Timeline


and

May 10, 2012:

5:15 PM Holes edited by spl
Minor typo (diff)
4:15 PM Changeset in ghc [4674a23]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdtc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Max Bolingbroke <batterseapower@…>
Fix bug in expandTypeSynonyms that could rarely cause problems
3:55 PM Changeset in nofib [9812ada]supercompiler by Max Bolingbroke <batterseapower@…>
Change how default supercompliation options are setup
3:55 PM Changeset in nofib [b65222f]supercompiler by Max Bolingbroke <batterseapower@…>
Playing with Bernouilli
3:55 PM Changeset in nofib [d77043e]supercompiler by Max Bolingbroke <batterseapower@…>
Ignore common OS junk
3:55 PM Changeset in nofib [5d0863d]supercompiler by Max Bolingbroke <batterseapower@…>
First draft of supercompilation nofib tests
3:44 PM Changeset in base [19f01f9]supercompiler by Max Bolingbroke <batterseapower@…>
Mark some List modules SUPERINLINABLE
3:27 PM Holes edited by spl
Revise description of variations in response to SLPJ (diff)
3:14 PM Changeset in ghc [007e710]supercompiler by Max Bolingbroke <batterseapower@…>
Tweak dead type defaulting
3:14 PM Changeset in ghc [e7cd0ba]supercompiler by Max Bolingbroke <batterseapower@…>
Tweaks to MSG, especially WRT handling of let-bound heapbindings
3:14 PM Changeset in ghc [d9eac575]supercompiler by Max Bolingbroke <batterseapower@…>
Fix major brainfart in msgVarBndr
3:14 PM Changeset in ghc [f20b2f7b]supercompiler by Max Bolingbroke <batterseapower@…>
FIXME only
3:14 PM Changeset in ghc [0b86b7a8]supercompiler by Max Bolingbroke <batterseapower@…>
Some FIXMEs about common-heap-vars optimisation in presence of MSG
3:14 PM Changeset in ghc [25a4e4b]supercompiler by Max Bolingbroke <batterseapower@…>
Further musing about the common-heap-vars hack for MSG
3:14 PM Changeset in ghc [f79e0c5]supercompiler by Max Bolingbroke <batterseapower@…>
Preserve the common-heap-vars invariant in MSG, exploit it more from Process3 (just use InScopeSet)
3:14 PM Changeset in ghc [596159a]supercompiler by Max Bolingbroke <batterseapower@…>
Comment only
3:14 PM Changeset in ghc [80a4201f]supercompiler by Max Bolingbroke <batterseapower@…>
Use MSG to implement matching in Process3 (experimentally)
3:14 PM Changeset in ghc [8b3355f3]supercompiler by Max Bolingbroke <batterseapower@…>
Comments and commoning-up of InstanceMatching
3:14 PM Changeset in ghc [ebe007c]supercompiler by Max Bolingbroke <batterseapower@…>
Restrict assertion
3:14 PM Changeset in ghc [1685b29]supercompiler by Max Bolingbroke <batterseapower@…>
Deal with non-matching heap binders by forcing generalisation
3:14 PM Changeset in ghc [a1355a8c]supercompiler by Max Bolingbroke <batterseapower@…>
Slight cleanup in applyAbsVars: loop with an InScopeSet instead of VarSet
3:14 PM Changeset in ghc [de1964e]supercompiler by Max Bolingbroke <batterseapower@…>
Add the missing PendingTerm handling to MSG
3:14 PM Changeset in ghc [13c4e4d]supercompiler by Max Bolingbroke <batterseapower@…>
Add some FIXMEs
3:14 PM Changeset in ghc [2b41c3d]supercompiler by Max Bolingbroke <batterseapower@…>
A marvellous method for allowing MSG to try to avoid generalisation for (x,e) pairs
3:14 PM Changeset in ghc [95fd4a1]supercompiler by Max Bolingbroke <batterseapower@…>
Add FIXME reminding me to remove PrimOp
3:14 PM Changeset in ghc [8b96ed2]supercompiler by Max Bolingbroke <batterseapower@…>
Tweak comment only
3:14 PM Changeset in ghc [5ea5aa66]supercompiler by Max Bolingbroke <batterseapower@…>
Extensive notes on match loops due to term-term matching
3:14 PM Changeset in ghc [e736d605]supercompiler by Max Bolingbroke <batterseapower@…>
Allow duplication of cheap stuff (not just values) when preparing term unfoldings
3:14 PM Changeset in ghc [094d50e]supercompiler by Max Bolingbroke <batterseapower@…>
Add comment about Indirect
3:14 PM Changeset in ghc [4b7623b]supercompiler by Max Bolingbroke <batterseapower@…>
Brackets only
3:14 PM Changeset in ghc [50e7af2e]supercompiler by Max Bolingbroke <batterseapower@…>
Unfailing term MSGing
3:14 PM Changeset in ghc [9df69812]supercompiler by Max Bolingbroke <batterseapower@…>
Experiment: allow common heap binders to clash with left/right heap binders
3:14 PM Changeset in ghc [0a1f003]supercompiler by Max Bolingbroke <batterseapower@…>
Fix a number of potential shadowing bugs in MSG, improve choice of Uniques at the same time
3:14 PM Changeset in ghc [021e35b]supercompiler by Max Bolingbroke <batterseapower@…>
Mark lambda contexts and be selective about what MSG floats from them
3:14 PM Changeset in ghc [170cee5]supercompiler by Max Bolingbroke <batterseapower@…>
Preliminary thoughts about renaming via MSG, invertRenaming
3:14 PM Changeset in ghc [245bdd8]supercompiler by Max Bolingbroke <batterseapower@…>
Incorporate the 'common heap vars' optimisation into MSG
3:14 PM Changeset in ghc [07e29c30]supercompiler by Max Bolingbroke <batterseapower@…>
Add additional check to the common-heap-vars match optimisation
3:14 PM Changeset in ghc [46840285]supercompiler by Max Bolingbroke <batterseapower@…>
Comments about binder MSG only
3:14 PM Changeset in ghc [ead945cd]supercompiler by Max Bolingbroke <batterseapower@…>
Fix stupid variable shadowing bug
3:14 PM Changeset in ghc [91565646]supercompiler by Max Bolingbroke <batterseapower@…>
Remove near-useless trace
3:14 PM Changeset in ghc [355a9c9]supercompiler by Max Bolingbroke <batterseapower@…>
MSGing *changes* FVs, so it is invalid to just traverse the Anned constructor. Doh.
3:14 PM Changeset in ghc [6b68150]supercompiler by Max Bolingbroke <batterseapower@…>
Space only
3:14 PM Changeset in ghc [7e111d4]supercompiler by Max Bolingbroke <batterseapower@…>
MSG type instantiations
3:14 PM Changeset in ghc [81555a5b]supercompiler by Max Bolingbroke <batterseapower@…>
Make type MSG unfailing
3:14 PM Changeset in ghc [7454df3]supercompiler by Max Bolingbroke <batterseapower@…>
Unfailingly MSG coercions
3:14 PM Changeset in ghc [1721b87c]supercompiler by Max Bolingbroke <batterseapower@…>
Comment only
3:14 PM Changeset in ghc [541667e0]supercompiler by Max Bolingbroke <batterseapower@…>
Had the same bugs in matching recursive lets as I did in the MSGer
3:14 PM Changeset in ghc [97d2022]supercompiler by Max Bolingbroke <batterseapower@…>
Tidy some traces, imports
3:14 PM Changeset in ghc [e3cc6d6]supercompiler by Max Bolingbroke <batterseapower@…>
Addendum to previous free vars fix: the InScopeSet also has to include the added FVs
3:14 PM Changeset in ghc [e524ff3]supercompiler by Max Bolingbroke <batterseapower@…>
Ensure we add renamings to both sides when MSGing lambda bound heap bindings
3:14 PM Changeset in ghc [40663c3]supercompiler by Max Bolingbroke <batterseapower@…>
Tidy error message
3:14 PM Changeset in ghc [37d1c06]supercompiler by Max Bolingbroke <batterseapower@…>
Better setting of generalisation flags in MSG
3:14 PM Changeset in ghc [56c89c40]supercompiler by Max Bolingbroke <batterseapower@…>
Ensure we can copy in lambdaBound bindigns for update frames in MSG
3:14 PM Changeset in ghc [8be1741]supercompiler by Max Bolingbroke <batterseapower@…>
Comment only
3:14 PM Changeset in ghc [fab4a36]supercompiler by Max Bolingbroke <batterseapower@…>
Tweak computation of MSG stack generalisation flag
3:14 PM Changeset in ghc [f3a333a2]supercompiler by Max Bolingbroke <batterseapower@…>
Remove MonadFix instance for MSG
3:14 PM Changeset in ghc [95869662]supercompiler by Max Bolingbroke <batterseapower@…>
Add option for history tree
3:14 PM Changeset in ghc [e8691a0]supercompiler by Max Bolingbroke <batterseapower@…>
Fix the rollback check, which appeared to be totally broken
3:14 PM Changeset in ghc [3a164c5]supercompiler by Max Bolingbroke <batterseapower@…>
Current best guess at incorporating MSG
3:14 PM Changeset in ghc [56e6ca55]supercompiler by Max Bolingbroke <batterseapower@…>
Checkpoint changes to error reporting
3:14 PM Changeset in ghc [ab79ca3]supercompiler by Max Bolingbroke <batterseapower@…>
Pick up free variables of FV unfoldings correctly when constructing heap
3:14 PM Changeset in ghc [daf45f1]supercompiler by Max Bolingbroke <batterseapower@…>
Correctly deal with recursive binding when MSGing, report better errors upon failure
3:14 PM Changeset in ghc [996d4144]supercompiler by Max Bolingbroke <batterseapower@…>
More debug traces
3:09 PM Changeset in ghc [5aa31d13]supercompiler by Max Bolingbroke <batterseapower@…>
Disable rules when simplifying stable unfolding to work around GHC.List non-termination
3:09 PM Changeset in ghc [db8fd4dc]supercompiler by Max Bolingbroke <batterseapower@…>
