GitLab

The test now actually passes

- as ghc 6.9 flags are beginning to diverge, we need
  a simpler way of specifying version-specific flags
  or settings: have two entirely separate sections,
  instead of one section with too many patches.

  even if that doubles the script size, it should
  be easier to edit. also add a warning that 
  ghci024.stdout is a generated file.

- default for -fprint-bind-result has changed

- -fno-ignore-breakpoints is gone

- -fprint-bind-contents is new

 
 - for unregisterised platforms, default is '-fno-asm-mangling'
 - powerpc_apple_darwin fails on ':set -package ghc' (#1845)
 - for ghc 6.8, -fno-run-cps and -fno-convert-to-zipper-and-back
   do not exist 

Test is drvfail013

these tests validate on win/xp.

  - if any issues with dos-style line-endings remain,
    please apply dos2unix for now, and let me know

  - ghci024 has ghci024.py instead of ghci024.stdout;
    the prepare024 function defined in there generates
    a platform- and version-specific ghci024.stdout
    for the test (see all.T)

  - ghci025 uses the (non-advertised) new -s option
    to get sorted :browse output for stable tests
    (to work around #1799). if this isn't sufficient,
    please set normal to ignore_output in all.T for
    now (until #1799 is fixed), and let me know

Revert to getting an error from the renamer