A number of small improvements all more or less related to rebase. The
last one is independent, but rebase suspend -v was the motivation.

5 patches for repository http://darcs.net/screened:

patch e4be0d0b5192eee51dbd5810c61e1f016b007781
Author: Ben Franksen <ben.franksen@online.de>
Date:   Sat Apr 22 02:16:48 CEST 2017
  * rebase: print status messages to stderr to avoid messing up shell
completion

patch 0c5e01cb43ec1edb07995f6b918f6a93ceadf9fc
Author: Ben Franksen <ben.franksen@online.de>
Date:   Thu Aug 17 10:06:13 CEST 2017
  * rebase suspend: use red text when verbosely displaying conflicting hunks

patch af208c8b7b8fc56da514a034248529fc229438fe
Author: Ben Franksen <ben.franksen@online.de>
Date:   Thu Aug 17 10:08:38 CEST 2017
  * rebase suspend: don't display hint to use --verbose when already in
effect

patch bc3036e2e41e4ba603dc2382cb89191b96e94da2
Author: Ben Franksen <ben.franksen@online.de>
Date:   Fri Aug 18 11:13:05 CEST 2017
  * rebase fix: suppress --verbose when adding the rebase patch
  
  Otherwise tentativelyAddPatch will show the rebase patch when we run with
  --verbose.

patch 949c03e67db7d11d61073e0f41fb32f0d58b0f8f
Author: Ben Franksen <ben.franksen@online.de>
Date:   Thu Aug 17 10:02:11 CEST 2017
  * use fancyPrinters in putVerbose and putInfo
  
  These are for user display, so we want coloring if available.
  For instance, rebase suspend -v now shows conflicting hunks colored.

Attachments

• rebase_-print-status-messages-to-stderr-to-avoid-messing-up-shell-completion.dpatch

Added a second bundle with a single patch that fixes the rebase tests so
that they no longer expect the rebase status message on stdout but on
stderr.

Attachments

• rebase-tests_-no-longer-expect-rebase-status-from-stdout.dpatch

Seem all good, accepted.