Due to unprincipled adding and removing the `-haddock` flag duringcompilation and recompilation checking, we were performing more workthan necessary.We avoid this by compiling everything with `-haddock` by default. Thisis safe nowadays, we have essentially been doing this for many releases,and know this is fine.For the occasion where we actually want to parse without the `-haddock`flag, we keep explicitly disabling it.We enable `-haddock` flag during session loading, since we alreadyperform a number of DynFlags tweaks.This behaviour is dependent on the `OptHaddockParse` opton, which can,currently, only be modified at compile-time.

…onRule

This work is based on these two commits which should either be mergedbefore this is merged, or it should be rebased.

This is related to#4598.This changes the file to module associating logic done during dependencygraph building.Before, each time a module `Foo.Bar` is found, HLS is testing inside allthe import path for the existence of a relevant fiel.. It means that for`i` import paths and `m` modules to locate, `m * n` filesystemoperations are done. Note also that this involves a lot of complexstring concatenation primitive to build the `FilePath`.A module is tested for each `import` for each of the file of theproject. We also test for `boot` files, doubling the number of test.In#4598 we have a project with `1100` modules, in more than 250 importpaths and we count more than `17000` `import` statments, resulting onover 6 millions test for file existences. This project was blocking formore than 3 minutes during HLS startup.This commit changes the way this is computed:- At startup, a `Map ModuleName FilePath` (the real type is a bit more  involved for performance, multiples unit and boot files handling) is  built by scanning all the import paths for files representing the  different modules.- Directory scanning is efficient and if import path only contains  haskell module, this will never do more job that listing the files of  the project.- The lookup is now simplify a `Map` lookup.The performance improvement is as follows:- The number of IO operation is dramatically reduced, from multiples  millions to a few recursive directories listing.- A lot of the boilerplate of converting path had be removed.- TODO: add an RTS stats before / after with number of allocations- On my project, the graph building time is reduced from a few minutes  to 3s.Limitations:- How to rebuild the `Map` if the content of one directory change?- If one directory is filled with millions of files which are not of  interested, performance can be damaged. TODO: add a diagnostic during  this phase so the user can learn about this issue.Code status:- The `lookup` is not fully restored, especially it does not include the  handling of home unit as well as reexport.- The initialisation phase is cached inside a `TVar` stored as a top  level identifier using `unsafePerformIO`. This is to be improved.A note about performanceMost users won't see the benefits of these change, but I think theyapply to everbody:- We are still doing 1 lookup per `import` per module. But the lookup  result is not multiples IO, so it should be faster by a large amount.- Most project only have 1 (or a few) import paths so won't benefit as  dramatically as me from this.TODO for allocations

It contains a fix which dramatically increase performance when doingcradle initialisation.haskell/hie-bios#464

It does not seem to be used and apparently had never been used (after aquick look in the repository diff).Ask m.pickering about this.