Martin Schaffstall wrote: > What is the time frame in which we can expect the official put back of > the ksh93 integration into OpenSolaris? The "Roadmap at OpenSolaris" > document (http://opensolaris.org/os/about/roadmap/) lacks any > information nor do any of the archived messages in the > ksh93-integration list. > Can anyone give us an estimation when we can expect this to happen?
That depends on many many factors, including 1. Getting ksh93 integrated into OS/Net's bulid system. That work is mainly done now except the following items - Get SPARC working (which depend on that I can somehow access a Solaris 10 SPARC build machine with Sun Forte 11 installed to generate the headers) - Create spec files for libdll (almost done, just needs to be commited) - Create spec files for libast (needs to be done (this weekend)) Note: - I've excluded 64bit for now since this seems to require updated headers (64bit itself works fine but seems to require a different set of autogenerated headers (ksh93 people are currently fixing that)) - I am not sure which version we should take for OS/Net integration: "ksh93r" is the currnet version we're working with, "ksh93r+" will be the next version (likely including tons of patches for Solaris (like the i18n one)), then "ksh93s-" will follow (being a preview/beta version of "ksh93s" with new features). "ksh93r"/"ksh93r+" are likely more stable - but in theory taking "ksh93s-" may be nice as this would avoid going through PSARC again for the new features (and maintain API stabilty over a longer period (which may be preferred by Sun... :-) )) 2. Request initial review (or better: inofficial review to catch all the dumb mistakes) of the integration and integrate feedback (I have no clue how long this will take) 3. Create the documentation for PSARC (I have no clue how long this will take, April may know that part better) 4. Request official review of the changes to OS/Net (which are shattered over dozends of files and directories in the source tree) 5. Convince PSARC to accept ksh93 as /bin/ksh (which is not trivial because ksh93 is (intentionally (to describe it shortly: Undocumented features and design bugs (such as the dynamic scope of variables (which is a nightmare)) of ksh88 are gone)) not 100% compatible to old ksh88) April: Did I forget anything ? ---- Bye, Roland -- __ . . __ (o.\ \/ /.o) roland.mainz at nrubsig.org \__\/\/__/ MPEG specialist, C&&JAVA&&Sun&&Unix programmer /O /==\ O\ TEL +49 641 7950090 (;O/ \/ \O;)