Messing with the heuristics
3:09 PM Changeset in ghc [bc6a8832]supercompiler by Max Bolingbroke <batterseapower@…>
Generalise residualisation code
3:09 PM Changeset in ghc [6f53c83c]supercompiler by Max Bolingbroke <batterseapower@…>
More feedback from ghcHeuristics
3:09 PM Changeset in ghc [5c99f5b]supercompiler by Max Bolingbroke <batterseapower@…>
Start computing unfoldings for GHC's heuristics at the occurrence sites (more reliable)
3:09 PM Changeset in ghc [a7cf8d9d]supercompiler by Max Bolingbroke <batterseapower@…>
Start paying attention to loop-breaker-ness again
3:09 PM Changeset in ghc [7be7c81]supercompiler by Max Bolingbroke <batterseapower@…>
Garbage collect in evaluator to detect used-once
3:09 PM Changeset in ghc [511ec132]supercompiler by Max Bolingbroke <batterseapower@…>
Add ability to set default inline pragmas for the whole module (can be overriden by specific ones)
3:09 PM Changeset in ghc [ec2f7fc]supercompiler by Max Bolingbroke <batterseapower@…>
Collect reasons as to why unfoldings couldn't be used, experimentally allow non-loopbreaker inlinings
3:09 PM Changeset in ghc [6c499e6]supercompiler by Max Bolingbroke <batterseapower@…>
Collect tags from AltCons as well because of positive information propagation
3:09 PM Changeset in ghc [d437772]supercompiler by Max Bolingbroke <batterseapower@…>
Check shouldExposeUnfolding in the evaluator too
3:09 PM Changeset in ghc [98ac9396]supercompiler by Max Bolingbroke <batterseapower@…>
Tag primops specially, skip wrappers for saturated occurrences
3:09 PM Changeset in ghc [859677f]supercompiler by Max Bolingbroke <batterseapower@…>
Comment only
3:09 PM Changeset in ghc [fdc77e4]supercompiler by Max Bolingbroke <batterseapower@…>
Push normalisation in (in the evaluator) in preparation for speculative inlining experiment
3:09 PM Changeset in ghc [5e91a454]supercompiler by Max Bolingbroke <batterseapower@…>
Grow some primitive inlining heuristics in the evaluator
3:09 PM Changeset in ghc [52e6777]supercompiler by Max Bolingbroke <batterseapower@…>
A whole mess of inlining heuristics
3:09 PM Changeset in ghc [2d199e5c]supercompiler by Max Bolingbroke <batterseapower@…>
Fix possible out-of-scope error in CoreBinds to CoreTerm translation
3:09 PM Changeset in ghc [3506e622]supercompiler by Max Bolingbroke <batterseapower@…>
Fix possible kinding error associated with defaulting dead type variables
3:09 PM Changeset in ghc [92cbcf2]supercompiler by Max Bolingbroke <batterseapower@…>
Rejig AbsVar handling so that dead TyVars get substituted in even at immediate tieback sites
3:09 PM Changeset in ghc [4f6c5f5]supercompiler by Max Bolingbroke <batterseapower@…>
Fix intensely stupid bug in binding partitioning
3:09 PM Changeset in ghc [15031a34]supercompiler by Max Bolingbroke <batterseapower@…>
Fix shadowing bug in DFunUnfoldings that showed up as a lint failure
3:09 PM Changeset in ghc [8cf2d732]supercompiler by Max Bolingbroke <batterseapower@…>
Use localiseId when manufacturing Id for positive information This prevents us creating some local bindings with external names, which messes up later simplifier runs
3:09 PM Changeset in ghc [83e11552]supercompiler by Max Bolingbroke <batterseapower@…>
Try to improve resid-tag tracking
3:09 PM Changeset in ghc [2ce5cd2]supercompiler by Max Bolingbroke <batterseapower@…>
Comment only about evaluatePrim
3:09 PM Changeset in ghc [6040009e]supercompiler by Max Bolingbroke <batterseapower@…>
Add SUPERINLINABLE pragma
3:09 PM Changeset in ghc [6972dc81]supercompiler by Max Bolingbroke <batterseapower@…>
Note about GCing before match, fixes a termination bug
3:09 PM Changeset in ghc [cbbb6fab]supercompiler by Max Bolingbroke <batterseapower@…>
Correct the commenting-out of the GC call
3:09 PM Changeset in ghc [f62af23]supercompiler by Max Bolingbroke <batterseapower@…>
Garbage collect when determining stateAbsVars since we stopped doing it in reduceForMatch
3:09 PM Changeset in ghc [4cf0709a]supercompiler by Max Bolingbroke <batterseapower@…>
In the evaluator, trim stack frames that are unreachable due to bottoming Ids
3:09 PM Changeset in ghc [b52b1b4]supercompiler by Max Bolingbroke <batterseapower@…>
Small comment with data about bernouilli
3:09 PM Changeset in ghc [2479681]supercompiler by Max Bolingbroke <batterseapower@…>
Improve TagBag pretty-printing
3:09 PM Changeset in ghc [251b7ae4]supercompiler by Max Bolingbroke <batterseapower@…>
Make children summary info shorter
3:09 PM Changeset in ghc [5a514101]supercompiler by Max Bolingbroke <batterseapower@…>
Change lookupIdSubst to rename global variables if a mapping is available
3:09 PM Changeset in ghc [8da007b2]supercompiler by Max Bolingbroke <batterseapower@…>
Deforest everything in sight
3:09 PM Changeset in ghc [521b5228]supercompiler by Max Bolingbroke <batterseapower@…>
Put the bulk of the tracing under flag control
3:09 PM Changeset in ghc [3234ffdb]supercompiler by Max Bolingbroke <batterseapower@…>
Carefully rethink what the tag in a CastBy actually means and update the code with the new story
3:09 PM Changeset in ghc [ddde588c]supercompiler by Max Bolingbroke <batterseapower@…>
Fix small bug in speculation and add multiple comments
3:09 PM Changeset in ghc [977fd43]supercompiler by Max Bolingbroke <batterseapower@…>
Pretty-print indirection specially
3:09 PM Changeset in ghc [fa539530]supercompiler by Max Bolingbroke <batterseapower@…>
Add SUPERCOMPILE pragma
3:09 PM Changeset in ghc [85be4686]supercompiler by Max Bolingbroke <batterseapower@…>
Correct flag name for memo cache preinitializatio
3:09 PM Changeset in ghc [e144ebd0]supercompiler by Max Bolingbroke <batterseapower@…>
Make use of SUPERCOMPILE pragmas when deciding what to SC
3:09 PM Changeset in ghc [c027f2e5]supercompiler by Max Bolingbroke <batterseapower@…>
Correct the circumstances in which SC IdInfo gets ppred
3:09 PM Changeset in ghc [f5ead43c]supercompiler by Max Bolingbroke <batterseapower@…>
Fix selective supercompilation: suck in all things *referred to*, not *referring to us*
3:09 PM Changeset in ghc [cabf31d6]supercompiler by Max Bolingbroke <batterseapower@…>
When abstracting over variables in the supercompiler, abstract over dictionaries first
3:09 PM Changeset in ghc [e0adb99]supercompiler by Max Bolingbroke <batterseapower@…>
SCC annotations all over the supercompiler
3:09 PM Changeset in ghc [54315875]supercompiler by Max Bolingbroke <batterseapower@…>
Preinitalization changes, normalise in eta-expansion: still a bit buggy, see GHC.Base.>>= in exp3_8
3:09 PM Changeset in ghc [ec50705d]supercompiler by Max Bolingbroke <batterseapower@…>
Fix a major bug with no-preinit where too many bindings were being eliminated post-speculation + some comments
3:09 PM Changeset in ghc [8402d155]supercompiler by Max Bolingbroke <batterseapower@…>
The to_binds when not preinitalizing are of mixed liftedness
3:09 PM Changeset in ghc [b896fa8d]supercompiler by Max Bolingbroke <batterseapower@…>
Another attempt at Binds->Term translation
3:09 PM Changeset in ghc [8f3fe711]supercompiler by Max Bolingbroke <batterseapower@…>
Non-preinitialization binds must go outside h functions
3:09 PM Changeset in ghc [3b6037c]supercompiler by Max Bolingbroke <batterseapower@…>
Zap binders in Binds->Term to work around GHC simplifier weirdness
3:09 PM Changeset in ghc [89be615]supercompiler by Max Bolingbroke <batterseapower@…>
Only abstract over RealWorld# conditionally
3:08 PM Changeset in ghc [0db230a]supercompiler by Max Bolingbroke <batterseapower@…>
Fix some instances where I had forgotten to consider varBndrFreeVars
3:08 PM Changeset in ghc [c30fc1dc]supercompiler by Max Bolingbroke <batterseapower@…>
Redefine tcSplitDFunTy so that it returns the types of the theta as well
3:08 PM Changeset in ghc [76e2d6ae]supercompiler by Max Bolingbroke <batterseapower@…>
Put Core-specific type-finding functions in the core module hierarchy
3:08 PM Changeset in ghc [3d7a4476]supercompiler by Max Bolingbroke <batterseapower@…>
Fix the DFunUnfolding elaboration so it isn't totally bogus
3:08 PM Changeset in ghc [0190e414]supercompiler by Max Bolingbroke <batterseapower@…>
Remove unused import
3:08 PM Changeset in ghc [7f4593a2]supercompiler by Max Bolingbroke <batterseapower@…>
Fix longstanding bug in stateType calculation for primop stack frames
3:08 PM Changeset in ghc [3ebcf844]supercompiler by Max Bolingbroke <batterseapower@…>
Changes to supercompiler code required to build after merge
3:08 PM Changeset in ghc [8b70add]supercompiler by Max Bolingbroke <batterseapower@…>
Explicit import lists from Data.Monoid to prevent <> clash
