and/or between filters

[Prahitha]

The functionality is kind of limited and still needs some work. For example, it assumes that the user will not enter: coverage > 30 AND/OR coverage > 40 AND/OR coverage > 50

I would also like to know if the data structure used here is okay or if I should implement an array with priority like earlier. Any feedback is appreciated. Thank you!

[mglb]

This doesn't seem to work in basic cases:

• (empty/default value) tool is icarus -> full table is visible;
  removing the empty value and leaving either OR or AND as the default should remove this problem.
• (AND) tool is icarus -> OK
• (OR) tool is icarus -> all tools visible with "no matching records"

"Remove all filters" doesn't remove filter from the table. Only filter fields are removed.

Why did you remove support for relation operators in entries (rows)? It has been useful to e.g. search "coverage <= 0 OR >= 100" AND "tool is X".

Code structure seems to be OK at first glance - lets make it work well first and optionally do some small optimization/refactoring later.

[Prahitha]

Thanks for the feedback, I'll work on those issues. Also, remove all filters works fine for me. It takes a awhile to redraw the table though.

Why did you remove support for relation operators in entries (rows)? It has been useful to e.g. search "coverage <= 0 OR >= 100" AND "tool is X".

@tgorochowik suggested that we have only one filter per row. So we agreed to first implement AND/OR between rows as a global relation and then extend it to AND/OR between every row.

[mglb]

Also, remove all filters works fine for me. It takes a awhile to redraw the table though.

Case that don't work in Chrome and Firefox:

• Select OR
• Add
• Select: tool is icarus
• Apply → "No matching records"
• Remove all filters → still "No matching records"

I did wait a few minutes for results.

[Prahitha]

Ah yes, it doesn't seem to work when there are no matching records.

[Prahitha]

I think I'm missing something here. When I print result in the console it shows (coverage, tool, types) => (true/false). I believe this is supposed to only return a boolean though. So, I was wondering what I was doing wrong and wanted some help. Thank you!

[mglb]

Let's break it:

var filter_result = eval(filter_expression) // <- boolean value
var function_code = "(coverage, tool, types) => (" + filter_result + ")";
var result = eval(function_code);

the last line is effectively the same as:

var result = (coverage, tool, types) => (filter_result); // either `... => (true)` or `... => (false)`

So this returns a function and the console prints its code.

(Spoiler alert: hints/solutions for further steps ahead ;) )

As a side note:
I guess filter_expression is a string with javascript expression code that uses coverage, tool, types variables. If so, you actually only need this inner eval (filter_result) to get result of the expression.
However, the whole function code looks like you want to actually generate a filter_cell function - in this case:

• You only need the outer eval.
• Instead of filter_result value, you should use the expression string (filter_expression) itself in the string concatenation.
• The function filter_result doesn't need any of its arguments (or, for readability purposes, could take filter_expression string as an argument). More suitable name for the function would be something like compile_filter_cell_function. It should be used this way: var filter_cell = compile_filter_cell_function(filter_expression).

The difference between two solutions (using the function that does eval of the expression vs compiling this function) is performance - filter_cell will be called on a few hundred (thousand?) cells in a loop, so it's better to run eval once instead of running it for every cell.

[github-actions]

Compared test results

tool	new_failures	new_passes	added	removed	not_affected
Slang	0	0	0	0	4974
Verilator	2	1	0	0	4986
Yosys	0	1	0	0	4973
moore	0	0	0	0	4746
Icarus	0	0	0	0	4655
UhdmYosys	1	1	0	0	4972
Surelog	2	0	0	0	4972
Verible	0	0	0	0	4655
VeribleExtractor	0	0	0	0	4655
sv_parser	0	0	0	0	4746
Sv2v_zachjs	2	0	0	0	4972
tree_sitter_verilog	0	0	0	0	4655
Odin	0	0	0	0	4746
UhdmVerilator	0	0	0	0	4989

Download an archive containing all the details