PDF documents miss title

[rrafal@…]

I downloaded documentation PDFs for Boost 1.41.0 from SourceForge. Some of the documents had title set correctly, ex. "regex.pdf" title is "Boost.Regex". However, most of the files have title "Untitled".

I'm using Mac OS X and I used Preview->Tools->Inspector to check the title.

This bug causes problems when I try to import the files into Calibre ebook library. I would greatly appreciate a fix for the next release.

Rafal

[Daniel James]

I'm not sure if I'll be able to get this into the new release, but I'll see what I can do.

[John Maddock]

I'll look into this as well - I've never investigated where that meta-data come from exactly - but I think this will require editing the documentation source to ensure they have the necessary metadata in their headers.

[John Maddock]

The difference appears to be whether the documentation begins:

[library foobar

or

[article foobar

The latter seems to be correctly setting the required metadata in the docbook XML, while the former doesn't. Any ideas? I'm guessing it's because "[library" is intended to be part of a larger document where as "[article" isn't?

John.

[Daniel James]

That's what I first thought, but then I noticed that Boost.Unordered has the correct metadata. My second thought was that it was because libraries have their title in the wrong place ('library' is part of boostbook, not docbook, and the boostbook dtd requires that title and libraryinfo are in the opposite order to the normal docbook document types), but that doesn't make a difference either. As far as I can tell, it's the xml parameters that determine this.

[John Maddock]

If it's the xsl params that determine this, then it may be as simple as changing the project requirements in doc/pdf/Jamfile.v2, I'm running a rebuild with this change now, but it may take a while....

John.

[John Maddock]

Still building but...

Adding

project

: requirements <format>pdf:<xsl:param>xep.extensions=1 <format>pdf:<xsl:param>fop.extensions=0 <format>pdf:<xsl:param>fop1.extensions=0 ;

to the doc/pdf jammfile fixes things for array and any, but not accumulators or string_algo... unless something didn't get rebuilt maybe?

Will investigate some more, again... John.

[anonymous]

OK, I can fix this by either:

• remembering to add the necessary xsl:param's to the command line when I build the docs.
• Changing the default in our stylesheet wrappers to use xep rather than fop1 extensions.

I don't particularly like either option at present, I've asked on the Boost.Build list if there's a better way.

John.

[rrafal@…]

Thanks a lot for working on it. I never expected such a quick response. Many thanks! Rafal

[Daniel James]

Reassigning to John as he's doing all the work.

[John Maddock]

Fixed in TRunk by adding a build script to take care of the necessary options. I'll do the same on the release branch when it reopens. 1.42.0 PDF builds should have this fixed too.