3:08 PM Changeset in ghc [fa91ea7e]supercompiler by Max Bolingbroke <batterseapower@…>
Remove unused variable
3:08 PM Changeset in ghc [cf3ba89]supercompiler by Max Bolingbroke <batterseapower@…>
Fill out the PtrRep case for when manufacturing dead argument fillers
3:08 PM Changeset in ghc [caa251bc]supercompiler by Max Bolingbroke <batterseapower@…>
Comment out match failure trace (very verbose)
3:08 PM Changeset in ghc [d3bd75e0]supercompiler by Max Bolingbroke <batterseapower@…>
Remove unused var warnings
3:08 PM Changeset in ghc [a246e38]supercompiler by Max Bolingbroke <batterseapower@…>
Support supplying a set of common variables to the matcher
3:08 PM Changeset in ghc [bd7d252]supercompiler by Max Bolingbroke <batterseapower@…>
Let the splitBy functions report failure
3:08 PM Changeset in ghc [4b56943]supercompiler by Max Bolingbroke <batterseapower@…>
Thread speculation history in attempt to tame huge speculation times
3:08 PM Changeset in ghc [60e024c]supercompiler by Max Bolingbroke <batterseapower@…>
Half-hearted attempt to deal with rollback to non-ancestors
3:08 PM Changeset in ghc [924f142e]supercompiler by Max Bolingbroke <batterseapower@…>
Implement instance-matching, including splitter support. This does indeed break map-map
3:08 PM Changeset in ghc [48ae26af]supercompiler by Max Bolingbroke <batterseapower@…>
Small tweaks only
3:08 PM Changeset in ghc [2d2e20b]supercompiler by Max Bolingbroke <batterseapower@…>
GC less agressively in speculation
3:08 PM Changeset in ghc [5bc0b5a]supercompiler by Max Bolingbroke <batterseapower@…>
New implementation of explicit promise tree (cleaner)
3:08 PM Changeset in ghc [595e672]supercompiler by Max Bolingbroke <batterseapower@…>
Use Train data type for promises tree instead (more precise), allow ancestor instance match
3:08 PM Changeset in ghc [883dcb69]supercompiler by Max Bolingbroke <batterseapower@…>
Full (possibly working) instance matching on ancestors, marking of generalised heap bindings/stack
3:08 PM Changeset in ghc [b24884b]supercompiler by Max Bolingbroke <batterseapower@…>
Missing lazy pattern match in eta
3:08 PM Changeset in ghc [5688858a]supercompiler by Max Bolingbroke <batterseapower@…>
Add parent tracing to speculation
3:08 PM Changeset in ghc [254acbf]supercompiler by Max Bolingbroke <batterseapower@…>
Be less verbose when showing MatchLRs (terms are too large to include)
3:08 PM Changeset in ghc [5297bd7]supercompiler by Max Bolingbroke <batterseapower@…>
Rethink skipping in Process3 memoisation
3:08 PM Changeset in ghc [01c11dd]supercompiler by Max Bolingbroke <batterseapower@…>
Put eager value splitting under flag control
3:08 PM Changeset in ghc [b85fbf24]supercompiler by Max Bolingbroke <batterseapower@…>
Have prepareTerm return a preinitialized version of the term as well
3:08 PM Changeset in ghc [d1b49822]supercompiler by Max Bolingbroke <batterseapower@…>
Memo table preinitialization in Process3
3:08 PM Changeset in ghc [7092279]supercompiler by Max Bolingbroke <batterseapower@…>
First version of heap instance-matching
3:08 PM Changeset in ghc [00c8ffc0]supercompiler by Max Bolingbroke <batterseapower@…>
Try to deal with the stack in instantiation-copying logic
3:08 PM Changeset in ghc [fbe7ad2]supercompiler by Max Bolingbroke <batterseapower@…>
Pretty print on instance match
3:08 PM Changeset in ghc [8a11d51]supercompiler by Max Bolingbroke <batterseapower@…>
Weaken assertion in gc function
3:08 PM Changeset in ghc [bcf49e8]supercompiler by Max Bolingbroke <batterseapower@…>
Checkpoint work on true instance-matching
3:08 PM Changeset in ghc [02566ff]supercompiler by Max Bolingbroke <batterseapower@…>
Make StaticFlags actualy work as command line arguments
3:08 PM Changeset in ghc [de136afe]supercompiler by Max Bolingbroke <batterseapower@…>
I was preventing match too eagerly by counting too many bindings as 'lambdaish'
3:08 PM Changeset in ghc [854e2762]supercompiler by Max Bolingbroke <batterseapower@…>
Fix a bug where the talis of a generalised state would not be manifest
3:08 PM Changeset in ghc [e5fa1ba]supercompiler by Max Bolingbroke <batterseapower@…>
Add utility functions for working with mixed type/value applications
3:08 PM Changeset in ghc [6cf81dd7]supercompiler by Max Bolingbroke <batterseapower@…>
Redundant import
3:08 PM Changeset in ghc [32805552]supercompiler by Max Bolingbroke <batterseapower@…>
Fix Process summary stats for shallow trees
3:08 PM Changeset in ghc [6a8fd8bf]supercompiler by Max Bolingbroke <batterseapower@…>
Some commented prettyprints for speculation
3:08 PM Changeset in ghc [324b1eb]supercompiler by Max Bolingbroke <batterseapower@…>
GC state in speculator to make termination check more lenient
3:08 PM Changeset in ghc [ee54e4fe]supercompiler by Max Bolingbroke <batterseapower@…>
Trim imports
3:08 PM Changeset in ghc [d6689d9f]supercompiler by Max Bolingbroke <batterseapower@…>
Show all deepest paths, not just one of them
3:08 PM Changeset in ghc [eb767cfb]supercompiler by Max Bolingbroke <batterseapower@…>
Add depth histogram, data structure summariser for leaf nodes
3:08 PM Changeset in ghc [932905e2]supercompiler by Max Bolingbroke <batterseapower@…>
Comment only
3:08 PM Changeset in ghc [235e966e]supercompiler by Max Bolingbroke <batterseapower@…>
Put positive information propagation under Staticflags control and turn it off
3:08 PM Changeset in ghc [a0e1613]supercompiler by Max Bolingbroke <batterseapower@…>
Checkpoint adding exceptions to Process3
3:08 PM Changeset in ghc [1219132]supercompiler by Max Bolingbroke <batterseapower@…>
Rollback
3:08 PM Changeset in ghc [0f394cdd]supercompiler by Max Bolingbroke <batterseapower@…>
Introduce CheckOnly memoisation mode
3:08 PM Changeset in ghc [6c485f4e]supercompiler by Max Bolingbroke <batterseapower@…>
Filter case alternatives in Split to take advantage of type refinement and eliminate stupid default cases
3:08 PM Changeset in ghc [dac1dfbd]supercompiler by Max Bolingbroke <batterseapower@…>
Show deepest path in SC at end of supercompilation
3:08 PM Changeset in ghc [38f41ba2]supercompiler by Max Bolingbroke <batterseapower@…>
Static-argument-transform some functions that are likely in the inner loop of term matching
3:08 PM Changeset in ghc [c5eac68f]supercompiler by Max Bolingbroke <batterseapower@…>
Fix SAT
3:08 PM Changeset in ghc [480ac43e]supercompiler by Max Bolingbroke <batterseapower@…>
Remove warning
3:08 PM Changeset in ghc [ed02c26b]supercompiler by Max Bolingbroke <batterseapower@…>
Improve deepest-path printing
3:08 PM Changeset in ghc [16d25e3]supercompiler by Max Bolingbroke <batterseapower@…>
Put case-alt refinement under flag control since it may be causing even more divergence in Main5
3:08 PM Changeset in ghc [d5374d3]supercompiler by Max Bolingbroke <batterseapower@…>
Record generalisation info in children tree
3:08 PM Changeset in ghc [cf0f6dd1]supercompiler by Max Bolingbroke <batterseapower@…>
Fix map-map fusion >_>
3:08 PM Changeset in ghc [6f08f8b]supercompiler by Max Bolingbroke <batterseapower@…>
Fix long-standing bug in zipBracketeds, exposed by eager value splitting
3:08 PM Changeset in ghc [854c774]supercompiler by Max Bolingbroke <batterseapower@…>
Tweaks to eager value splitting code only
3:08 PM Changeset in ghc [ee61318]supercompiler by Max Bolingbroke <batterseapower@…>
Fix a horrible matching bug that was causing too many binder pairs to be matched rigidly
3:08 PM Changeset in ghc [4549ab08]supercompiler by Max Bolingbroke <batterseapower@…>
Debugging aid: summary of node descendant count
3:08 PM Changeset in ghc [dd15e91]supercompiler by Max Bolingbroke <batterseapower@…>
Comment about indirections only
3:08 PM Changeset in ghc [fa5cfc0]supercompiler by Max Bolingbroke <batterseapower@…>
Comment only
3:08 PM Changeset in ghc [5851773e]supercompiler by Max Bolingbroke <batterseapower@…>
Add dEPTH_LIIMT
3:08 PM Changeset in ghc [4fc83553]supercompiler by Max Bolingbroke <batterseapower@…>
Make impossible-alternative-finding code more reusable Makes the following changes: 1. Generalises the type signatures of some functions relating to alternatives so that the type of "variables" and "expression" is not specified 2. Puts the bulk of the alternative-filtering code into a new function filterAlts (in CoreUtils) that can be used outside of the SimplM monad 3. Allows prepareAlts to return a null alternatives list if none are applicable - it turns out that this case was already handled by the caller (in the simplifier). This should result in a modest optimisation improvement in some cases.
