Fix #311: resolve log path against filteredRuns to support local time mode

[cj36457]

Closes #311.

Problem

When the "convert to local timezone" toggle is on, clicking a run on any chart shows:

Log file error: this output didn't have a path in the database so the log file cannot be found!

…even though path is non-empty in the DB and log.html exists on disk.

Root cause

In js/log.js open_log_file, the lookup compared the clicked label against the raw runs global:

var output = runs.find(run => run.run_start.slice(0, 19) === runStart.slice(0, 19))

But runs[i].run_start holds the original UTC string from the DB, while chart labels (and the clicked value) come from filteredRuns, whose run_start has been rewritten by convert_timezone / remove_timezones in filter.js to match the displayed local time. The substring compare never matches → output = undefined → path = "" → alert.

Fix

Look up against filteredRuns first (whose run_start mirrors what the chart actually shows), then fall back to the raw runs array so filtered-out runs still resolve. The downstream update_log_path_with_id path keeps using raw runs/suites/tests — they share the original UTC timestamps internally so suite/test id lookup remains consistent.

Evidence

Added tests/javascript/log.test.js with 4 cases. Reverting just log.js while keeping the new test reproduces the issue's exact alert text:

FAIL  tests/javascript/log.test.js > open_log_file (issue #311) > resolves log path against filteredRuns when timezone is converted to local time
  Received:
    1st vi.fn() call: [ "Log file error: this output didn't have a path in the database so the log file cannot be found!" ]

With the fix applied:

✓ resolves log path against filteredRuns when timezone is converted to local time
✓ still resolves correctly in UTC mode (filteredRuns == runs run_start)
✓ falls back to raw runs when the clicked run was filtered out
✓ alerts only when no matching run exists anywhere

Test Files  12 passed (12)
Tests       232 passed (232)

Full JS suite (npm run test:js) green — no regressions.

[timdegroot1996]

Ah you already started an implementation. I will look it over and add some comments if I have any. If you make any updates be sure to ping me so I can trigger the tests. Since right now you are a first time contributor I have to manually start them. The next PR will start them automatically 👍