|
Hello ! I have some remarks on the WMF transcoder
package.
I noticed that the WMF transcoding
process don't handle the charset in which the fonts are defined in the WMF
file, the result being that, for example, greek characters are not properly
handled : When some text is encountered (META_TEXTOUT), the WMFRecordStore
uses the standard charset to decode the incoming bytes. It should use the
charset ID encountered for the last font (META_CREATEFONTINDIRECT), and decode the bytes
accordingly.
Another problem I encountered is that the decoding
of WMF text (META_TEXTOUT) seems to depend on the fact that there
is exactly as many bytes as there is characters in the text. In fact, WMF files
code strings in U16 cells, each cell storing two characters, so if there is
an odd number of characters, it is necessary to read the last remaining half of
U16, or the reading of the file will stop abruptly after
that...
|