3:08 PM Changeset in ghc [167fed1]supercompiler by Max Bolingbroke <batterseapower@…>
Ensure we localise AbsVars when building binding sites
3:08 PM Changeset in ghc [27024dc]supercompiler by Max Bolingbroke <batterseapower@…>
Fix free variables when deeply splitting values
3:08 PM Changeset in ghc [043abaeb]supercompiler by Max Bolingbroke <batterseapower@…>
Do not externalise non-Ids
3:08 PM Changeset in ghc [7ced16b]supercompiler by Max Bolingbroke <batterseapower@…>
Fix Entered information when eager splitting
3:08 PM Changeset in ghc [4f815ddf]supercompiler by Max Bolingbroke <batterseapower@…>
Changes with Simon
3:08 PM Changeset in ghc [d6128ff]supercompiler by Max Bolingbroke <batterseapower@…>
Abstract all h-functions over RealWorld#
3:08 PM Changeset in ghc [7b37a89]supercompiler by Max Bolingbroke <batterseapower@…>
Use a VoidRep placeholder that doesn't crash the simplifier
3:08 PM Changeset in ghc [02ca04fa]supercompiler by Max Bolingbroke <batterseapower@…>
Was removing too many FVs from the shell
3:08 PM Changeset in ghc [343b8ac8]supercompiler by Max Bolingbroke <batterseapower@…>
Small code cleanup to improve the output of RealWorld# arguments
3:08 PM Changeset in ghc [259e6f6c]supercompiler by Max Bolingbroke <batterseapower@…>
Track residualised tags for syntax dup stats
3:08 PM Changeset in ghc [e53710f6]supercompiler by Max Bolingbroke <batterseapower@…>
Better pretty-printing for lambdas (more horizontally compact)
3:08 PM Changeset in ghc [6ced9eb4]supercompiler by Max Bolingbroke <batterseapower@…>
Avoid overflow when using Deeds maxBound
3:08 PM Changeset in ghc [c302fb93]supercompiler by Max Bolingbroke <batterseapower@…>
Teach the splitter that the body of TyLambdas do not duplicate work
3:08 PM Changeset in ghc [921490e]supercompiler by Max Bolingbroke <batterseapower@…>
Stuff
3:08 PM Changeset in ghc [2c3692f7]supercompiler by Max Bolingbroke <batterseapower@…>
Eagerly split values to solve the false-tieback problem (I hope)
3:08 PM Changeset in ghc [3899bb7]supercompiler by Max Bolingbroke <batterseapower@…>
Fix warnings
3:08 PM Changeset in ghc [8a71f2f]supercompiler by Max Bolingbroke <batterseapower@…>
Give DataCons better tags
3:08 PM Changeset in ghc [39719535]supercompiler by Max Bolingbroke <batterseapower@…>
Add stop count
3:08 PM Changeset in ghc [de258ef]supercompiler by Max Bolingbroke <batterseapower@…>
Debugging trace upon matching
3:08 PM Changeset in ghc [0702b1f]supercompiler by Max Bolingbroke <batterseapower@…>
Trace depth, trace less
3:08 PM Changeset in ghc [a8a000d4]supercompiler by Max Bolingbroke <batterseapower@…>
Match after reduction
3:08 PM Changeset in ghc [8115435e]supercompiler by Max Bolingbroke <batterseapower@…>
Record reduction count in reduce'
3:08 PM Changeset in ghc [670493a9]supercompiler by Max Bolingbroke <batterseapower@…>
Change Deeds into abstract type, add step limit to it
3:08 PM Changeset in ghc [90188406]supercompiler by Max Bolingbroke <batterseapower@…>
Revert "Record reduction count in reduce'" This reverts commit a25cb09d0dc819c283a21d5bb29ea39abeadfe41.
3:08 PM Changeset in ghc [de85001]supercompiler by Max Bolingbroke <batterseapower@…>
Claim steps when reducing, ignore deeds for reduce-matching
3:08 PM Changeset in ghc [48aa16d]supercompiler by Max Bolingbroke <batterseapower@…>
Add TODOs, debug outputs..
3:08 PM Changeset in ghc [2ce8f693]supercompiler by Max Bolingbroke <batterseapower@…>
Add explicit Parent parameter
3:08 PM Changeset in ghc [b6bc3a0d]supercompiler by Max Bolingbroke <batterseapower@…>
Blah blah blah
3:08 PM Changeset in ghc [e7e1af8]supercompiler by Max Bolingbroke <batterseapower@…>
Tracing only
3:08 PM Changeset in ghc [70ad91ad]supercompiler by Max Bolingbroke <batterseapower@…>
Absolute rubbish
3:08 PM Changeset in ghc [c181a9c]supercompiler by Max Bolingbroke <batterseapower@…>
Generalise the hack which prevented self-recursive thunks being residualised
3:08 PM Changeset in ghc [b3e61c02]supercompiler by Max Bolingbroke <batterseapower@…>
Add Process3
3:08 PM Changeset in ghc [fa0ea797]supercompiler by Max Bolingbroke <batterseapower@…>
SPeculate in Process3 so we can do (f . g)
3:08 PM Changeset in ghc [c925d1d]supercompiler by Max Bolingbroke <batterseapower@…>
Prevent knot-tying bug in Match
3:08 PM Changeset in ghc [4efebd18]supercompiler by Max Bolingbroke <batterseapower@…>
Disable deeds by default
3:08 PM Changeset in ghc [27a163e]supercompiler by Max Bolingbroke <batterseapower@…>
Remove redundancies from Process2, fix a bug
3:08 PM Changeset in ghc [0d340c99]supercompiler by Max Bolingbroke <batterseapower@…>
Remove redundant export from Process1
3:08 PM Changeset in ghc [c55125f]supercompiler by Max Bolingbroke <batterseapower@…>
Make process2 the defaulrt
3:08 PM Changeset in ghc [02523b61]supercompiler by Max Bolingbroke <batterseapower@…>
Try not to panic in bindCapturedFloats (still a hack)
3:08 PM Changeset in ghc [0429934]supercompiler by Max Bolingbroke <batterseapower@…>
Add speculation and rollback to Process2
3:08 PM Changeset in ghc [0e2dbf4e]supercompiler by Max Bolingbroke <batterseapower@…>
Implement graphical history, thread speculation properly
3:08 PM Changeset in ghc [9bce977]supercompiler by Max Bolingbroke <batterseapower@…>
Refactor in terms of terminateM combinator
3:08 PM Changeset in ghc [b4103f89]supercompiler by Max Bolingbroke <batterseapower@…>
Just rearrange stuff a bit, change the meaning of generated_key
3:08 PM Changeset in ghc [8f5c280]supercompiler by Max Bolingbroke <batterseapower@…>
I may have the perfect refactoring of Process2
3:08 PM Changeset in ghc [802e6010]supercompiler by Max Bolingbroke <batterseapower@…>
Checkpoint intermediate DelayM refactoring
3:08 PM Changeset in ghc [8088569]supercompiler by Max Bolingbroke <batterseapower@…>
Checkpoint building version of new process with function to delay folded in
3:08 PM Changeset in ghc [9775401]supercompiler by Max Bolingbroke <batterseapower@…>
Use a functor type parameter to DelayStructure rather than a GADT index
3:08 PM Changeset in ghc [63c09448]supercompiler by Max Bolingbroke <batterseapower@…>
Fill out sc' definition: history still not propagated
3:08 PM Changeset in ghc [c16f4e6]supercompiler by Max Bolingbroke <batterseapower@…>
Generalise the type of memo, for now
3:08 PM Changeset in ghc [d17c607d]supercompiler by Max Bolingbroke <batterseapower@…>
Fulfilments in Process2 (well, it typechecks)
3:08 PM Changeset in ghc [604b3b33]supercompiler by Max Bolingbroke <batterseapower@…>
Bring both histories into scope and complete implementation for threaded history
3:08 PM Changeset in ghc [6db05a3]supercompiler by Max Bolingbroke <batterseapower@…>
Good-enough implementation of MonadStatics for now
3:08 PM Changeset in ghc [1182acec]supercompiler by Max Bolingbroke <batterseapower@…>
Small tweaks
3:08 PM Changeset in ghc [50b8d4bd]supercompiler by Max Bolingbroke <batterseapower@…>
Checkpoint an experiment in process-tree refactoring
3:04 PM Changeset in ghc [b61b975b]supercompiler by Max Bolingbroke <batterseapower@…>
Wasn't releasing deeds when destroying identity CastIt
3:04 PM Changeset in ghc [ee4969d]supercompiler by Max Bolingbroke <batterseapower@…>
Allow statics to not be printed (cleaner output)
3:04 PM Changeset in ghc [97e7300]supercompiler by Max Bolingbroke <batterseapower@…>
PRety-printing tweaks
3:04 PM Changeset in ghc [fe12ea1f]supercompiler by Max Bolingbroke <batterseapower@…>
Bug fixes detected by Lint
3:04 PM Changeset in ghc [ee1f9640]supercompiler by Max Bolingbroke <batterseapower@…>
Turn deeds on permanently
3:04 PM Changeset in ghc [52e27bd]supercompiler by Max Bolingbroke <batterseapower@…>
Fix various bugs to do with IdInfo
3:04 PM Changeset in ghc [77f89c7]supercompiler by Max Bolingbroke <batterseapower@…>
Small tweaks, still doesn't work
3:04 PM Changeset in ghc [f88d9eb]supercompiler by Max Bolingbroke <batterseapower@…>
Simplify a lot of bullshit
3:04 PM Changeset in ghc [2cbdf8b8]supercompiler by Max Bolingbroke <batterseapower@…>
Change how we deal with dead variable FV improvement
3:04 PM Changeset in ghc [bd1cc03]supercompiler by Max Bolingbroke <batterseapower@…>
Fill in pretty-printer
3:04 PM Changeset in ghc [fc9440d1]supercompiler by Max Bolingbroke <batterseapower@…>
Checkpoint pretty printing with ctrl-c handling
3:04 PM Changeset in ghc [e27886b9]supercompiler by Max Bolingbroke <batterseapower@…>
More normalised coercions
3:04 PM Changeset in ghc [753d5d3b]supercompiler by Max Bolingbroke <batterseapower@…>
Fix stupidity added to OptCoercion
3:04 PM Changeset in ghc [7754719]supercompiler by Max Bolingbroke <batterseapower@…>
Fixes for latest GHC, eliminate refls more agressively, real data type for Coerced
3:04 PM Changeset in ghc [d7cf9dbf]supercompiler by Max Bolingbroke <batterseapower@…>
Eliminate trivial CastIt frames as well
3:04 PM Changeset in ghc [2e837ef4]supercompiler by Max Bolingbroke <batterseapower@…>
TODOs only
3:04 PM Changeset in ghc [7a40ef0b]supercompiler by Max Bolingbroke <batterseapower@…>
Warnings only
3:04 PM Changeset in ghc [a91def9c]supercompiler by Max Bolingbroke <batterseapower@…>
Ensure we don't wrap answers in spurious refl coercions, for the sake of rule matching
3:04 PM Changeset in ghc [89e949db]supercompiler by Max Bolingbroke <batterseapower@…>
Checkpoint broken explicit process tree
3:04 PM Changeset in ghc [05592937]supercompiler by Max Bolingbroke <batterseapower@…>
Correct partitionFulfilments to only consider each fulfilment at most once. Record holes in the FulfilmentTree to match dropped fulfilments
3:04 PM Changeset in ghc [51e7ab9d]supercompiler by Max Bolingbroke <batterseapower@…>
Vastly simplify the FV-sensitive driving loop in Split
3:04 PM Changeset in ghc [43aa937]supercompiler by Max Bolingbroke <batterseapower@…>
Changed some Maybe into real data types
3:04 PM Changeset in ghc [f099c36]supercompiler by Max Bolingbroke <batterseapower@…>
Checkpoint pretty-printing
3:04 PM Changeset in ghc [3831bdd]supercompiler by Max Bolingbroke <batterseapower@…>
Don't rely on Foldable to enumerate eligible fulfilments
3:04 PM Changeset in ghc [9ac9f96]supercompiler by Max Bolingbroke <batterseapower@…>
Fix to KPush that somehow got missed when I put universal type args in Data
3:04 PM Changeset in ghc [81a8c51]supercompiler by Max Bolingbroke <batterseapower@…>
Use my own pPrintBndr rather than the noisy pprBndr
3:04 PM Changeset in ghc [028bb610]supercompiler by Max Bolingbroke <batterseapower@…>
Add coercionKindNonRepr to work around coercionKind bug
3:04 PM Changeset in ghc [36daa53]supercompiler by Max Bolingbroke <batterseapower@…>
When residualising a term to SDoc, wrap angle brackets around the focus
3:04 PM Changeset in ghc [ac5c619]supercompiler by Max Bolingbroke <batterseapower@…>
Comment only
3:04 PM Changeset in ghc [1849424]supercompiler by Max Bolingbroke <batterseapower@…>
Tweak to Evaluate to prevent a cast answer evaluating to itself in one step (triggers reduce-stop). Still needs work.
