Hi, Am Freitag, den 03.06.2011, 15:40 +0200 schrieb Tomáš Janoušek: > On Fri, Jun 03, 2011 at 12:02:30AM +0200, Joachim Breitner wrote: > > here on my machine and with ghc-7.0.3, things are reasonably fast. Can > > you check whether it is still too slow for your, or whether maybe things > > have improved with the new ghc version? > > Things are very fast indeed, because the documentation for type-level isn't > indexed. The reason for this is that: > > $ dpkg -L libghc-type-level-dev | grep 'conf$' > /var/lib/ghc-7.0.3/package.conf.d/type-level-0.2.4.conf > $ grep haddock /var/lib/ghc-7.0.3/package.conf.d/type-level-0.2.4.conf > haddock-interfaces: > /usr/lib/ghc-7.0.3/haddock/type-level-0.2.4/type-level.haddock > haddock-html: /usr/share/doc/libghc-type-level-doc/html/ > $ dpkg -L libghc-type-level-doc | grep 'haddock$' > /usr/lib/ghc-7.0.2/haddock > /usr/lib/ghc-7.0.2/haddock/type-level-0.2.4/type-level.haddock > > Notice that the files that should be in /usr/lib/ghc-7.0.3 are in > /usr/lib/ghc-7.0.2.
ah, ok. Yes, that is a regression introduced by the fact that we try to re-use .haddock files across compatible ghc versions. The next ghc package will put the files in a version-agnostic directory. > If I modify /usr/lib/ghc-doc/gen_contents_index to replace the path for > type-level, then it gets slow again (24 seconds vs. 2 seconds if type-level is > ignored). Ok, thanks for the check. Not sure what to do about it, though. Maybe someone can profile the indexing and look for some low-hanging optimization fruit. Tagging this as "help" to encourage that. Greetings, Joachim -- Joachim "nomeata" Breitner Debian Developer nome...@debian.org | ICQ# 74513189 | GPG-Keyid: 4743206C JID: nome...@joachim-breitner.de | http://people.debian.org/~nomeata
signature.asc
Description: This is a digitally signed message part
