[EMAIL PROTECTED] wrote, On 2007-12-21 21:30: > [snip] I wanted to know the purpose of the macro > FULL_DUPLEX_CAPABLE & how to enable it for my (Perhaps by setting it > to PR_TRUE, whatever.. while building the library...??)
That feature test macro appears in only one place in NSS. http://lxr.mozilla.org/security/source/security/nss/cmd/selfserv/selfserv.c#684 SSL is not inherently a half-duplex protocol. It is capable of doing two-way simultaneous (a.k.a. "full duplex") communication, and NSS's libSSL is capable of supporting such communication (although that is not its default mode of operation). No products are known to use the full duplex feature of libSSL. There is no known demand for that feature. At one time, long ago, NSS's test clients and servers tested libSSL's full duplex capability. But then it was discovered that selfserv's design had some serious performance limitations, and work was done to eliminate those limitations. To hasten the completion of that performance work, the FULL_DUPLEX_CAPABLE macro was put into selfserv.c to temporarily(!) stop building the code whose purpose was to test libSSL's full-duplex capability. When the high speed testing work was completed, the work to re-enable full duplex testing in selfserv was never done. I suspect that the code inside that ifdef no longer builds (if the symbol is defined), or (if it does) that the code is never invoked, or does not work properly. I have filed Bug 410686 to track this issue. _______________________________________________ dev-tech-crypto mailing list dev-tech-crypto@lists.mozilla.org https://lists.mozilla.org/listinfo/dev-tech-crypto