3:04 PM Changeset in ghc [a1ba5fe]supercompiler by Max Bolingbroke <batterseapower@…>
Cast things are cheap as well
3:04 PM Changeset in ghc [f50d8497]supercompiler by Max Bolingbroke <batterseapower@…>
Add pprPreview2 for comparing certain SDocs within an editor
3:04 PM Changeset in ghc [bb3a0cbe]supercompiler by Max Bolingbroke <batterseapower@…>
Tracing logic for reduce-stop fiddle
3:04 PM Changeset in ghc [e8b174b]supercompiler by Max Bolingbroke <batterseapower@…>
We have to include the universal type arguments in Data or we can't derive a value's type
3:04 PM Changeset in ghc [89f53f7]supercompiler by Max Bolingbroke <batterseapower@…>
More comprehensive treatment of unlifted let bindings: in particular, stop residualising them as LetRecs
3:04 PM Changeset in ghc [345832aa]supercompiler by Max Bolingbroke <batterseapower@…>
Huge refactoring of the Bracketed data type so we can thread the correct branch type through to case_ if we use pushStack
3:04 PM Changeset in ghc [80d2669b]supercompiler by Max Bolingbroke <batterseapower@…>
Whitespace sanity in zipBracketeds
3:04 PM Changeset in ghc [6b0f6e3]supercompiler by Max Bolingbroke <batterseapower@…>
Fix subtle bugs introduced by recent changes (bindMany was bogus, for example..)
3:04 PM Changeset in ghc [4782266]supercompiler by Max Bolingbroke <batterseapower@…>
Enforce the invariant that Data never includes a newtype DataCon
3:04 PM Changeset in ghc [3d9f5246]supercompiler by Max Bolingbroke <batterseapower@…>
Do not assume that the RHS of an unlifted Let is ok-for-speculation
3:04 PM Changeset in ghc [b3a4c9b]supercompiler by Max Bolingbroke <batterseapower@…>
Fix warning, remove dead code
3:04 PM Changeset in ghc [10f7674e]supercompiler by Max Bolingbroke <batterseapower@…>
Improve pretty-printing of binders in Core (show their types)
3:04 PM Changeset in ghc [4c43261]supercompiler by Max Bolingbroke <batterseapower@…>
sc-rollback was totally busted in my porting effort
3:04 PM Changeset in ghc [a1c5f5b]supercompiler by Max Bolingbroke <batterseapower@…>
Be more fastidious about checking list-length-equality in the supercompiler
3:04 PM Changeset in ghc [4d50df6]supercompiler by Max Bolingbroke <batterseapower@…>
Correct the form of implicit datacon workers: only apply existential TVs
3:04 PM Changeset in ghc [e1cf97d2]supercompiler by Max Bolingbroke <batterseapower@…>
The case wildcard binding was totally bogus, and also triggered a deeds assertion
3:04 PM Changeset in ghc [e195faac]supercompiler by Max Bolingbroke <batterseapower@…>
The evaluator was not releasing deeds from coercions inside an Answer
3:04 PM Changeset in ghc [bfde2ee]supercompiler by Max Bolingbroke <batterseapower@…>
More cleanups in the evaluator, fix a bug where too many deeds would be released for cast application
3:04 PM Changeset in ghc [d981ac3]supercompiler by Max Bolingbroke <batterseapower@…>
I was accidentally duplicating coercions on indirections
3:04 PM Changeset in ghc [6bc0e4bd]supercompiler by Max Bolingbroke <batterseapower@…>
General cleanup of Core syntax predicates and dead code elimination
3:04 PM Changeset in ghc [3b675a8]supercompiler by Max Bolingbroke <batterseapower@…>
Eliminate some dodgy const fmaps with traverses
3:04 PM Changeset in ghc [f71dfb7]supercompiler by Max Bolingbroke <batterseapower@…>
Must abstract over type varibles as well when refining fulfilment FVs
3:04 PM Changeset in ghc [3774d47]supercompiler by Max Bolingbroke <batterseapower@…>
Improve assertion formatting in step'
3:04 PM Changeset in ghc [0130952]supercompiler by Max Bolingbroke <batterseapower@…>
Try to handle coercion variable substitutions properly
3:04 PM Changeset in ghc [319a1107]supercompiler by Max Bolingbroke <batterseapower@…>
Bite the bullet and break out a separate CoApp constructor, fix bugs exposed by its introduction (quite a few)
3:04 PM Changeset in ghc [ddfbcbe4]supercompiler by Max Bolingbroke <batterseapower@…>
Comment only
3:04 PM Changeset in ghc [113770f4]supercompiler by Max Bolingbroke <batterseapower@…>
Remove some redundant insertion operations on Renaming
3:04 PM Changeset in ghc [89175e0]supercompiler by Max Bolingbroke <batterseapower@…>
Improve splitSubst hack to ensure we don't lose CoVar->CoVar and TyVar->TyVar renamings
3:04 PM Changeset in ghc [8db5e678]supercompiler by Max Bolingbroke <batterseapower@…>
Fix build of plugin code, write script for installing it inplace (for testing)
3:04 PM Changeset in ghc [80a02e5]supercompiler by Max Bolingbroke <batterseapower@…>
Fix scrutination of cast data with Simon's new dealWithCoercion function
3:04 PM Changeset in ghc [6585d454]supercompiler by Max Bolingbroke <batterseapower@…>
Just in case, unconditionally export primop/dc unfoldings even if we shouldn't
3:04 PM Changeset in ghc [71462ae]supercompiler by Max Bolingbroke <batterseapower@…>
Fix construction of the initial heap from the unfoldings: FVs of unfoldings were not bound
3:04 PM Changeset in ghc [6d92f39f]supercompiler by Max Bolingbroke <batterseapower@…>
Eliminate strings from assertions, since they pretty print like lists
3:04 PM Changeset in ghc [eb542955]supercompiler by Max Bolingbroke <batterseapower@…>
Fix subtle bug where annotated variable free variables would not be renamed
3:04 PM Changeset in ghc [e2b3fa1]supercompiler by Max Bolingbroke <batterseapower@…>
Rename free variables in annotation correctly in termToAnswer
3:04 PM Changeset in ghc [b9e6c7c9]supercompiler by Max Bolingbroke <batterseapower@…>
Supress warning in Split
3:04 PM Changeset in ghc [ffa5c71]supercompiler by Max Bolingbroke <batterseapower@…>
Don't lose deeds when the evaluator builds a term from a bare Var
3:04 PM Changeset in ghc [6f6463a2]supercompiler by Max Bolingbroke <batterseapower@…>
Only bind exported things into the big tuple
3:04 PM Changeset in ghc [f209ed7]supercompiler by Max Bolingbroke <batterseapower@…>
Totally rewrite matcher: detect rigid binder inequality earlier
3:04 PM Changeset in ghc [7bff7ac4]supercompiler by Max Bolingbroke <batterseapower@…>
Stop building a ridiculous loop in the matcher
3:04 PM Changeset in ghc [a4658d9a]supercompiler by Max Bolingbroke <batterseapower@…>
Only expose unfoldings visible in phase 2
3:04 PM Changeset in ghc [88b7f83]supercompiler by Max Bolingbroke <batterseapower@…>
Cleanup code outside supercompiler directory
3:04 PM Changeset in ghc [192e207]supercompiler by Max Bolingbroke <batterseapower@…>
Add instances for non-1 variants derived from 1 variants, remove some uses of PrettyFunction and Wrapper1
3:04 PM Changeset in ghc [498cb10a]supercompiler by Max Bolingbroke <batterseapower@…>
Comment only
3:04 PM Changeset in ghc [0bbe9d73]supercompiler by Max Bolingbroke <batterseapower@…>
Remove my generalisation of VarEn
3:04 PM Changeset in ghc [7ccd6ea]supercompiler by Max Bolingbroke <batterseapower@…>
Initial cut of plugin functionality. Plugin likely won't build
3:04 PM Changeset in ghc [ce15491]supercompiler by Max Bolingbroke <batterseapower@…>
Checkpoint better FV calculation in splitter
3:04 PM Changeset in ghc [559784ab]supercompiler by Max Bolingbroke <batterseapower@…>
Fix lurking bugs in free variable of Bracketed code, fix splitting of type lambdas
3:04 PM Changeset in ghc [961326e4]supercompiler by Max Bolingbroke <batterseapower@…>
Dump debugging code
3:04 PM Changeset in ghc [f5002061]supercompiler by Max Bolingbroke <batterseapower@…>
Match types whenever we match Ids: SC now loops because we don't type-abstract enough
3:04 PM Changeset in ghc [cd931d04]supercompiler by Max Bolingbroke <batterseapower@…>
Fix lots of lurking badness in Renaming. InScopeSets are still wrong
3:04 PM Changeset in ghc [93be0d5]supercompiler by Max Bolingbroke <batterseapower@…>
Fixed some missed type renaming in the evaluator primop case
3:04 PM Changeset in ghc [6c138f5]supercompiler by Max Bolingbroke <batterseapower@…>
