Upgrade LTS, generic derivations, SQL quasi quoter #50
Conversation
I moved `convertError` from `Database.MySQL.Simpe.QueryResults` into
`Database.MySQL.Simple.Result` to avoid a cyclic dependency between
the `Database.MySQL.Simpe.QueryResults` and
`Database.MySQL.Simpe.QueryResults.Generic` that both make use of said
function.
Note that this change requires GHC Haskell. We may want to hide this
behind a preprocessor flag for that reason.
I don't quite know if it makes sense to derive instances for
`U1` (types with a single constructor and no fields). Although it is
certainly helpful during development to not get a type error when
trying to derive instances for a type that you are in the process of
defining, like:
data User = User
I have also noted a concern with my generic implementation for
`(:*:)`. I'll reproduce it here:
I'm concerned about this implementation since it's biased
towards one side. Meaning that it's expecting the chain of
`(:*:)` to lean to the right. As an example of the problem
consider the following:
> data T = T Int Int Int deriving (Generic)
> from (T 1 2 3)
M1 {unM1 = M1 {unM1 = M1 {unM1 = K1 {unK1 = 1}} :*: (M1 {unM1 = K1 {unK1 = 2}} :*: M1 {unM1 = K1 {unK1 = 3}})}}
If the result in stead had been like this (note the re-bracketing):
M1 {unM1 = (M1 {unM1 = M1 {unM1 = K1 {unK1 = 1}} :*: M1 {unM1 = K1 {unK1 = 2}}}) :*: M1 {unM1 = K1 {unK1 = 3}}}
Then the generic derivation for `T` would fail.
Furthermore, to be able to supply the arity in the call to
`convertError` I think we might need a `Typeable (a :*: b)`
constraint.
Lastly, the existing implementations for `QueryResults` uses
strictness annotations in certain places. I don't know if these are
strictly (pun!) necessary and haven't used them in the generic
implementation.
fredefox
force-pushed
the
fredefox/generic-quasi
branch
from
03d3e56
to
ee3bba3
Compare
|
Thanks - I think this is a good idea. Obviously it needs to be fixed to pass the tests on older LTS versions. |
|
I was wondering if you have any take on my concerns mentioned in the commit message of 2e1619f. |
Goodbye to these lovely GHC extensions :( * TypeApplications * KindSignatures Also turns on -Wall and -Werror in some modules.
fredefox
force-pushed
the
fredefox/generic-quasi
branch
from
735c3d5
to
2766106
Compare
|
EDITED
The "slots in the target type" are now being inferred generically. |
fredefox
force-pushed
the
fredefox/generic-quasi
branch
from
62c33c2
to
75ebd98
Compare
| convertResults err (x:xs) (y:ys) = | ||
| -- I'm concerned about this implementation since it's biased | ||
| -- towards one side. Meaning that it's expecting the chain of | ||
| -- ':*:' to lean to the right. As an example of the problem |
There was a problem hiding this comment.
I'm not clear on how this all fits together, but that's an extremely bad assumption. Once there are more than three fields, the generic representation will be a tree.
There was a problem hiding this comment.
I didn't manage to describe the problem properly in this comment. The issue is not what the generic representation generated by from is. The problem is the generic representation that we generate and then later try and convert back to some type using to. Although the problem may be related since I suspect from and to are inverses.
There was a problem hiding this comment.
I can't really figure out how to give a sensible definition for this. If someone wants to take a stab at it this snippet may come in handy for testing. https://gist.github.com/fredefox/79e8bcae045eada98934df88bdad7783
There was a problem hiding this comment.
I've added a test-case for this.
|
Thanks for the comments @treeowl! I like your type error patch for |
…rity Credit to David Feuer.
fredefox
force-pushed
the
fredefox/generic-quasi
branch
2 times, most recently
from
c55f628
to
8ca2778
Compare
fredefox
force-pushed
the
fredefox/generic-quasi
branch
from
8ca2778
to
bc73b14
Compare
This change incorporate three changes
QueryParamsandQueryResultPlease see the relevant commit messages for more context.