We should merge this fix to the 7.10 branch.

On Jan 8, 2015 11:52 PM, "Peter Wortmann" <scpmw@leeds.ac.uk> wrote:

(sorry for late answer)

Yes, that's pretty much what this would boil down to. The patch is trivial:

https://github.com/scpmw/ghc/commit/29acc#diff-1

I think this is a good idea anyways. We can always re-introduce the data for higher -g<n> levels.

Greetings,
  Peter


On 05/01/2015 00:59, Johan Tibell wrote:
What about keeping exactly what -g1 keeps for gcc (i.e. functions,
external variables, and line number tables)?

On Sun, Jan 4, 2015 at 5:48 PM, Peter Wortmann <scpmw@leeds.ac.uk
<mailto:scpmw@leeds.ac.uk>> wrote:



    Okay, I ran a little experiment - here's the size of the debug
    sections that Fission would keep (for base library):

       .debug_abbrev:          8932 - 0.06%
       .debug_line:          374134 - 2.6%
       .debug_frame:         671200 - 4.5%

    Not that much. On the other hand, .debug_info is a significant
    contributor:

       .debug_info(full):   4527391 - 30%

    Here's what this contains: All procs get a corresponding DWARF
    entry, and we declare all Cmm blocks as "lexical blocks". The latter
    isn't actually required right now - to my knowledge, GDB simply
    ignores it, while LLDB shows it as "inlined" routines. In either
    case, it just shows yet more GHC-generated names, so it's really
    only useful for profiling tools that know Cmm block names.

    So here's what we get if we strip out block information:

       .debug_info(!block): 1688410 - 11%

    This eliminates a good chunk of information, and might therefore be
    a good idea for "-g1" at minimum. If we want this as default for
    7.10, this would make the total overhead about 18%. Acceptable? I
    can supply a patch if needed.

    Just for comparison - for Fission we'd strip proc records as well,
    which would cause even more extreme savings:

       .debug_info(!proc):    36081 - 0.2%

    At this point the overhead would be just about 7% - but without
    doing Fission properly this would most certainly affect debuggers.

    Greetings,
       Peter

    On 03/01/2015 21:22, Johan Tibell wrote:
    > How much debug info (as a percentage) do we currently generate? Could we just keep it in there in the release?

    _________________________________________________
    ghc-devs mailing list
    ghc-devs@haskell.org <mailto:ghc-devs@haskell.org>
    http://www.haskell.org/__mailman/listinfo/ghc-devs
    <http://www.haskell.org/mailman/listinfo/ghc-devs>




_______________________________________________
ghc-devs mailing list
ghc-devs@haskell.org
http://www.haskell.org/mailman/listinfo/ghc-devs


_______________________________________________
ghc-devs mailing list
ghc-devs@haskell.org
http://www.haskell.org/mailman/listinfo/ghc-devs