Re: Inlining errors...
On Wed, 19 Apr 2000, Jan-Willem Maessen wrote: I agree so wholeheartedly that I just write a (very!) short paper on the subject for ICFP (having discovered to my surprise that no such write-up existed). It describes how to identify such expressions and hoist them out so they don't end up getting inlined. It's still being refereed is thus likely to be revised, but if you're interested in a pre-print take a look (there are no pointers to it from elsewhere at the moment): It's interesting reading. However, it seems to me that it would interact badly with minimal typing derivations (mtd). Mtd is an algorithm for type checking which computes the least general type instead of the most general. This is used by some compilers to guide other optimisations. Some optimisations, like certain representation analysises, do a better job on less polymorphic code and therefore it is desirable to have less general types. Does ghc use mtd or some other heuristics to remove redundant polymorphism? /Josef PS. Sorry for not having any pointers to mtd.
Minimal typing derivations and free theorems...
Josef Sveningsson [EMAIL PROTECTED] writes: It's interesting reading. However, it seems to me that it would interact badly with minimal typing derivations (mtd). Mtd is an algorithm for type checking which computes the least general type instead of the most general. This is used by some compilers to guide other optimisations. My instinct is that any analysis which makes use of free theorems is going to require a most general type. This is because the free theorems are based upon the most general type---and of course the more polymorphism there is, the stronger the free theorems that can be derived in general. That being said, I don't see any difficulty with using both kinds of type information---the information we obtain with the two analyses is different, and both types constrain program behavior in different ways. Mtd is unlikely to be useful for extracting bottoms, but there is no reason bottom extraction should stop us from assigning a more specific minimal type to an error-handling expression. Actually, though, I wonder if mtd are useful at all for expressions known to be bottom---this knowledge is an assertion that we need not represent the value at all, and consequently seems strictly more flexible than any representation we might choose by eliminating polymorphism. The most important observation, really, is that we can still assign a minimal type (like "Int") to a bottom expression, and thus that we will not contaminate representation analysis or other mtd-based optimizations. And, in fact, its bottom-ness means we can assign such a minimal type to each occurrence, so that the sharing of the bottom expression need not introduce spurious polymorphism. -Jan-Willem Maessen
ghc-4.06 installation -- libgmp.so.2?
I've been trying to install the last couple of ghc releases (from various linux binaries). Installation itself goes OK, but ghc dies on the lack of libgmp.so.2. Unfortunately, the only libgmp releases I've seen build only a static library. Am I (no doubt) missing something? Thanks, --ag -- Artie Gold, Austin, TX (finger the cs.utexas.edu account for more info) mailto:[EMAIL PROTECTED] or mailto:[EMAIL PROTECTED] -- A: Look for a lawyer who speaks Aramaic...about trademark infringement.
Bcc: Re: ghc-4.06 installation -- libgmp.so.2?
"Arthur H. Gold" [EMAIL PROTECTED] wrote, I've been trying to install the last couple of ghc releases (from various linux binaries). Installation itself goes OK, but ghc dies on the lack of libgmp.so.2. Unfortunately, the only libgmp releases I've seen build only a static library. Am I (no doubt) missing something? Which distribution are you running? If you use Red Hat, just install the gmp-devel package that is on the binaries CD. Otherwise, a symbolic link from libgmp.so.2.x.y to libgmp.so should also do the job. Manuel
