got a DCL failure when running the cover files of a book while generating DITA. It turns out that when files do not produce any topics at all, there are problems writing the empty output file. We now delete the empty file, with a log message: W1: No content in output file, not written Brad Currier was using the drop-down text facility in MS HTML Help, but he found that his drop-down link wasn't working because we were writing an anchor for the link para... *inside* the ... dropdown link. We have fixed this invalid code. Mark Dempsey had a crash at the start of JavaHelp generation, and so did Becca Heins. We tracked it down to some memory-freeing errors that were unlikely to affect anything but JavaHelp, and fixed them. Becca's test case also revealed a problem that bit when a file couldn't be opened for writing 1p.gif (because the specified path was invalid); we fixed that too. Mainly for DITA, we added a new setting which sets the outputclass attribute (or the class attribute, in HTML) for a DITA xref (or HTML ) link element according to the Frame xref or char format: [CSS] ; XrefFormatIsXrefClass = No (default) or Yes (for xrefs, use the Frame ; xref format name as the xref class name; for hyperlinks, use the char ; format name instead. Mainly for DITA @outputclass use.) ; Default is reversed to Yes if UseCSS=Yes, and for DITA. XrefFormatIsXrefClass=No **For RTF (build 292), we added support for the Frame 9 <$sectionnum> and <$subsectionnum> building blocks in autonumbers and in other system variables. Oliver Sterczyk reported that underscores in char formats in Word RTF were not turned off at the end of the format when that format delimits a hyperlink. (Underscores have a special meaning in WinHelp RTF related to links, and the code failed to differentiate WinHelp and Word outputs in that respect at one point.) We fixed that bug. **For MIF (build 206), Seraphim Larsen got a DCL failure when he ran some files to produce DITA. We found a minor logic error (freeing a variable where its content was not allocated) and fixed it. Bjarne Melin also reported this bug. We also added support for the Frame 9 <$sectionnum> and <$subsectionnum> building blocks in autonumbers and in other system variables. **For the plugin (m2rbook build 110), Stephan Will reported troubles when his paths included German characters (with values above U+007F). We confirmed that such characters in filenames also caused problems (different ones), and fixed both of these to support all upper-ASCII characters for which Frame has a representation (nearly all of them). We've also added support for Frame 9, which no longer accepts requests to write Frame 8 MIF files. The Frame 9 files are virtually identical, and that is what we now use by default for Frame 9 output. You can, however, force output of Frame 7 MIF files instead with: [FDK] UseFrame9MIF=No We do not recommend this unless you need to send the MIF files to a person using Frame 7 or earlier; among other problems, it removes the Frame 8/9 Unicode support. Julie Bruce needed to copy graphic from several places to the wrap directory after processing but before compiling and archiving. We added two more settings to do this: [BookFileCommands] ; BookWrapCommand = command line run at end, before automation processing ; FileWrapCommand = command line run at end, before automation processing =======================================================================