Use InScopeSet as a real InScopeSet, not an InScopeMap
3:04 PM Changeset in ghc [931de9fd]supercompiler by Max Bolingbroke <batterseapower@…>
Fix positive information propagation
3:04 PM Changeset in ghc [c43b42d]supercompiler by Max Bolingbroke <batterseapower@…>
Add comment about unfoldings
3:04 PM Changeset in ghc [6b1a46e1]supercompiler by Max Bolingbroke <batterseapower@…>
The Process module now compiles
3:04 PM Changeset in ghc [a6778cbf]supercompiler by Max Bolingbroke <batterseapower@…>
Support strict lets, coercions as values, write term->CoreSyn conversion
3:04 PM Changeset in ghc [fa91c6f1]supercompiler by Max Bolingbroke <batterseapower@…>
Deal with [CoreBind] by going via CoreExpr
3:04 PM Changeset in ghc [fce2c403]supercompiler by Max Bolingbroke <batterseapower@…>
Implement wrappers properly, give PrimOps type arguments
3:04 PM Changeset in ghc [0bbb48c2]supercompiler by Max Bolingbroke <batterseapower@…>
Split DataAlt binder list, fix FV calculation, fix code generation for binder->term embedding
3:04 PM Changeset in ghc [ca76f32]supercompiler by Max Bolingbroke <batterseapower@…>
Port over the splitter
2:53 PM Status/May12 edited by diatchki
(diff)
2:53 PM Status/May12 edited by diatchki
(diff)
2:47 PM Status/May12 edited by diatchki
(diff)
2:23 PM Status/May12 edited by bgamari
add note about ARM GHCi support (diff)
12:46 PM Ticket #4889 (Ignore broken pipe error when writing to a died process) closed by pcapriotti
fixed: […]
12:41 PM Changeset in process [76fb4b7]ghc-7.6ghc-7.8 by Paolo Capriotti <p.capriotti@…>
Ignore broken pipe error in readProcessWithExitCode (#4889)
12:39 PM Changeset in ghc [e4168865]supercompiler by Max Bolingbroke <batterseapower@…>
More small fixes
12:39 PM Changeset in ghc [e51b461]supercompiler by Max Bolingbroke <batterseapower@…>
Checkpoint (broken)
12:39 PM Changeset in ghc [bf2ba32]supercompiler by Max Bolingbroke <batterseapower@…>
A more plausible checkpoint, but still broken overall
12:39 PM Changeset in ghc [5b88481]supercompiler by Max Bolingbroke <batterseapower@…>
The evaluator finally compiles
12:39 PM Changeset in ghc [b1011f2]supercompiler by Max Bolingbroke <batterseapower@…>
Checkpoint work on matcher
12:39 PM Changeset in ghc [a444857]supercompiler by Max Bolingbroke <batterseapower@…>
Eliminate tabs
12:39 PM Changeset in ghc [bf8fae9]supercompiler by Max Bolingbroke <batterseapower@…>
Matching compiles and should handle new syntax
12:39 PM Changeset in ghc [50c06811]supercompiler by Max Bolingbroke <batterseapower@…>
Initial supercompiler commit
12:39 PM Changeset in ghc [1ea7885]supercompiler by Max Bolingbroke <batterseapower@…>
Port stuff over up until the evaluator itself (infrastructure is in place)
12:39 PM Changeset in ghc [89091c11]supercompiler by Max Bolingbroke <batterseapower@…>
Misc fixes to supercompiler, moving it forward
12:39 PM Changeset in ghc [86909c04]supercompiler by Max Bolingbroke <batterseapower@…>
Commit more work on the porting
12:39 PM Changeset in ghc [13ed44ce]supercompiler by Max Bolingbroke <batterseapower@…>
Small fixes
12:32 PM Status/May12 edited by simonmar
(diff)
12:30 PM Status/May12 edited by simonmar
(diff)
12:28 PM DeferErrorsToRuntime edited by dreixel
(diff)
12:06 PM Status/SLPJ-Tickets edited by simonpj
(diff)
12:05 PM Ticket #6087 (Join points need strictness analysis) created by simonpj
I came across this code in the nofib/spectral progam knights, in the …
11:46 AM Ticket #5623 (GHC 7.2.1 Performance Regression: Vector) closed by simonpj
fixed: I've added a test for Roman's plusPtr example. I'm not sure how to …
11:45 AM Changeset in ghc [b2d98cc3]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Simon Peyton Jones <simonpj@…>
Test Trac #5623
11:45 AM Changeset in testsuite [ff33bd8]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Simon Peyton Jones <simonpj@…>
Test Trac #5623
10:42 AM Status/May12 edited by simonpj
(diff)
10:42 AM Status/May12 edited by simonpj
(diff)
10:41 AM Status/May12 edited by simonpj
(diff)
10:16 AM Status/May12 edited by simonmar
(diff)
9:57 AM Building/RunningTests/Adding edited by simonmar
(diff)
9:56 AM Building/RunningTests/Adding edited by simonmar
add info about performance tests (diff)
8:45 AM Changeset in ghc [c1e928e]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Simon Peyton Jones <simonpj@…>
Comments only
8:32 AM Status/SLPJ-Tickets edited by simonpj
(diff)
7:52 AM Changeset in ghc [4faed348]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Simon Peyton Jones <simonpj@…>
Accept debug-print wibbles
7:52 AM Changeset in testsuite [a65db87]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Simon Peyton Jones <simonpj@…>
Accept debug-print wibbles
6:40 AM Changeset in ghc [007bfeb]type-nats by Iavor S. Diatchki <iavor.diatchki@…>
Merge with latest updated to constraint solver.

May 9, 2012:

11:53 PM Changeset in ghc [59e1f068]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdtc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simplewip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Ian Lynagh <igloo@…>
Fix build on Win64
6:04 PM Ticket #6086 (Cross compilation fails using system linker for other architecture ...) created by mtjm
Building ghc from git revision 4b089dbacd2a14f38a335103cf8ac0779d92f937
5:03 PM Changeset in ghc [43e1ff2d]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Simon Peyton Jones <simonpj@…>
Merge branch 'master' of http://darcs.haskell.org/testsuite
5:03 PM Changeset in testsuite [b982c7c]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Simon Peyton Jones <simonpj@…>
Merge branch 'master' of http://darcs.haskell.org/testsuite
5:02 PM Ticket #6048 (Exponential inlining code blowup) closed by simonpj
fixed: Really nice example. Thank you for making it so compelling. Fixed!
5:01 PM Changeset in ghc [af0bf03c]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Simon Peyton Jones <simonpj@…>
Test Trac #6048
5:01 PM Changeset in testsuite [fda160e]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Simon Peyton Jones <simonpj@…>
Test Trac #6048
4:54 PM Ticket #6078 (GHC panic encountered for unlifted let bindings) closed by simonpj
fixed: Thanks. Fixed.
4:52 PM Changeset in ghc [1cec00d]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Simon Peyton Jones <simonpj@…>
Merge branch 'master' of http://darcs.haskell.org//ghc
4:50 PM Changeset in ghc [3528d0a]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Simon Peyton Jones <simonpj@…>
Merge branch 'master' of http://darcs.haskell.org/testsuite
4:50 PM Changeset in testsuite [d9553c1]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Simon Peyton Jones <simonpj@…>
Merge branch 'master' of http://darcs.haskell.org/testsuite
4:49 PM Changeset in ghc [e4212c29]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Simon Peyton Jones <simonpj@…>
Wibble to simplified output
4:49 PM Changeset in testsuite [52fd080]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Simon Peyton Jones <simonpj@…>
Wibble to simplified output
4:49 PM Changeset in ghc [7d3a6ae]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Simon Peyton Jones <simonpj@…>
Better performance for #3064, but #5030 is now expect-broken
4:49 PM Changeset in testsuite [ffc4384]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Simon Peyton Jones <simonpj@…>
Better performance for #3064, but #5030 is now expect-broken
4:48 PM Changeset in ghc [9dfd8cd7]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Simon Peyton Jones <simonpj@…>
Test Trac #6078
4:48 PM Changeset in testsuite [37769ce]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Simon Peyton Jones <simonpj@…>
Test Trac #6078
4:45 PM Ticket #4452 (Graphics.Win32.GDI.Clip.enumClipboardFormats fails on success.) closed by pcapriotti
fixed: Pushed as …
4:39 PM Changeset in ghc [93d853b]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Paolo Capriotti <p.capriotti@…>
Move win32 tests to win32 repository.
