Declare UTF-8 in HTML layouts; byte order mark on packaged CSVs

Packaged HTML opened from disk had no charset, so browsers guessed
Windows-1252 and em dashes showed as mojibake; Excel does the same with
CSVs that lack a BOM.

Co-Authored-By: Claude Opus 5.5 <noreply@anthropic.com>
This commit is contained in:
rickyandClaude Opus 5.5 committed 2026-11-07 17:24:53 -05:00
1 parent 5471b9317f
commit 350914bddc
5 files changed
+10 -2

No files matched your search

+4 -2
View File
@@ -91,7 +91,8 @@ class RecordsIntegrityTest < ActionDispatch::IntegrationTest
html = zip.text("status-report.html")
assert_includes html, "Maintenance status · 14 CFR 91.417(a)(2)"
refute_includes html, "<script"
assert_includes html, "<style>"
assert_match(/<head>\s*<meta charset="utf-8">/, html) # files opened from disk need it (em dashes etc.)
assert_includes html, "<style>"
end
assert @a.events.exists?(event_name: "Records package downloaded")
@@ -116,7 +117,8 @@ class RecordsIntegrityTest < ActionDispatch::IntegrationTest
engine = create_component(@a, tbo_hours: 2000, hours_since_overhaul: 126, on_condition: true)
path = Rails.root.join("tmp", "pkg-#{SecureRandom.hex(4)}.zip")
RecordsPackage.new(@a, source: "test").write_to(path)
html = Zip::File.open(path) { |z| z.read("status-report.html").force_encoding("UTF-8") }
html, csv = Zip::File.open(path) { |z| [ z.read("status-report.html").force_encoding("UTF-8"), z.read("signatures.csv").force_encoding("UTF-8") ] }
assert csv.start_with?("\uFEFF"), "CSV has a byte order mark for Excel"
File.delete(path)
assert_includes html, "<style>"
refute_includes html, "\\[</head>" # escaped "&" in CSS selectors left intact