Hi all, I'm using PdfPTables to build the PDF output to mimic on-screen (i.e. HTML) tabular report output but with much better printed rendering. Our style guide calls for different cell background and border colors in the table header and the table body. In addition, the table header borders are strictly interior to the header. In other words, the outside edge of the header area is the same color as the background of those cells. I started out doing what I thought was relatively obvious. I set a border color for the entire cell (using PdfPCell.setBorderColor) and then used different border colors to override the left, right, top, and/or bottom depending on the position of the cell in the header. This didn't quite come out as I'd planned. All the border colors were the color set using setBorderColor. So, I started looking at the code and quickly noticed that if I had variable width borders, or called setUseVariableWidthBorders(true), then completely different border rendering code was invoked. It's also quite clear that there are paths thru the code that will either not draw the border color (if borderColor is null but hasBorder(Rectangle.BOX) is true) or will not use the side-specific border colors (if hasBorder(Rectangle.BOX) is false). If the variable width border mode is in effect, there are several differences to how the borders are drawn. The colors are taken from each of the sides as are the widths. The code also adjusts the line caps and joins. Further, it seems to interpret the border width value differently. In some cases it seems to give double the width of the border that would have been drawn if variable width border mode is off.
I'm curious about what is the best path forward. I did an experiment that forced hasBorder(Rectangle.BOX) to be false (I disabled the bottom side border in the table header) and added code to each left, right, top, and bottom portions of the else block to get the side-specific border color. This works fine for me. But, while my table is clearly different than others have used for iText, I can imagine that the variable width behaviors (specifically, the cap and join calculations) would be better if more color variation were used in the borders. I have code I could submit if this outcome was thought to be appropriate. I'm about to have a much closer look at PdfContentByte.variableRectangle to see if I can figure out a way to make it not double the border widths where I don't want that behavior. I'm a bit concerned that with variable width borders, I'll need to specially handle all outside edge cells (not just the table header outside edge cells) to get the drawn width of the border to be visually the same on the outside edges as it is in the interior of the table body. But maybe that's unavoidable. Thoughts anyone? Have I been clear or should I provide some additional details? I don't have a small reproduction program but can build one if needed. Thanks, Rick Berman ------------------------------------------------------- This SF.net email is sponsored by: Splunk Inc. Do you grep through log files for problems? Stop! Download the new AJAX search engine that makes searching your log files as easy as surfing the web. DOWNLOAD SPLUNK! http://ads.osdn.com/?ad_id=7637&alloc_id=16865&op=click _______________________________________________ iText-questions mailing list iText-questions@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/itext-questions