4:39 PM Changeset in testsuite [8edee7a]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Paolo Capriotti <p.capriotti@…>
Move win32 tests to win32 repository.
3:45 PM Changeset in ghc [cd954499]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdtc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simplewip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Ian Lynagh <igloo@…>
Fix build on Win64
3:31 PM Status/SLPJ-Tickets edited by simonpj
(diff)
3:27 PM Ticket #6085 (I found that it cannot be written such as [False..]) closed by simonpj
invalid: The trouble is that "False.." is being lexed as a qualified name. You …
3:22 PM Changeset in ghc [980372f3]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Simon Peyton Jones <simonpj@…>
Re-do the "function application discount" (fixes Trac #6048) * Undoes Max's very aggressive function-inlining change (see comments with Trac #6048) * Resticts function application discount to functions that occur just once in the body. It was the multiple occurrences that led to the exponential behavour in Trac #6048. See Note [Function application discount] in CoreUnfold. Module binary sizes are down 2% on average, which is good. Allocations wobble about a bit, but only on a few benchmarks and not by much, so it seems a price worth paying to avoid exponential behaviour! Allocs Min -1.2% Max +2.8% Geometric Mean +0.0%
3:03 PM Changeset in ghc [4c550307]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Simon Peyton Jones <simonpj@…>
Take care not to mix polymorphic and unlifted bindings in a group Fixes Trac #6078
3:03 PM Changeset in ghc [e08cad7]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Simon Peyton Jones <simonpj@…>
Wibbles to 'simplify the SimplCont data type'
3:02 PM Changeset in ghc [4bbd6ba]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Simon Peyton Jones <simonpj@…>
Simplify the SimplCont data type * Put the result type in the Stop continuation * No need for the alts type in Select The result is a modest but useful simplification
2:50 PM Changeset in ghc [f7e246b5]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Simon Peyton Jones <simonpj@…>
Merge branch 'ghc-new-flavor'
2:26 PM Status/May12 edited by igloo
(diff)
1:49 PM Changeset in ghc [203afa1c]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdtc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simplewip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Ian Lynagh <igloo@…>
Don't pointlessly cast sizeof's result to size_t
1:47 PM Ticket #5727 (Unclear documentation about .eventlog's spark information flag) closed by duncan
worksforme: I cannot reproduce the behaviour. It works correctly for me with …
1:12 PM Changeset in ghc [ff10612]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-new-flavorghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Simon Peyton Jones <simonpj@…>
Make fresh variables when decomposing Givens This turns out to be important becuase we don't have a form for superclass selection in TcCoercion (we could but we don't). Se comments with xCtFlavor_cache, the Given case.
1:11 PM Changeset in ghc [d63e81b]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-new-flavorghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Simon Peyton Jones <simonpj@…>
Use fresh uniques when unboxing coercions in the desugarer This is kosher, and turns out to be vital when we have more complicate evidence terms.
1:11 PM Changeset in ghc [a606905]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-new-flavorghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Simon Peyton Jones <simonpj@…>
Cosmetics
10:06 AM Changeset in ghc [2112f43]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Simon Peyton Jones <simonpj@…>
Be a little less aggressive about inlining (fixes Trac #5623) When inlining, we are making a copy of the expression, so we have to be careful about duplicating work. Previously we were using exprIsCheap for that, but it is willing to duplicate a cheap primop -- and that is terribly bad if it happens inside some inner array loop (Trac #5623). So now we use a new function exprIsWorkFree. Even then there is some wiggle room: see Note [exprIsWorkFree] in CoreUtils This commit does make wheel-sieve1 allocate a lot more, but we decided that's just tough; it's more important for inlining to be robust about not duplicating work.
7:03 AM Ticket #6085 (I found that it cannot be written such as [False..]) created by phi16
Please see the following results: Prelude> [False..] <interactive>:0:9: …

May 8, 2012:

7:17 PM Changeset in base [89e58b9]data-proxydbcsghc-7.6ghc-7.8imp-param-classsupercompilertype-reasoningwindows-iocp by Paolo Capriotti <p.capriotti@…>
Fix build.
7:17 PM Changeset in ghc [46314891]wip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/orf by Paolo Capriotti <p.capriotti@…>
Fix build.
5:31 PM Changeset in ghc [4b089dba]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Ian Lynagh <igloo@…>
Merge branch 'master' of darcs.haskell.org:/srv/darcs//ghc
3:51 PM Ticket #5967 (incompatible implicit declaration of function 'memcpy') closed by pcapriotti
fixed: Merged as 0e069310cec2ef3df1d1928fa7683a87aebd9eed.
3:51 PM Ticket #6006 (hs_init(NULL, NULL) segfaults on latest snapshot) closed by pcapriotti
fixed: Merged as d7c6368938c77644da7c8410df4d4fa440aca7e8.
3:50 PM Ticket #6007 (ghci: Importing specific non-existent names from a module breaks later ...) closed by pcapriotti
fixed: Merged as 81885c1c34af716ed3dad159e927c943270eb74b.
3:50 PM Changeset in ghc [cb384a6]ghc-7.4 by Paolo Capriotti <p.capriotti@…>
Build fix.
3:04 PM Changeset in ghc [81885c1c]ghc-7.4 by Paolo Capriotti <p.capriotti@…>
Catch illegal imports earlier (#6007) MERGED from commit b6e28266cf29bfdf07fe08e894df77b400da8a04
2:51 PM Changeset in ghc [d7c6368]ghc-7.4 by Paolo Capriotti <p.capriotti@…>
hs_init(): cope with argc and/or argv being NULL (#6006) MERGED from commit 4ca281829c70331571291ed3dcf813a6028cc904
2:46 PM Ticket #4363 (openFile sharing permissions are inconsistent across platforms) closed by pcapriotti
fixed
2:04 PM Commentary/CodingStyle edited by danbst
typo (diff)
1:24 PM Changeset in ghc [0e06931]ghc-7.4 by Paolo Capriotti <p.capriotti@…>
Cast memory primops in the C backend (#5976) To prevent conflicts with GCC builtins, generate identical code for calls to mem primos and FFI calls. Based on a patch by Joachim Breitner. MERGED from commit 40c1106c338e209f07023d165f32bff0f75e2e54
1:09 PM Changeset in ghc [9fb12e14]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Paolo Capriotti <p.capriotti@…>
Enable FileLock for win32 (#4363)
1:08 PM Changeset in base [ae7d0c7]data-proxydbcsghc-7.6ghc-7.8imp-param-classsupercompilertype-reasoningwindows-iocp by Paolo Capriotti <p.capriotti@…>
Use in-process file locking on Windows (#4363)
1:08 PM Changeset in ghc [9cb3d363]wip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/orf by Paolo Capriotti <p.capriotti@…>
Use in-process file locking on Windows (#4363)
12:47 PM Status/May12 edited by chak
(diff)
12:34 PM Ticket #6061 (threadDelay broken on Windows) closed by pcapriotti
fixed: Pushed to GHC: […] and base: […]
12:23 PM Changeset in ghc [bb5447c]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Ian Lynagh <igloo@…>
More Win64 test fixes
12:23 PM Changeset in testsuite [b5e2a59]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Ian Lynagh <igloo@…>
More Win64 test fixes
12:19 PM Changeset in ghc [79e56d3]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Ian Lynagh <igloo@…>
3837 now passes on Windows: #3929 is fixed
12:19 PM Changeset in testsuite [c4c39be]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Ian Lynagh <igloo@…>
3837 now passes on Windows: #3929 is fixed
12:15 PM Changeset in ghc [5f9c642]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Ian Lynagh <igloo@…>
Some test fixes on Win64
12:15 PM Changeset in testsuite [8fce1d0]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Ian Lynagh <igloo@…>
Some test fixes on Win64
12:12 PM Status/May12 edited by chak
(diff)
12:06 PM Changeset in base [29ef12e]data-proxydbcsghc-7.6ghc-7.8imp-param-classsupercompilertype-reasoningwindows-iocp by Paolo Capriotti <p.capriotti@…>
Use RTS version of getMonotonicNSec on Windows (#6061)
12:06 PM Changeset in ghc [c031aec]wip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/orf by Paolo Capriotti <p.capriotti@…>
Use RTS version of getMonotonicNSec on Windows (#6061)
12:05 PM Changeset in ghc [c0461976]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Paolo Capriotti <p.capriotti@…>
Move getMonotonicUSec from base to the RTS.
11:49 AM Ticket #6084 (Add stg_ap_pnnv and related call patterns) created by SimonMeier
In the implementation of bytestring encoding and decoding libraries, the …
11:45 AM Ticket #3090 (ghc-pkg update should fail if dependent packages might break) closed by simonpj
wontfix: Closing because * cabal install already checks * Sooner or later …
11:35 AM LightweightConcurrency edited by kc
(diff)
11:27 AM Status/SLPJ-Tickets edited by simonpj
(diff)
11:26 AM Status/SLPJ-Tickets edited by simonpj
(diff)
10:42 AM Changeset in ghc [07fadb7a]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Ian Lynagh <igloo@…>
Driver tweak
10:42 AM Changeset in testsuite [bb5cfdf]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Ian Lynagh <igloo@…>
Driver tweak
10:36 AM Ticket #6077 (Respect XDG_CONFIG_HOME) closed by simonmar
duplicate: Thanks; already reported as #5966
9:04 AM Status/May12 edited by gmainland
(diff)
5:31 AM Ticket #6083 (GeneralizedNewtypeDeriving: "deriving ALL") created by clinton
I've found I've wanted to do: newtype X1 = X1 ( X ) deriving ( ALL ); …

May 7, 2012:

10:51 PM Repositories edited by pcapriotti
Updated Cabal, mtl, added transformers (diff)
10:24 PM Changeset in ghc [4614cf6]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Paolo Capriotti <p.capriotti@…>
Adapt to changes in Cabal.
