Malayalam text formating Rendering Problem with LibreOffice Writer 5.3.6 and 5.4.2

While using LibreOffice Writer 5.3.6 and 5.4.2 with Malayalam texts following problems are facing.

  • When applying paragraph Justification trailing space appearing in some lines.
  • When applying paragraph Justification Full Stop/ Unicode character U+002E , Comna/Unicode character U+002C followed by word without space separated from word.
  • Unicode character U+200D containing words breaking to next line in paragraph.

image description

  • No Problem facing with LibreOffice version 5.2.7 (obsolete version)
  • How to solve these problems.

File a bug report to bug tracker. Also, please mark the problematic areas on screenshots, to enable developers see them (they might not know the language and its rules). Please also attach a sample ODT file to your report.

Did you make sure that this problem is not specific to a particular font? I have copied some text from the Malayalam Wikipedia to LO Writer 5.3.3.2 (portable) on MS Windows and do not notice any issues with commas and points. The font is Times New Roman, but LibreOffice obviously gets Malayalam glyphs from some other font. As for U+200D, this seems to be correct behavior, at least the description at codepoint does not tell anything otherwise.

Changed textlayoutengine string value ‘new’ to ‘old’ using Expert Configuration option in LibeOffice; Now problem related with unwanted space is solved.
Please download and open .odt file (LO version 5.3.6) enclosed with https://drive.google.com/uc?export=download&id=0BwfLqao5ypwEQTlKNVJiUVFwZms for understanding word breaking problem related with Unicode character U+200D.

As mentioned above, please use bugtracker to file a bug report and attach the file there. Developers don’t use AskLibO to track bugs.

With 5.4.0.3 on Windows and Meera regular font, I can confirm the first problem. For example, when a line ends in U+0D26U+0D46 there is extra space. Perhaps when the glyphs are drawn in reverse order, the engine fails to determine where to add spacing.

Apache OpenOffice 4.1.3 does not show extra space. LibreOffice introduced a new layout engine starting in 5.3, as explained at https://wiki.documentfoundation.org/ReleaseNotes/5.3#Text_Layout. One of the main goals was to use a single engine across all platforms, rather than Uniscribe only on Windows.

Kartika, the font I normally use for Malayalam, seems to be worse than Meera. Nirmala UI is somewhat cleaner, but Meera looks better. Anyway, changing to a different font does not fix the problem.

I also tried to reproduce the other issues you described but could not. U+002E, U+002C and U+200D seemed to behave correctly. Perhaps you could edit the question and provide example data. Not just a screenshot, but actual Unicode text. However, the answer is probably still the same.

To solve these problems, use a version before 5.3. Or, reportedly in 5.3 but not 5.4 and newer, it is possible to disable the new engine with the SAL_NO_COMMON_LAYOUT environment variable. As mentioned in comments, it may also help to file a bug.

EDIT:

Your example document clearly demonstrates the problem with U+200D. In recent versions of LibreOffice, the word വാല്‍ക്കണ്ണാടി (vāl‍kkaṇṇāṭi – U+200D is the fifth character in this word) can be incorrectly split up by end-of-line wrapping.

…വാല്‍

ക്കണ്ണാടി

The problem does not occur in Apache OpenOffice 4.1.3, so it seems you must use an older version. These are significant problems for anyone who uses Malayalam script, so I recommend that you file a bug report and include this example document.

Changed textlayoutengine string value ‘new’ to ‘old’ using Expert Configuration option in LibeOffice; Now problem related with unwanted space is solved.
Please download and open .odt file (LO version 5.3.6) enclosed with https://drive.google.com/uc?export=download&id=0BwfLqao5ypwEQTlKNVJiUVFwZms for understanding word breaking problem related with Unicode character U+200D.

Good example. See edited answer.