Sam Derbyshire retweeted
We are delighted to announce that Well-Typed is receiving funding to work on Cabal as part of the #ContributeBackChallenges by @sovtechfund! For more details on how we plan to improve Haskell tooling please read our blog: well-typed.com/blog/2023/10/…
Sam Derbyshire retweeted
Find out what the GHC team at Well-Typed have been working on recently!
well-typed.com/blog/2023/10/…
Sam Derbyshire retweeted
Check out what the Well-Typed GHC team have been up to over the last couple of months: well-typed.com/blog/2023/08/…
Sam Derbyshire retweeted
Read about how we're making Haskell builds faster! Some great work by @mpickering_ and @samderbyshire thanks to support from @HasuraHQ. well-typed.com/blog/2023/08/…
Sam Derbyshire retweeted
New blog post: GHC activities report for April-May 2023. Check out what the Well-Typed GHC developers have been up to recently! well-typed.com/blog/2023/06/…
Sam Derbyshire retweeted
New blog post by @mpickering_: Interface Files with Core Definitions. Speed up Template Haskell compilation times by up to 40%! well-typed.com/blog/2023/02/…
Sam Derbyshire retweeted
New blog post: GHC activites report for October–November 2022 well-typed.com/blog/2022/12/…
Sam Derbyshire retweeted
computer advice! you can “paste without formatting” in most apps with the handy keyboard shortcut combo:
Ctrl+V
(Deep sigh)
Ctrl+Z
Ctrl+Shift+V
Sam Derbyshire retweeted
sending my grad students an email indicating we are about to “go hardcore” as a research group — all results must work in characteristic 2
Sam Derbyshire retweeted
I'm not sure I look forward to the days of “oh, this proof worked with version 4.6.3-beta4 of étale cohomology but it broke when we upgraded to 4.6.3-final”. 😅
Sam Derbyshire retweeted
Add diagnostic codes
ift.tt/ZyUdv3V
Sam Derbyshire retweeted
GHC removing its Make build system haskell.org/ghc/blog/2022080…

ALT The GHC build system includes all of the above. To illustrate the consequences, here is one rule from GHC’s current build system, which uses Make: $1/$2/build/%.$$($3_osuf) : \ $1/$4/%.hs $$(LAX_DEPS_FOLLOW) \ $$$$($1_$2_HC_DEP) $$($1_$2_PKGDATA_DEP) $$(call cmd,$1_$2_HC) $$($1_$2_$3_ALL_HC_OPTS) \ -c $$< -o $$@ \ $$(if $$(findstring YES,$$($1_$2_DYNAMIC_TOO)), \ -dyno $$(addsuffix .$$(dyn_osuf),$$(basename $$@)) ) $$(call ohi-sanity-check,$1,$2,$3,$1/$2/build/$$∗) Yes, there are four dollar signs in a row! If it wasn’t so tragic, im- penetrable code like this would be hilarious. This kind of thing goes on for thousands of lines. The result is a nighmare to understand, maintain and modify.
Sam Derbyshire retweeted
*Mitsuba 3* is now available! It's a major redesign of the lab's infrastructure for differentiable rendering building on the Dr.Jit just-in-time compiler announced yesterday. Full video link: tinyurl.com/mitsuba-tut 1/10
Sam Derbyshire retweeted
New blog post: GHC activities report: April–May 2022 well-typed.com/blog/2022/06/…
Sam Derbyshire retweeted
GHC 9.2.2 is now available! There are many great fixes in this release; users of GHC 9.2.1 are encouraged to upgrade quickly not only for these fixes but also some great compiler performance improvements.
discourse.haskell.org/t/anno…
Sam Derbyshire retweeted
Kill derived constraints
ift.tt/aAL5qkT
Sam Derbyshire retweeted
Here's the final post in our series on writing type-checking plugins for GHC! tweag.io/blog/2022-02-17-tcp…
Sam Derbyshire retweeted
We are told that Haskell infers the most general type for functions. Except that GHC doesn't do this, because it supports unboxed types; representation affects compilation. So GHC has to make choices. It does so at strange times, as @RaeHaskell shows: youtu.be/ovX7VE21FBM
Sam Derbyshire retweeted
Want to contribute to GHC? I came across a ticket today that should be nicely addressable by a newcomer. Hint: look in libraries/template-haskell/Language/Haskell/TH/Ppr.hs gitlab.haskell.org/ghc/ghc/-…












