In Stackage Nightly, cc @ocharles :
Building benchmark 'benchmark-logging-effect' for logging-effect-1.3.0..
[1 of 1] Compiling Main ( Benchmark.hs, dist/build/benchmark-logging-effect/benchmark-logging-effect-tmp/Main.o )
Benchmark.hs:20:96: error:
• Couldn't match type ‘prettyprinter-1.2.0.1:Data.Text.Prettyprint.Doc.Internal.Doc
ann3’
with ‘PP.Doc’
NB: ‘PP.Doc’
is defined in ‘Text.PrettyPrint.Leijen.Text’
in package ‘wl-pprint-text-1.1.1.1’
‘prettyprinter-1.2.0.1:Data.Text.Prettyprint.Doc.Internal.Doc’
is defined in ‘Data.Text.Prettyprint.Doc.Internal’
in package ‘prettyprinter-1.2.0.1’
arising from a functional dependency between:
constraint ‘LoggingEffect.MonadLog
(LoggingEffect.WithSeverity PP.Doc)
(LoggingEffect.LoggingT
(LoggingEffect.WithSeverity
(prettyprinter-1.2.0.1:Data.Text.Prettyprint.Doc.Internal.Doc
ann3))
IO)’
arising from a use of ‘loggingEffectLog’
instance ‘LoggingEffect.MonadLog
message (LoggingEffect.LoggingT message m)’
at <no location info>
• In the first argument of ‘LoggingEffect.runLoggingT’, namely
‘loggingEffectLog’
In the first argument of ‘nfIO’, namely
‘(LoggingEffect.runLoggingT
loggingEffectLog loggingEffectStdoutHandler)’
In the second argument of ‘bench’, namely
‘(nfIO
(LoggingEffect.runLoggingT
loggingEffectLog loggingEffectStdoutHandler))’
|
20 | main = defaultMain [ bgroup "log10k" [ bench "logging-effect" (nfIO (LoggingEffect.runLoggingT loggingEffectLog loggingEffectStdoutHandler))
| ^^^^^^^^^^^^^^^^
:
:
:
Not sure if this is a Stackage issue or not?
http://hdiff.luite.com/cgit/logging-effect/diff?id=1.3.0&id2=1.2.6
The prettyprinter was changed, but it wasn't updated in the benchmarks afaics.
Ah yes, I should have filed in the upstream project then.
@ocharles do you need me to do that?
I would love a PR! I will try and prep a fix but it might take me a bit of time (work is crazy right now). Is it possible to just disable benchmarks? Not entirely sure why Stackage builds them.
@ocharles there is no extreme rush - just wanted to make sure it doesn't go forgotten that's all.
(Earlier I mistook it as a potential import ambiguity - that is why I filed it here cautiously by mistake.)
I believe this is released and fixed now. Anything I need to do?
Thanks!