10:24 PM Changeset in ghc [d7bb8cd2]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Paolo Capriotti <p.capriotti@…>
Follow upstream mtl and Cabal, add transformers package (#5958)
9:45 PM Changeset in ghc [041d1bc]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Ian Lynagh <igloo@…>
Add support for .text.startup sections to the Linker We get these in a full build on Win64
7:06 PM Status/May12 edited by danbst
typo (diff)
4:40 PM Changeset in ghc [dd7522c]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-new-flavorghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Simon Peyton Jones <simonpj@…>
Yet another major refactoring of the constraint solver This is the result of Simon and Dimitrios doing a code walk through. There is no change in behaviour, but the structure is much better. Main changes: * Given constraints contain an EvTerm not an EvVar * Correspondingly, TcEvidence is a recursive types that uses EvTerms rather than EvVars * Rename CtFlavor to CtEvidence * Every CtEvidence has a ctev_pred field. And use record fields consistently for CtEvidence * The solved-constraint fields of InertSet (namely inert_solved and inert_solved_funeqs) contain CtEvidence, not Ct There is a long cascade of follow-on changes.
4:14 PM Changeset in ghc [fdc19e6]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Ian Lynagh <igloo@…>
Merge branch 'master' of darcs.haskell.org:/srv/darcs//ghc
3:02 PM Ticket #6082 (Program compiled with 7.4.1 runs many times slower than compiled with ...) created by gchrupala
The morfette program (http://hackage.haskell.org/package/morfette), when …
3:01 PM Ticket #6080 (dph-classes-vseg-fast fails: Vectorisation error) closed by chak
fixed
2:27 PM Changeset in ghc [85ae01b]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Ian Lynagh <igloo@…>
Fix alignment when loading .a files on Win64
2:18 PM Changeset in ghc [ada13856]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Ian Lynagh <igloo@…>
Test for misaligned sections, and fail if we find one
2:17 PM Changeset in ghc [b35ad0a0]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Manuel M T Chakravarty <chak@…>
Fixed #6080
2:17 PM Changeset in testsuite [d6ef6b2]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Manuel M T Chakravarty <chak@…>
Fixed #6080
2:14 PM Changeset in ghc [209e3750]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Manuel M T Chakravarty <chak@…>
Fix #6080 & house keeping in Vectorise.Exp
9:04 AM LightweightConcurrency edited by kc
edits (diff)
8:41 AM LightweightConcurrency edited by kc
edits (diff)
5:13 AM Changeset in ghc [ceb9f7f6]type-nats by Iavor S. Diatchki <iavor.diatchki@…>
Implement a simple version of 'useRule' and redo 'solve' to use the rules.
3:53 AM Changeset in ghc [37838c0c]type-nats by Iavor S. Diatchki <iavor.diatchki@…>
Add forward rules. Convert Proofs back into Coercions.

May 6, 2012:

8:41 PM Status/May12 edited by simonmar
(diff)
8:30 PM Status/May12 created by simonmar
1st draft
8:01 PM Changeset in ghc [fd832ea]type-nats by Iavor S. Diatchki <iavor.diatchki@…>
Make rule instantiation more consistent (and bit more clean-up).
7:49 PM Changeset in ghc [b26a1b3]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Ian Lynagh <igloo@…>
Build fix
7:39 PM Ticket #6058 (trac: can't register a new user) closed by igloo
wontfix: Thanks for the report. This happens when the server is under load. Trying …
7:01 PM Changeset in ghc [fcb19db]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Ian Lynagh <igloo@…>
Follow changes in haskeline I'm not sure if we still need to encode the error, but validate is happy with this fix, at least.
6:42 PM Changeset in ghc [e7b3a157]type-nats by Iavor S. Diatchki <iavor.diatchki@…>
Reorganize file and add some comments.
6:39 PM Changeset in ghc [9949db0a]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Ian Lynagh <igloo@…>
Fix the way we call strip on DLLs On Win64, "install -s" calls a strip that doesn't understand 64bit binaries. For some reason, this means the DLLs end up non-executable, which means executables that use them just segfault.
2:54 PM Changeset in ghc [f09812de]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Ian Lynagh <igloo@…>
Fix RTS DLL references on Win64
1:22 AM Changeset in ghc [1a860da]type-nats by Iavor S. Diatchki <iavor.diatchki@…>
Some more functionality for matching rules.
1:21 AM Changeset in ghc [5872bf5]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Ian Lynagh <igloo@…>
Working towards fixing DLLs on Win64

May 5, 2012:

10:52 PM Changeset in ghc [6dcf0012]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Ian Lynagh <igloo@…>
Install DLLs as executable Otherwise the dynHelloWorld test segfaults when run from an installed compiler on cygwin.
9:11 PM Ticket #6081 (Kind variables not allowed in class instance declarations) created by goldfire
Consider the following code: […] The last line fails to compile, …
9:07 PM Changeset in ghc [3ce738b]type-nats by Iavor S. Diatchki <iavor.diatchki@…>
Some functionality for matching rules directly (rather than generating them)
9:06 PM Changeset in ghc [b6c8c442]type-nats by Iavor S. Diatchki <iavor.diatchki@…>
Remove (<=) as a class, and add (<=?) as a type function.
5:22 PM LightweightConcurrency edited by kc
(diff)
2:24 PM LightweightConcurrency edited by kc
(diff)
2:01 PM LightweightConcurrency edited by kc
More text (diff)
1:54 PM Changeset in ghc [a8787ece]atomicsbetter-ho-cardinalitycardinalitycoerciblecross-compiler-alienlessdata-kind-syntaxdecision-procedureghc-7.6ghc-7.8ghc-lwc2ghc-parmake-gsocimp-param-classlate-dmdlate-lam-liftnew-demand-to-mergenewcgoverlapping-tyfamssilent-sc-argssimdsupercompilertc-arrowstc-untouchablesth-newth-new-7.6ticky-for-all-letstype-holes-branchtype-natstype-nats-simpleunboxed-tuple-arguments2wip/Cabal-1.20wip/T7704wip/T8545-ghc-7.8wip/T8592wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/cbv-conv-thunkwip/common-contextwip/coveritywip/cpr-vs-jpwip/drop-containers-dep-from-thwip/exprAritywip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compatwip/simdwip/th-new by Ian Lynagh <igloo@…>
Use lnats to avoid overflowing when allocating large amounts Stops outofmem segfaulting on Win64
1:53 PM Changeset in ghc [a6ab2290]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Ian Lynagh <igloo@…>
Add Win64 output for outofmem
1:53 PM Changeset in testsuite [01bd03c]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Ian Lynagh <igloo@…>
Add Win64 output for outofmem
1:18 PM LightweightConcurrency edited by kc
Minor edit (diff)
1:17 PM Changeset in ghc [e9e791d4]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Ian Lynagh <igloo@…>
dph-classes-vseg-fast is broken (#6080)
1:17 PM Changeset in testsuite [3201cb4]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Ian Lynagh <igloo@…>
dph-classes-vseg-fast is broken (#6080)
1:16 PM Ticket #6080 (dph-classes-vseg-fast fails: Vectorisation error) created by igloo
[…]
1:12 PM LightweightConcurrency edited by kc
Image added (diff)
1:10 PM Changeset in ghc [c8b715c0]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Ian Lynagh <igloo@…>
Mark SEH tests as broken on Win64 (#6079)
1:10 PM Changeset in testsuite [24afd1d]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Ian Lynagh <igloo@…>
Mark SEH tests as broken on Win64 (#6079)
1:07 PM Ticket #6079 (SEH exception handler not implemented on Win64) created by igloo
In RtsMain.c we only enable BEGIN_CATCH/END_CATCH on Win32. I think …
12:53 PM GHC_LWC_Key.jpg attached to LightweightConcurrency by kc
key_idea
12:07 PM Ticket #6067 (regression: cgrun051(ghci) failing in HEAD) closed by simonpj
fixed
12:05 PM Ticket #6075 (Incorrect interpretation of scoped type variable declaration in GADT ...) closed by simonpj
invalid: The errror mesage (with GHC 7.4) is this: […] And that seems right to …
11:32 AM Changeset in ghc [95978ed0]ghc-7.8wip/Cabal-1.20wip/T8545-ghc-7.8wip/T8776wip/T8995-level-generalisationwip/T9016wip/ampwip/coveritywip/drop-containers-dep-from-thwip/nested-cprwip/orfwip/pattern-synonymswip/recurs-compat by Ian Lynagh <igloo@…>
Fix ghcilink00{1,4} when the path to AR contains spaces
11:32 AM Changeset in testsuite [2a91733]atomicsdata-kind-syntaxghc-7.6overlapping-tyfamstc-untouchablesth-newtype-natstype-nats-simpleunboxed-tuple-arguments2wip/nested-cprwip/pattern-synonymswip/th-new by Ian Lynagh <igloo@…>
Fix ghcilink00{1,4} when the path to AR contains spaces
11:18 AM LightweightConcurrency edited by kc
more edits (diff)
9:13 AM LightweightConcurrency edited by kc
Edits (diff)
12:54 AM Changeset in ghc [cd34194]type-nats by Trevor Elliott <trevor@…>
Fix a copy/paste bug in the comment for the type-nat cancellation axioms
12:53 AM Changeset in ghc [c665136]type-nats by Trevor Elliott <trevor@…>
Start checking left/right cancellation rules for type nats in CoreLint.lhs It would be nice to specify the rules as a datatype, then use an interpreter to verify that the proofs follow the structure that they describe.
Note: See TracTimeline for information about the timeline view.