-
jbkgwr: while you're in there, it might be nice to get rid of all the #ifdef soup for K&R compilers
-
jbkI think we can safely require a compiler no older than 37 years at this point :)
-
jbk(it won't get rid of it entirely, but would at least reduce the complexity a bit)
-
tsoome+1
-
tsoomethats really good idea:)
-
jbkthere's already so many layers of #ifdefs in some of those places that I find it very hard to follow what's actually going on
-
gwrjbk: while I'm in where? which #ifdef soup? There are so many choices:)
-
jbkthe xdr.h stuff
-
gwrAh.
-
jbkand yes, we have a plethora of choices there :)
-
gwrWhich #idefs in there do you think we could do without?
-
gwrI only see the old K&R declarations. Everything else looks like it's needed.
-
jbkthat's what I mean.. the old K&R stuff...
-
gwrSo __STDC__. Ok.
-
jbkyeah... (and I just meant if you felt like it while messing with xdr.h)
-
jbkthere are other files with similar things I'm sure, but that could be done separately
-
gwrOh, we have a LOT of __STDC__ stuff in uts/common/rpc. I think I'll leave that for a project to do "just that".
-
gwrIs that everwhere in our headers?
-
gwrLooks like mostly there.
-
richlowegwr: my only comment -- and it took me ages just now to try to find and remember -- is places like cpuvar.h seem to imply they use __STDC__ to hide from genassym
-
richlowe(which I'm not even sure works?!)
-
gwrThanks. I guess we'll find out when we get around to removing not __STDC__ stuff.
-
alancbut you'll break building with the bundled K&R-only SunOS 4 compiler that was barely only useful for rebuilding the kernel!
-
alanc(...and not a single tear will be shed)
-
LXGHTNXNG:D