Migration guide for multiple home units
Erdi, Gergo
Gergo.Erdi at sc.com
Mon Jun 20 05:15:45 UTC 2022
PUBLIC
I managed to get this working, but I would like some feedback from Matt or others on whether this is the intended way of doing this.
1. The `extendMG` change can be brute-forced just to keep things going, by looking up dependency `ModSummary`s by `ModName` (we don't have full `Module`s at this point yet!):
registerModSummary :: (GhcMonad m) => ModSummary -> m () registerModSummary ms = modifySession \env ->
let mg = hsc_mod_graph env
-- TODO: of course with a bit more housekeeping we can do better than this...
edges = [ NodeKey_Module $ msKey ms' | dep <- deps, ms' <- mgModSummaries mg, ms_mod_name ms' == dep ]
mg' = extendMG mg edges ms
in env{ hsc_mod_graph = mg' }
where
deps = map (unLoc . snd) $ ms_textual_imps ms
2. `registerModule` can be kept mostly as-is, since it only uses `modifyUnitState` to change the active unit:
registerModule :: (GhcMonad m) => ModDetails -> ModIface -> m () registerModule details iface = modifySession $ extendHpt . addModule
where
hmi = HomeModInfo iface details Nothing
mod = mi_module iface
modOrig = ModOrigin (Just True) [] [] True
addModule = modifyUnitState $ \us -> us
{ moduleNameProvidersMap = M.insert (moduleName mod) (M.singleton mod modOrig) $ moduleNameProvidersMap us
}
extendHpt = hscUpdateHUG $ addHomeModInfoToHug hmi
modifyUnitState :: (UnitState -> UnitState) -> HscEnv -> HscEnv modifyUnitState f env = env
{ hsc_unit_env = let ue = hsc_unit_env env in
let units = ue_units ue
units' = f units
in ue_setUnits units' ue
}
3. The tricky part is getting `addUnit` right. For this, based on the implementation of `initUnits`, I came up with the following:
modifyUnitEnv :: (UnitEnv -> UnitEnv) -> HscEnv -> HscEnv modifyUnitEnv f env = env
{ hsc_unit_env = let ue = hsc_unit_env env in f ue
}
addUnit :: DynFlags -> UnitId -> HscEnv -> HscEnv addUnit dflags unitId = modifyUnitEnv $ \ue ->
let dbs = ue_unit_dbs ue
unit_state = ue_units ue
home_unit = ue_homeUnit ue
in flip updateHug ue $ unitEnv_insert unitId $ HomeUnitEnv
{ homeUnitEnv_units = unit_state
, homeUnitEnv_unit_dbs = dbs
, homeUnitEnv_dflags = dflags
, homeUnitEnv_hpt = emptyHomePackageTable
, homeUnitEnv_home_unit = home_unit
}
setCurrentUnit :: UnitId -> HscEnv -> HscEnv setCurrentUnit unitId = modifyUnitEnv $ ue_setActiveUnit unitId
So my questions about this:
1. How does setting the home unit make sense? By doing this, I am effectively setting the home unit to `main` for all units, since that's the initial `ue_homeUnit` of the initial unit environment. Or does it not matter because after `addUnit`, I call `setCurrentUnit` anyway? I've found that I can't use the unit I am just adding as its own home unit, because that then leads to module name resolution problems in `main`: every imported module from `main` is searched for in `main` instead of its correct unit.
2. Speaking of `main`, why is it that when adding units, I have to skip `mainUnitId`, otherwise module resolution breaks again?
3. Unlike the previous version, I am no longer creating and putting `UnitInfo`s anywhere. Where is this going to bite me? Where (if anywhere) should I put `UnitInfo`s with the new setup?
Thanks,
Gergo
-----Original Message-----
From: ÉRDI Gergő <gergo at erdi.hu>
Sent: Sunday, June 19, 2022 12:32 PM
To: Erdi, Gergo <Gergo.Erdi at sc.com>
Cc: GHC Devs <ghc-devs at haskell.org>; Montelatici, Raphael Laurent <Raphael.Montelatici at sc.com>
Subject: [External] Re: Migration guide for multiple home units
On Thu, 16 Jun 2022, Erdi, Gergo via ghc-devs wrote:
> Is there a migration guide for GHC API clients for the new “multiple home units”
> feature?
OK so in concrete terms, please see my attached program which is a heavily cut-down, standalone version of my real program. On commit fd42ab5fa1df847a6b595dfe4b63d9c7eecbf400^ (i.e.
3219610e3ba6cb6a5cd1f4e32e2b4befea5bd384) it compiles and works as expected. On commit fd42ab5fa1df847a6b595dfe4b63d9c7eecbf400 onwards, two problems pop up:
1. `extendMG` has changed and now requires manually specifying outgoing dependency edges. I thought the whole point of `summariseFile` was to collect this information? The reason I need to `extendMG` at that point is to get intra-unit orphan instances working.
2. `modifyUnitState` and its two uses (`addUnit` and `registerModule`) need to be updated to the new API. I think it makes sense that these need changing, since they touch exactly on the issue of which units are being compiled right now. However, I don't know how to update these. Also, I guess `setHomeUnit` should change the `CurrentUnit` instead of the `HomeUnit` now?
Thanks,
Gergo
This email and any attachments are confidential and may also be privileged. If you are not the intended recipient, please delete all copies and notify the sender immediately. You may wish to refer to the incorporation details of Standard Chartered PLC, Standard Chartered Bank and their subsidiaries at https: //www.sc.com/en/our-locations
Where you have a Financial Markets relationship with Standard Chartered PLC, Standard Chartered Bank and their subsidiaries (the "Group"), information on the regulatory standards we adhere to and how it may affect you can be found in our Regulatory Compliance Statement at https: //www.sc.com/rcs/ and Regulatory Compliance Disclosures at http: //www.sc.com/rcs/fm
Insofar as this communication is not sent by the Global Research team and contains any market commentary, the market commentary has been prepared by the sales and/or trading desk of Standard Chartered Bank or its affiliate. It is not and does not constitute research material, independent research, recommendation or financial advice. Any market commentary is for information purpose only and shall not be relied on for any other purpose and is subject to the relevant disclaimers available at https: //www.sc.com/en/regulatory-disclosures/#market-disclaimer.
Insofar as this communication is sent by the Global Research team and contains any research materials prepared by members of the team, the research material is for information purpose only and shall not be relied on for any other purpose, and is subject to the relevant disclaimers available at https: //research.sc.com/research/api/application/static/terms-and-conditions.
Insofar as this e-mail contains the term sheet for a proposed transaction, by responding affirmatively to this e-mail, you agree that you have understood the terms and conditions in the attached term sheet and evaluated the merits and risks of the transaction. We may at times also request you to sign the term sheet to acknowledge the same.
Please visit https: //www.sc.com/en/regulatory-disclosures/dodd-frank/ for important information with respect to derivative products.
More information about the ghc-devs
mailing list