News Archive 8

The PBC library now has a logo, a result of an attempt to liven up these otherwise text-only pages.

I had thought that releasing version 0.4.0 and two new libraries during the Thanksgiving holiday would give me ample time to fix a few things and release the successor before too many people noticed. After all, who does any work during this time?

I was wrong. I have already received two emails pointing out a number of issues, mainly to do with documentation and the organization of files. This release should address these problems.

I removed the gitweb link. I belatedly realized that a web spider could harvest email addresses buried in PBC code and documentation files via the interface. (How do other projects deal with this? Perhaps there are options I can enable to mangle email addresses?) I also removed all email addresses from the AUTHORS file for good measure. They probably should not have been there in the first place.

I wrote a simple script that automatically chunks five news items per page and generates navigation links. This is much better than the manual editing I was doing before. The oldest items appear their own page and not here.

I’m trying to hype up PBC. The "About" page now looks more like a sales pitch, and I started a list of projects that use the PBC library. If you’re using PBC for something please let me know!

See the mailing list press release or the NEWS file for details. One big change is that the signature and broadcast encryption code have been spun off into new libraries. Together with new debugging features and file renaming (and the usual slew of minor fixes and cleanups), these constitute substantial modifications which justifies incrementing the minor version number.

The new libraries have their own webpages:

PBC_sig: http://crypto.stanford.edu/pbc/sig/

PBC_bce: http://crypto.stanford.edu/pbc/bce/

but share the same mailing list. They are in a highly immature state.

I’m following the cumbersome but informative satellite library naming convention of the SDL library. (Some SDL-based libraries have names such as SDL_Image or SDL_net.)

The roles of the three different places to get news on the library are becoming clearer. I changed the NEWS file to look more like the one in GMP and less like the ChangeLog. The latter is now the only place containing a play-by-play analysis.

One drawback of this new regimen is that I only started maintaining the ChangeLog from around version 0.3.0. Before that, details of changes can only be found an old verson of the NEWS file, though I have posted a webpage summarizing early PBC history.

I now use Valgrind to find memory leaks, and I have jettisoned my own clumsy kludgy leak detection code.

Various bugfixes. The most serious one was reported by Paul Miller: element_from_hash() was not returning the same point each time.