Hi Simon,

for me it happens on many files that are compiled with

 $ make GhcStage2HcOpts="-O1 -g"

while the stage1 compiler is active.

Tomorrow I can add a concrete invocation.

Cheers,

     Gabor

Em seg, 21 de ago de 2017 às 17:49, Simon Peyton Jones <simonpj@microsoft.com> escreveu:
|  Oh dear, then I've really misunderstood the intent of this warning. I've
|  been noticing these during the GHC build for quite some time now.

Really?

After stumbling on #14142, I removed -dno-debug-output from mk/flavours/validate.mk and re-validated.  Not a single occurrence of

|  > | WARNING: file compiler/simplCore/OccurAnal.hs, line 2695
|  > |   OccurAnal failed to rediscover join point(s): [go58_akKa]

How can I reproduce what you are seeing?

Simon

|  -----Original Message-----
|  From: Ben Gamari [mailto:ben@smart-cactus.org]
|  Sent: 19 August 2017 16:54
|  To: Simon Peyton Jones <simonpj@microsoft.com>; Gabor Greif
|  <ggreif@gmail.com>
|  Cc: ghc-devs <ghc-devs@haskell.org>
|  Subject: RE: Stg-lint and bootstrapping (fails?)
|
|  Simon Peyton Jones <simonpj@microsoft.com> writes:
|
|  > | WARNING: file compiler/simplCore/OccurAnal.hs, line 2695
|  > |   OccurAnal failed to rediscover join point(s): [go58_akKa]
|  >
|  > That shouldn't happen!  If you can make a repro case, please open a
|  ticket.
|  >
|  Oh dear, then I've really misunderstood the intent of this warning. I've
|  been noticing these during the GHC build for quite some time now.
|
|  Cheers,
|
|  - Ben