Skip to content

Changes

Page history
Update tying the knot authored by Ben Gamari's avatar Ben Gamari
Hide whitespace changes
Inline Side-by-side
commentary/compiler/tying-the-knot.md
View page @ e1381ea4
......@@ -21,87 +21,7 @@ How this knot tying works is a dark corner of GHC, but hopefully this wiki page
Knot-tying is all intimately tied up with the treatment of `hs-boot` files, so those tickets are listed here.
Use Keyword = `hs-boot` to ensure that a ticket ends up on these lists.
**Open Tickets:**
<table><tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/1012">#1012</a></th>
<td>ghc panic with mutually recursive modules and template haskell</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/8441">#8441</a></th>
<td>Allow family instances in an hs-boot file</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/9450">#9450</a></th>
<td>GHC instantiates Data instances before checking hs-boot files</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/9562">#9562</a></th>
<td>Type families + hs-boot files = unsafeCoerce</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/10333">#10333</a></th>
<td>hs-boot modification doesn&apos;t induce recompilation</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/12034">#12034</a></th>
<td>Template Haskell + hs-boot = Not in scope during type checking, but it passed the renamer</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/12063">#12063</a></th>
<td>Knot-tying failure when type-synonym refers to non-existent data</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/13069">#13069</a></th>
<td>hs-boot files permit default methods in type class (but don&apos;t typecheck them)</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/13180">#13180</a></th>
<td>Confusing error when hs-boot abstract data implemented using synonym</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/13299">#13299</a></th>
<td>Typecheck multiple modules at the same time</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/13322">#13322</a></th>
<td>Pattern synonyms in hs-boot files</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/13347">#13347</a></th>
<td>Abstract classes in hs-boot should not be treated as injective</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/13981">#13981</a></th>
<td>Family instance consistency checks happens too early when hs-boot defined type occurs on LHS</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/14092">#14092</a></th>
<td>hs-boot unfolding visibility not consistent between --make and -c</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/14103">#14103</a></th>
<td>Retypechecking the loop in --make mode is super-linear when there are many .hs-boot modules</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/14382">#14382</a></th>
<td>The &apos;impossible&apos; happened whilst installing gi-gtk via cabal</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/16127">#16127</a></th>
<td>Panic: piResultTys1 in compiler/types/Type.hs:1022:5</td></tr></table>
**Closed Tickets:**
<table><tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/2412">#2412</a></th>
<td>Interaction between type synonyms and .hs-boot causes panic &quot;tcIfaceGlobal (local): not found&quot;</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/4003">#4003</a></th>
<td>tcIfaceGlobal panic building HEAD with 6.12.2</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/7672">#7672</a></th>
<td>boot file entities are sometimes invisible and are not (semantically) unified with corresponding entities in implementing module</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/10083">#10083</a></th>
<td>ghc: panic! (the &apos;impossible&apos; happened)</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/11062">#11062</a></th>
<td>Type families + hs-boot files = panic (type family consistency check too early)</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/12035">#12035</a></th>
<td>hs-boot knot tying insufficient for ghc --make</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/12042">#12042</a></th>
<td>Infinite loop with type synonyms and hs-boot</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/12064">#12064</a></th>
<td>tcIfaceGlobal error with existentially quantified types</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/13140">#13140</a></th>
<td>Handle subtyping relation for roles in Backpack</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/13591">#13591</a></th>
<td>&quot;*** Exception: expectJust showModule&quot; in ghci with hs-boot</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/13710">#13710</a></th>
<td>panic with boot and -jX</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/13803">#13803</a></th>
<td>Panic while forcing the thunk for TyThing IsFile (regression)</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/14075">#14075</a></th>
<td>GHC panic with parallel make</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/14080">#14080</a></th>
<td>GHC panic while forcing the thunk for TyThing IsFile (regression)</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/14396">#14396</a></th>
<td>Hs-boot woes during family instance consistency checks</td></tr>
<tr><th><a href="https://gitlab.haskell.org/ghc/ghc/issues/14531">#14531</a></th>
<td>tcIfaceGlobal (local): not found</td></tr></table>
See the ~"hs-boot" label.
## Practical advice
......@@ -141,7 +61,7 @@ data IfaceRule data CoreRule
Taking `IfaceType` and `Type` as an example, we can see the big difference in a constructor for type constructor application:
```
```haskell
data Type
= ...
| TyConApp TyCon [KindOrType]
......@@ -149,6 +69,7 @@ data Type
data IfaceType
= ...
| IfaceTyConApp IfaceTyCon IfaceTcArgs
data IfaceTyCon
= IfaceTyCon { ifaceTyConName :: IfExtName
, ifaceTyConInfo :: IfaceTyConInfo }
......@@ -168,7 +89,7 @@ In `Type`, the type constructor application contains the full `TyCon` which cont
Consider the following Haskell file:
```
```haskell
data T = MkT S
data S = MkS T
```
......@@ -221,14 +142,16 @@ As we typecheck Haskell source code, we produce `TyCon`s and other type-checking
**An hs file which implements an hs-boot file.** This is the trickiest case of knot-tying during type checking, so let's look at a particular example:
```
```haskell
-- A.hs-boot
module A where
data T
-- B.hs
module B where
import {-# SOURCE #-} A
data S = MkS T
-- A.hs
module A where
import B
......