2014-07-10 16:56 GMT+02:00 Adam Wulkiewicz <adam.wulkiewicz@gmail.com>:
Hi Patrick,


Patrick J. Lopresti wrote:
I have checked out the latest "modular Boost" from github.

I followed the instructions in lib/geometry/doc/readme.txt carefully.

I am using fop 1.1 installed via "apt-get install fop" on Linux Mint 17.

I get the following error:
<snip>

I know precisely nothing about .fo files. But I do know a little XML,
and if I go to line 24012 offset 1094 in
bin.v2/libs/geometry/goc/gcc-4.8/debug/geometry.fo, I am in the middle
of a table that has three "fo:table-column" tags. If I add a fourth
such tag by hand and re-run "b2 pdf", the error goes away... only to
be replaced by a similar error on line 24531.

Proceeding in this way a few dozen times, I eventually wind up with a
successful run and a decent-looking PDF file. So this looks like a bug
in the .fo file generator (as opposed to the fop tool), although I
admit I am not exactly qualified to tell.

Please let me know if this report is insufficient or if I should send
it somewhere else.

Thanks for the report.
I was able to locate a bug in doxygen_xml2qbk resulting in generation of invalid number of columns in some tables.
The tables were not generated properly for member functions without the return description.
It's now fixed in the develop branch.

Unfortunately there is similar problem related with the reference matrix:
www.boost.org/doc/libs/1_55_0/libs/geometry/doc/html/geometry/indexes/matrix.html

It's generated from quickref.xml somehow, automatically.
It seems that the first occurence of a problem is inside the table describing Core functionalities.
For now I don't have an Idea how this should be fixed. I'll try to find a way.
Of course any help is appreciated.


Ok, this was quite obvious bug.

Both are now fixed in develop:
https://github.com/boostorg/geometry/commit/e04cc53c6da77b5f8b696d021deaae4d0c8d864b
https://github.com/boostorg/geometry/commit/76c394bd282664af4b0493f95ef8fc21f14327b8

So I can build PDF file on my machine.

The pictures in some cases aren't displayed nicely but this is a different story.

Regards,
Adam