Following up one making 'show dependencies' optional.

1 patch for repository http://darcs.net/screened:

patch 4c36c4c72ce4d0d1361b64d3737f7a7a7ed83708
Author: Ben Franksen <ben.franksen@online.de>
Date:   Mon Jan 27 15:39:08 CET 2020
  * move getDeps and SPatchAndDeps from D.P.Depends to D.UI.C.ShowDependencies
  
  This somewhat messy code has no relationship with anything else happening
  inside D.P.Depends. Besides, getting rid of the graphviz dependency may
  involve re-writing at least parts of that code, which will be easier if it
  is concentrated in D.UI.C.ShowDependencies.

Attachments

• move-getdeps-and-spatchanddeps-from-d_p_depends-to-d_ui_c_showdependencies.dpatch
• patch-preview.txt
• unnamed

It was surprisingly easy to re-implement the main functionality from first
principles. I also find it remarkable how much cleaner the code has become.
The output of the new 'show dependencies' command differs in two points: it
uses a hex rendering of the shortened (to 32 bits) patch ID as the node
identifier. It also contains edges for all direct dependencies which I find
more useful than reducing the graph to avoid transitive edges by arbitrarily
deleting them.

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

patch 4c36c4c72ce4d0d1361b64d3737f7a7a7ed83708
Author: Ben Franksen <ben.franksen@online.de>
Date:   Mon Jan 27 15:39:08 CET 2020
  * move getDeps and SPatchAndDeps from D.P.Depends to D.UI.C.ShowDependencies
  
  This somewhat messy code has no relationship with anything else happening
  inside D.P.Depends. Besides, getting rid of the graphviz dependency may
  involve re-writing at least parts of that code, which will be easier if it
  is concentrated in D.UI.C.ShowDependencies.

patch 3a8e57ef9fed776f62a3538f8842b6593546e368
Author: Ben Franksen <ben.franksen@online.de>
Date:   Tue Jan 28 22:37:44 CET 2020
  * remove graphviz dependency by re-implementing show dependencies

Attachments

• move-getdeps-and-spatchanddeps-from-d_p_depends-to-d_ui_c_showdependencies.dpatch
• patch-preview.txt
• unnamed

I'm having some issues interpreting the show dependencies output of this log.

It looks like the `re-enable XWayland support` depends on the `use `pkg-config` to locate wayland protocol files` twice since 
`409bbad0ff6955c4fa9bda58ccda123e58dc103b` is also the hash of that patch. 
patch eaeab0643fa1c7df014fc781293ff644b3fa4828
Author: raichoo@acmelabs.space
Date:   Fri Feb  7 22:18:23 CET 2020
  * re-enable XWayland support
  
  XWayland support has not been maintained for quite a while so we need to make it
  work in the new infrastructure. This patch introduces the view method `move_resize`
  which allows to move and resize a view atomically, this is needed for properly
  handling XWayland views.

patch a282380414bd498de8428bfdfc3cf510367adc40
Author: raichoo@acmelabs.space
Date:   Fri Feb  7 20:09:27 CET 2020
  * fix indicator damage on maximize
  
  When views are small enough so that their damage does not cover the indicator
  the latter can leave artifacts.

patch 409bbad0ff6955c4fa9bda58ccda123e58dc103b
Author: raichoo@acmelabs.space
Date:   Fri Feb  7 13:34:33 CET 2020
  * use `pkg-config` to locate wayland protocol files

patch 417a707301b75d99f796dd508edbde87f5c0fb7a
Author: raichoo@acmelabs.space
Date:   Thu Feb  6 21:47:16 CET 2020
  * clarify hikari support

patch 15a7a6359d76fa54f59e4fd0704a4986d4352f2d
Author: raichoo@acmelabs.space
Date:   Thu Feb  6 20:59:16 CET 2020
  tagged 1.0.0-alpha2

Attachments

• dependencies.png

Oops. I did not realize that the patch hashes need to be quoted to be
valid node identifiers in the dot language. Will fix.

Attachments

• pEpkey.asc

1 patch for repository http://darcs.net/screened:

patch 1fb9819b0d237af77f07387c98b95d59d758dc98
Author: Ben Franksen <ben.franksen@online.de>
Date:   Sun Feb  9 22:28:09 CET 2020
  * fix: need to quote hashes in dot output

Attachments

• fix_-need-to-quote-hashes-in-dot-output.dpatch
• patch-preview.txt
• unnamed

raichoo, is this working ok for you now?

I confirmed with raichoo on IRC that it's good now.

Yeah, sorry. Your reply got lost in my overflowing inbox :/

The code looks reasonable. I'm a bit surprised we don't have any tests
of the output, but that was already the case before your changes.