+1 maintenance report

[Bryce Harrington]

Here's some of the things I worked on for my +1 maintenance this week.

* Packages where only a single arch failed, I retriggered test runs.
  Many of these required additional triggers to include icu, gcc-10, etc.
  - sleekxmpp
  - jinja2
  - lua-luv
  - libhttp-dav-perl
  - node-jquery
  - gegl
  - bluez-qt
  - Bunch more, lost track...

* Went through packages listing 'missing builds' and re-triggered
  ones where the logs showed connection errors or other irregularities
  that might be build system errors.  Quite a few of these passed.

* Rebuilt / retested all packages with a riscv failure more than 30 days
  old.  It's relatively new, figured things change fast.  Over half of
  these passed.

* php-horde-*: Most of this stack has been blocked from being synced,
  but there were four more trying to sync.  These four are added now.
  (If any other php-horde-* packages show up, please add them as well.)

* openclipart: Debian is on libreoffice 7.x but we're on 6.x;
  openclipart was trying to depend on a new binary package from 7.x.  I
  restored that back to the 6.x-era dependency, and this migrated.

* nodejs: There was a syntax warning in an awk statement that was
  causing the autopkgtest's lint check to fail.  I fixed the warning;
  hopefully it will pass now.

* node-*: There were numerous autopkgtest regressions that were fixed by
  re-triggering against nodejs, icu, pegjs and various combinations of other
  dependencies.
  - node-chokidar
  - node-formidable
  - node-json-buffer
  - node-kew
  - node-clean-css
  - node-oauth-sign
  - node-opencv

### Still Needing Attention ###

Some notes for remaining work:

* node-node-sass:  Hits a TypeError, still needs investigated.

* node-sha.js, node-create-hash, node-crypto-browserfy:
  - See LP: #1887144

* node-xterm:  I think this needs updated to 4.x (see deb #964170).
  Debian (and Ubuntu) are on 3.8.  There seems to be some type conflict
  between Window and DOMWindow; might be due to API changes in nodejs?

* python-hypothesis: It errors on unreliable test timings.  The package
  seems to have a history of this.  I retried a couple times, but maybe
  the test should be skipped as flaky?

* gubbins: This is being built on arch: All, but depends on iqtree which
  is built only for a few arch's, so gubbins fails on other arch's.  I
  tried building iqtree with arch: All but no such luck.  Not sure what
  to do here...