[chore] Fix todo and fixme tags (#8365)

Slightly overlooked in #8312.

frontend/apps/reader/modules/readerhighlight.lua

@@ -1072,7 +1072,7 @@ function ReaderHighlight:lookup(selected_text, selected_link)
             for _, sbox in ipairs(selected_text.sboxes) do
                 local word = self.ui.document:getOCRWord(self.hold_pos.page, { sbox = sbox })
                 logger.dbg("OCRed word:", word)
-                -- @fixme This might produce incorrect results on RTL text.
+                --- @fixme This might produce incorrect results on RTL text.
                 if word and word ~= "" then
                     text = text .. word
                 end

plugins/japanese.koplugin/deinflector.lua

@@ -435,7 +435,7 @@ function Deinflector:init()
                                     DEFAULT_TEXT_CONVERSIONS
     if self.rules ~= nil then return end -- already loaded
 
-    -- @todo Maybe make this location configurable?
+    --- @todo Maybe make this location configurable?
     inflections = parsePluginJson("yomichan-deinflect.json")
 
     -- Normalise the reasons and convert the rules to the rule_types bitflags.

plugins/japanese.koplugin/main.lua

@@ -68,7 +68,7 @@ function Japanese:onWordLookup(args)
         return
     end
 
-    -- @todo Try to repeatedly reduce the text and deinflect the shortened text
+    --- @todo Try to repeatedly reduce the text and deinflect the shortened text
     --       to provide more candidates. This is particularly needed because
     --       JMDict has a habit of creating entries for compounds or phrases
     --       that do not exist in monolingual dictionaries (even in 大辞林 or
@@ -80,7 +80,7 @@ function Japanese:onWordLookup(args)
     results = self.deinflector:deinflect(text)
     logger.dbg("japanese.koplugin: deinflection of", text, "results:", results)
 
-    -- @todo Pass up the reasons list (formatted Yomichan style) to the
+    --- @todo Pass up the reasons list (formatted Yomichan style) to the
     --       dictionary pop-up so you can get some more information about the
     --       inflection. But this would require adding some kind of tag
     --       metadata that we have to pass through from the lookup to the
@@ -185,7 +185,7 @@ function Japanese:onWordSelection(args)
     -- Calling sdcv is fairly expensive, so reduce the cost by trying every
     -- candidate in one shot and then picking the longest one which gave us a
     -- result.
-    -- @todo Given there is a limit to how many command-line arguments you can
+    --- @todo Given there is a limit to how many command-line arguments you can
     --       pass, we should split up the candidate list if it's too long.
     local best_word
     local cancelled, all_results = self.dictionary:rawSdcv(all_words)

spec/unit/readerdictionary_spec.lua

@@ -52,7 +52,7 @@ describe("Readerdictionary module", function()
         assert.spy(s).was_called()
         assert.spy(s).was_called_with(match.is_ref(readerui.languagesupport), word)
         if readerui.languagesupport.plugins["japanese_support"] then
-            -- @todo This should probably check against a set or sorted list
+            --- @todo This should probably check against a set or sorted list
             --       of the candidates we'd expect.
             assert.spy(s).was_returned_with(match.is_not_nil())
         end