Hi folks, Release candidate distributions for 6.2.2 are now available. If you have the time, we'd really appreciate it if you could install and/or build one of these distributions and give us feedback. ghc-6.2.20040914 and later are 6.2.2 release candidates Please mention the exact version number, including the date, when submitting a bug report. Source distributions and Linux (RH 9) binary dists are here: http://www.haskell.org/ghc/dist/stable/dist/ and there's a Windows installer here: http://cvs.haskell.org/ephem/ghc-6-2-2.msi The actual release will probably be in a couple of weeks' time (next week is ICFP). Thanks! Simon
simonmar:
Hi folks,
Release candidate distributions for 6.2.2 are now available. If you have the time, we'd really appreciate it if you could install and/or build one of these distributions and give us feedback.
ghc-6.2.20040914 and later are 6.2.2 release candidates
Please mention the exact version number, including the date, when submitting a bug report.
Source distributions and Linux (RH 9) binary dists are here:
And OpenBSD binary dists are at: ftp://ftp.cse.unsw.edu.au/pub/users/dons/ghc/stable/dist/ -- Don
Hello, I thought we were going to get a new FFI function soon.. http://www.haskell.org/pipermail/ffi/2004-March/001740.html AFAICS it isn't in the libs for this release :-( Has this got lost or forgotten about? Regards -- Adrian Hey
I have some trouble building ghc from CVS HEAD. I'm not sure whether this applies to the release candidate version, too, but I figured I'd better report it anyway, just in case. 1) The build fails: | ../../ghc/compiler/ghc-inplace -M -optdep-f -optdep.depend | -optdep-s -optdepp -osuf o -H16m -O -optc-fno-pic | -optc-fno-stack-protector -fglasgow-exts -cpp -Iinclude | -"#include" HsBase.h -funbox-strict-fields -package-name | base -O -Rghc-timing Control/Arrow.hs Control/Concurrent.hs | Control/Concurrent/Chan.hs [...] | Data/FiniteMap.hs:111: token "@" is not valid in preprocessor expressions | Data/FiniteMap.hs:314: token "@" is not valid in preprocessor expressions | Data/FiniteMap.hs:420: token "@" is not valid in preprocessor expressions | Data/FiniteMap.hs:584: token "@" is not valid in preprocessor expressions | Data/FiniteMap.hs:733: token "@" is not valid in preprocessor expressions | Data/FiniteMap.hs:746: token "@" is not valid in preprocessor expressions | Data/FiniteMap.hs:807: token "@" is not valid in preprocessor expressions | <<ghc: 11638648 bytes, 6 GCs, 71976/71976 avg/max bytes | residency (1 samples), 25M in use, 0.02 INIT (0.00 | elapsed), 0.08 MUT (1.43 elapsed), 0.03 GC (0.04 elapsed) | :ghc>> | make[2]: *** [depend] Error 1 | make[1]: *** [boot] Error 1 | make[1]: Leaving directory `/usr/local/src/ghc-current/libraries' 2) "make distclean" fails: | $ make distclean | rm -f -rf autom4te.cache | rm -f *.CKP *.ln *.BAK *.bak [...] | ------------------------------------------------------------------------ | ===fptools== Recursively making `distclean' in glafp-utils happy alex haddock ghc libraries hslibs docs ... | PWD = /usr/local/src/ghc-current | ------------------------------------------------------------------------ | ------------------------------------------------------------------------ | ==fptools== make distclean -r; | in /usr/local/src/ghc-current/glafp-utils | ------------------------------------------------------------------------ | You haven't run ./../configure yet. | make[1]: *** [../config.status] Error 1 | make: *** [distclean] Error 1 3) In order to build GHC's user manual and the documentation for the older libraries, I use the following bit of shell scripting before I run configure: | DOCDIRS="alex ghc haddock happy hslibs libraries" | | for n in ${DOCDIRS}; do | echo >$n/mk/build.mk "SGMLDocWays := html" | done | | echo >mk/build.mk "SGMLDocWays := html" But apparently, this has stopped working! A "make install-docs" will neither install nor even build the documentation. (I have had this problem for a few weeks, but never really came around to investigate it.) Peter
Peter Simons wrote:
I have some trouble building ghc from CVS HEAD. I'm not sure whether this applies to the release candidate version, too,
Nope...
but I figured I'd better report it anyway, just in case.
Fine, we like to hear about bugs... :-}
1) The build fails:
My fault, already fixed.
2) "make distclean" fails:
I'll have a look into it.
3) In order to build GHC's user manual and the documentation for the older libraries, I use the following bit of shell scripting before I run configure:
| DOCDIRS="alex ghc haddock happy hslibs libraries" | | for n in ${DOCDIRS}; do | echo >$n/mk/build.mk "SGMLDocWays := html" | done | | echo >mk/build.mk "SGMLDocWays := html"
But apparently, this has stopped working! [...]
Yes, SGML docs are dead, long live XML. :-) To e.g. recursively build all HTML documentation simply use make html or to recursively build and install all documentation in HTML and PostScript: make XMLDocWays="html ps" install-docs This reminds me of the fact that I should update the documentation in the HEAD a little bit... Furthermore, with our current build system there is no need for a build.mk anymore IMHO. Cheers, S.
Sven Panne writes:
1) The build fails:
My fault, already fixed.
Great. Just built the compiler from the scratch; works nicely now.
2) "make distclean" fails:
I'll have a look into it.
Also works correctly now.
make XMLDocWays="html ps" install-docs
OK. :-) I changed SGMLDocWays to XMLDocWays in the build.mk files and now everything works as expected again.
Furthermore, with our current build system there is no need for a build.mk anymore IMHO.
I think it's nice to have a way to configure the build "permamently" so that I can just forget about it and say "make install", "make install-docs", etc., with all the proper flags being passed automatically. It's hard to remember which packages needs what additional settings, so I am quite fond of the build.mk files. Thanks for your help, Sven. Peter
participants (5)
-
Adrian Hey -
dons＠cse.unsw.edu.au -
Peter Simons -
Simon Marlow -
Sven Panne