On the matter of GCC compatibility, I've recently noticed a minor issue thanks to having more SELinux systems configured to be "enforcing" out in the wild. I haven't had time to run down the issue and write up a proper bug report, and my asm chops are not good enough to patch it myself.
We use crypto++ linked into a dynamic library that provides higher level PKI functions. We had previously been building it using the default settings in the makefile included in the source distribution, which do not enable generation of position independent code. Linking this into our shared library caused SELinux to get angry at runtime, and we had to change the SELinux context for our library every time we rebuilt. The easiest solution was to rebuild crypto++ with -fPIC. This failed on either x86 or ia64 or both unless assembly was disabled. Since only parts of the library we don't use had this problem, we simply removed them. (I think it was mainly wake... there may have been something else too though.) Sorry for the annoyingly vague bug report. It's been on my list for weeks to write a better one and I just haven't had time to produce that. If you build on RHEL 5.2 or a derivative with -fPIC you should see it right away, though. Some of the asm uses a register gcc wants when generating position independent code. Geoff On Thu, Nov 20, 2008 at 20:03, Wei Dai <[EMAIL PROTECTED]> wrote: > > I didn't have any specific plans, but the near future may be a good time to > do a new release, in order to fix the incompatibility with the latest GCC. > I'll also add GCM mode to the next release, given that it's being added to > various standards. > > I'm not aware of any other major issue that needs to be addressed. If you > have something else you'd like to see in the next version, please post your > suggestions. > > -------------------------------------------------- > From: "Alexei Evdokimov" <[EMAIL PROTECTED]> > Sent: Wednesday, November 19, 2008 8:57 PM > To: "Crypto++ Users" <[EMAIL PROTECTED]> > Subject: Are there plans for new version of cryptopp? If yes, when and what > changes would it include? > >> >> The subject says it all. >> I'd be especially interested in support for the GCM mode (AES-GCM if >> it has to be limited). >> Plus, any other improvements/fixes - whether for performance, >> reliability, or compatibility with other implementations. >> >> Thanks! >> > > > > > --~--~---------~--~----~------------~-------~--~----~ You received this message because you are subscribed to the "Crypto++ Users" Google Group. To unsubscribe, send an email to [EMAIL PROTECTED] More information about Crypto++ and this group is available at http://www.cryptopp.com. -~----------~----~----~----~------~----~------~--~---
