Fix various issues with parsing of Cabal files #459
Merged
Conversation
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Bodigrim
force-pushed
the
fix-cabal-parsing
branch
from
c6333d2 to
0468611
Compare
…vision parseCabalTarball relies on the assumption that the highest version of a package is the last in 01-index.tar.gz. This is not necessarily true in general (one can make a patch release in a long-term support major version), but especially not true since Hackage gained a concept of package revisions. Each package revision is a separate entry in 01-index.tar.gz and it is very common to revise old releases (because, say, they became incompatible with new versions of dependencies). Thus Hoogle ends up with a random, some times terribly outdated version of a package which just happened to be revised recently.
hoogle-*.warn file contains 1000+ warnings caused by non-compliant parsing of build-depends field in Cabal files. E. g., it warns that hinstaller.cabal: Import of non-existant package Cabal>=1.2.3.0 The actual problem is that Hoogle expects a package name to be separated from version bounds by a space, which is not a requirement imposed by Cabal. So Hoogle literally looks for a package named "Cabal>=1.2.3.0", not "Cabal". After this commit the number of warnings decreases to ~100 and they all become genuine. E. g., cabal-install.cabal: Import of non-existant package Cabal-QuickCheck Indeed, Cabal-QuickCheck is not on Hackage.
Now that we depend on Cabal-syntax anyway, why not use a proper type to represent package names?
At the moment Hoogle just reads the license field as is, which leads to suboptimal situation when `hoogle search license:BSD3` and `hoogle search license:BSD-3-Clause` return different (and disjoint!) results. We should do better and normalise licenses as Cabal does. By this point it's easier to migrate entire readCabal to reading fields of GenericPackageDescription instead of rolling out our own parser.
Bodigrim
force-pushed
the
fix-cabal-parsing
branch
from
0468611 to
72cc432
Compare
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
1 participant
Add this suggestion to a batch that can be applied as a single commit.
This suggestion is invalid because no changes were made to the code.
Suggestions cannot be applied while the pull request is closed.
Suggestions cannot be applied while viewing a subset of changes.
Only one suggestion per line can be applied in a batch.
Add this suggestion to a batch that can be applied as a single commit.
Applying suggestions on deleted lines is not supported.
You must change the existing code in this line in order to create a valid suggestion.
Outdated suggestions cannot be applied.
This suggestion has been applied or marked resolved.
Suggestions cannot be applied from pending reviews.
Suggestions cannot be applied on multi-line comments.
Suggestions cannot be applied while the pull request is queued to merge.
Suggestion cannot be applied right now. Please check back later.
This fixes various issues with Cabal files, see individual commit messages for details.