Re: module graphing

2007-03-23 Thread Malcolm Wallace
Simon Peyton-Jones <[EMAIL PROTECTED]> wrote: > Module cycles found: > -- Cycle 1 -- > GHC.Base {-# SOURCE #-} imports GHC.Err > Data.Either imports GHC.Base > Data.Tuple imports GHC.Base > ... FWIW, hmake reports the same cycles for me:

Re: module graphing

2007-03-22 Thread Sven Panne
On Thursday 22 March 2007 18:01, Simon Peyton-Jones wrote: > Each blob reported as a "cycle" is a strongly connected component. You can > get from every module in the cycle to every other module in the cycle. OK, but then the flag and the output should not talk about "cycles", but "SCCs". Curren

RE: module graphing

2007-03-22 Thread Simon Peyton-Jones
: Simon Peyton-Jones | Cc: cvs-ghc@haskell.org; Ian Lynagh | Subject: Re: module graphing | | On Thursday 22 March 2007 15:20, Simon Peyton-Jones wrote: | > It is topologically sorted! | | Hmmm, then I don't understand the underlying partial order... What does it | mean when a module is ment

Re: module graphing

2007-03-22 Thread Sven Panne
On Thursday 22 March 2007 15:20, Simon Peyton-Jones wrote: > It is topologically sorted! Hmmm, then I don't understand the underlying partial order... What does it mean when a module is mentioned later than another one? Or are several cycles lumped together? Let's take the first part of the outp

RE: module graphing

2007-03-22 Thread Simon Peyton-Jones
It is topologically sorted! | -Original Message- | From: Sven Panne [mailto:[EMAIL PROTECTED] | Sent: 22 March 2007 12:51 | To: cvs-ghc@haskell.org | Cc: Simon Peyton-Jones; Ian Lynagh | Subject: Re: module graphing | | On Thursday 22 March 2007 13:32, Simon Peyton-Jones wrote: | > I

Re: module graphing

2007-03-22 Thread Sven Panne
On Thursday 22 March 2007 13:32, Simon Peyton-Jones wrote: > I've just added a flag -ddump-mod-cycles to the behaviour of ghc -M. It's > specialised for displaying module cycles. Here's sample output for 'base' > [...] Great! That was fast... One further suggestion: If it doesn't involve too much

RE: module graphing

2007-03-22 Thread Simon Peyton-Jones
I've just added a flag -ddump-mod-cycles to the behaviour of ghc -M. It's specialised for displaying module cycles. Here's sample output for 'base' Open question: should the flag be "-ddump-mod-cycles" or "-optdep-dump-mod-cycles"? For this reason I have not yet added to the documentation. Y

Re: module graphing

2007-03-21 Thread Sven Panne
On Wednesday 21 March 2007 17:44, Ian Lynagh wrote: > [...] It doesn't help that the tool doesn't currently look for, and colour, > loops (source imports are coloured red, though, so you can spot them at > least). Could it be modified to simply print out all loops in textual form as an option? Th

Re: module graphing

2007-03-21 Thread Ian Lynagh
On Wed, Mar 21, 2007 at 09:40:13AM +, Simon Peyton-Jones wrote: > great! For module loops, a dotted line for the {-SOURCE-} imports would be > good. > > (give me a yell when it's working on GHC itself and the libs) OK, so the commandline below works for base (once it has been built by cabal;

RE: module graphing

2007-03-21 Thread Simon Peyton-Jones
great! For module loops, a dotted line for the {-SOURCE-} imports would be good. We should make it easy to get the tool. (give me a yell when it's working on GHC itself and the libs) S | -Original Message- | From: [EMAIL PROTECTED] [mailto:[EMAIL PROTECTED] On Behalf Of Ian Lynagh | Sen