Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 57.23%. Comparing base (d191997) to head (4a98986).
Report is 351 commits behind head on main.

@@            Coverage Diff             @@
##             main    #6349      +/-   ##
==========================================
- Coverage   57.23%   57.23%   -0.01%     
==========================================
  Files         507      507              
  Lines       31406    31406              
  Branches     4690     4690              
==========================================
- Hits        17976    17975       -1     
  Misses      12575    12575              
- Partials      855      856       +1     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

• Would quoting the label be too extra, as in [ "some node" ] (hopefully not)?
• Would it be a bad idea to put the label to the end of the line, so that varying length labels will not make the lines look weird? For example

[ some_label ] y = Add (x, y)
[ some_extra_extra_long_label__some_scope_1__some_scope_2 ] z = some::SomeNodeNames(y)

Alternatively putting them at the back:

y = Add (x, y) [ some_label ]
z = some::SomeNodeNames(y) [ some_extra_extra_long_label__some_scope_1__some_scope_2 ]

Or does this syntax support line breaks, so

[ some_label ]
y = Add (x, y)
[ some_extra_extra_long_label__some_scope_1__some_scope_2 ]
z = some::SomeNodeNames(y)

is valid?

• Would quoting the label be too extra, as in [ "some node" ] (hopefully not)?

Yes ... that is precisely the choice/tradeoff discussed between (a) and (b). Option (b) avoids it, at the cost of including any extra blanks before or after the label as part of the label. (Though we could skip that, but seems a bit complicated).

Yes, line breaks can be used.

Putting the labels at the end: hmmm. It has the advantage you mention. But it seems less natural (convention typically puts labels before the statement; and, with if/while this would push the label to appear after the entire if/while body). I am inclined to leave it as is (visualization tools can always choose to present it differently, with alignment/omission etc. being supported as pretty-printing options).

I see. "some node" is a valid identifier. I missed that part.

I see. "some node" is a valid identifier. I missed that part.

I am unclear what your preference is. I am planning to extend the PR to allow [some node] as opposed to the current requirement of ["some node"]. Please let me know if you think that's not a good idea.

I think it would be more straightforward to just assume a valid identifier so that it is consistent across.

Sorry I see the trade off now: node names may not be a valid identifier…

edit: but still users can still write [ "6" ], which fits well with the fact that node names are strings. I personally feel that option (a) is clean and nice.

Ok, I will leave this as is (using option 1).