fix: Prevent duplicate HTML IDs on photo page

[larseggert]

Fixes #6355

[codecov]

Codecov Report

Merging #6356 (f38b2e2) into main (32bb4ef) will decrease coverage by 0.01%.
Report is 35 commits behind head on main.
The diff coverage is 87.83%.

@@            Coverage Diff             @@
##             main    #6356      +/-   ##
==========================================
- Coverage   88.69%   88.68%   -0.01%     
==========================================
  Files         290      290              
  Lines       40390    40422      +32     
==========================================
+ Hits        35823    35848      +25     
- Misses       4567     4574       +7     

Files Changed	Coverage Δ
ietf/celeryapp.py	0.00% <0.00%> (ø)
ietf/help/views.py	63.63% <ø> (+8.46%)	⬆️
ietf/ipr/models.py	88.95% <ø> (ø)
ietf/urls.py	80.00% <ø> (-0.56%)	⬇️
ietf/ipr/views.py	87.96% <95.45%> (+0.24%)	⬆️
ietf/doc/models.py	88.98% <100.00%> (ø)
ietf/doc/utils.py	87.38% <100.00%> (+0.23%)	⬆️
ietf/doc/views_review.py	95.16% <100.00%> (+0.20%)	⬆️
ietf/doc/views_search.py	89.09% <100.00%> (-0.21%)	⬇️
ietf/ipr/feeds.py	93.33% <100.00%> (+0.22%)	⬆️
... and 3 more

... and 2 files with indirect coverage changes

[rjsparks]

This will break anyone that is used to typing http://localhost:8000/wg/photos/#h to get to the h's.
http://localhost:8000/wg/photos/#h8 is just not something they're going to guess.

It also just pushes the problem down the road (this problem exists on the current page) on doing the right thing with real names. Someone with a non [a-zA-Z] first character will end up with a section that is labelled with an [a-zA-Z] heading that is sorted in nothing matching any standardized collation. Are ids restricted to [a-zA-Z]? Could we just use the first unicode character instead and make sure the results are sorted with the system's collation setting? This template is only used for working group chairs, so we're not likely to have random garbage thrown at production.

[larseggert]

We could. Let me see if I can rework this.

[larseggert]

Turns out HTML5 IDs can be Unicode :-)