print: comment positions (same line or next line) are now preserved

FAQ.md

@@ -141,17 +141,12 @@ and ledger if you avoid ledger's most advanced features.
 Some ledger features are parsed but ignored, eg:
 
 - automated transactions ( = ... , ~ ... )
-- balance assertions ( AMT1=AMT2 )
 - fixed lot prices ( {...} )
 - historical prices ( P ... )
 
-Some features are not currently parsed and will cause an error, eg:
+Some features are not currently parsed and will cause an error, eg
-
+ledger's more recent top-level directives. There can also be subtle
-- balance assignments
+differences in parser behaviour.
-- some top level directives like "account"
-
-There can also be subtle differences in parser behaviour, eg
-comments may be permissible in different places. 
 
 ### What other functionality differences are there ?
 
@@ -184,10 +179,11 @@ comments may be permissible in different places.
   ledger print shows only the secondary date with --aux-date, but not
   vice versa.
 
-- hledger's default commodity directive (D) sets the commodity for
+- hledger's default commodity directive (D) sets the commodity to be
-  subsequent commodityless amounts, and sets that commodity's display
+  used for subsequent commodityless amounts, and also sets that
-  settings if such an amount is the first seen. ledger uses D only for
+  commodity's display settings if such an amount is the first
-  commodity display settings and for the entry command.
+  seen. ledger uses D only for commodity display settings and for the
+  entry command.
 
 - hledger generates a description for timelog sessions, instead of
   taking it from the clock-out entry

MANUAL.md

@@ -685,10 +685,10 @@ The most basic reporting commands are `print`, `register` and `balance`:
 
 #### print
 
-The print command displays full transactions from the journal file, tidily
+The print command displays full transactions from the journal file,
-formatted and showing all amounts explicitly. The output of print is
+tidily formatted and showing all amounts explicitly. The output of
-always a valid hledger journal, but it might not preserve the original
+print is always a valid hledger journal, but it does always not
-content absolutely intact (eg comments.)
+preserve all original content exactly (eg directives).
 
 hledger's print command also shows all unit prices in effect, or (with
 -B/--cost) shows cost amounts.

NEWS.md

@@ -6,6 +6,7 @@ title: hledger news
 
 ## unreleased
 
+- print: comment positions (same line or next line) are now preserved
 - register: `--average/-A` shows a running average, like ledger
 - queries: `sym:REGEXP` matches (whole) commodity symbols
 - queries: `amt` now uses the = operator by default, eg amt:50 finds amounts equal to 50

hledger-lib/Hledger/Data/Transaction.hs

@@ -106,19 +106,17 @@ tests_showTransactionUnelided = [
           pstatus=True,
           paccount="a",
           pamount=Mixed [usd 1, hrs 2],
-          pcomment="pcomment1\npcomment2\n",
+          pcomment="\npcomment2\n",
           ptype=RegularPosting,
           ptags=[("ptag1","val1"),("ptag2","val2")]
           }
        ]
       }
       `gives` unlines [
-      "2012/05/14=2012/05/15 (code) desc",
+      "2012/05/14=2012/05/15 (code) desc    ; tcomment1",
-      "    ;tcomment1",
       "    ; tcomment2",
       "                $1.00",
       "    * a          2.0h",
-      "    ;pcomment1",
       "    ; pcomment2",
       ""
       ]
@@ -128,29 +126,38 @@ tests_showTransactionUnelided = [
 showTransaction' :: Bool -> Transaction -> String
 showTransaction' elide t =
     unlines $ [descriptionline]
-              ++ multilinecomment
+              ++ newlinecomments
               ++ (postingsAsLines elide t (tpostings t))
               ++ [""]
     where
-      descriptionline = rstrip $ concat [date, status, code, desc, inlinecomment]
+      descriptionline = rstrip $ concat [date, status, code, desc, samelinecomment]
       date = showdate (tdate t) ++ maybe "" showedate (tdate2 t)
       showdate = printf "%-10s" . showDate
       showedate = printf "=%s" . showdate
       status = if tstatus t then " *" else ""
       code = if length (tcode t) > 0 then printf " (%s)" $ tcode t else ""
       desc = if null d then "" else " " ++ d where d = tdescription t
-      (inlinecomment, multilinecomment) = commentLines $ tcomment t
+      (samelinecomment, newlinecomments) =
+        case renderCommentLines (tcomment t) of []   -> ("",[])
+                                                c:cs -> (c,cs)
 
--- Render a transaction or posting's comment as indented, semicolon-prefixed comment lines -
+-- Render a transaction or posting's comment as indented, semicolon-prefixed comment lines.
--- an inline comment (when it's a single line) or multiple lines.
+renderCommentLines :: String -> [String]
-commentLines :: String -> (String, [String])
+renderCommentLines s  = case lines s of ("":ls) -> "":map commentprefix ls
-commentLines s
+                                        ls      -> map commentprefix ls
-    | null s = ("", [])
-    | length ls == 1 = (prefix $ head ls, [])
-    | otherwise = ("", (prefix $ head ls):(map prefix $ tail ls))
     where
-      ls = lines s
+      commentprefix = indent . ("; "++)
-      prefix = indent . (";"++)
+
+-- -- Render a transaction or posting's comment as semicolon-prefixed comment lines -
+-- -- an inline (same-line) comment if it's a single line, otherwise multiple indented lines.
+-- commentLines' :: String -> (String, [String])
+-- commentLines' s
+--     | null s = ("", [])
+--     | length ls == 1 = (prefix $ head ls, [])
+--     | otherwise = ("", (prefix $ head ls):(map prefix $ tail ls))
+--     where
+--       ls = lines s
+--       prefix = indent . (";"++)
 
 postingsAsLines :: Bool -> Transaction -> [Posting] -> [String]
 postingsAsLines elide t ps
@@ -161,11 +168,13 @@ postingsAsLines elide t ps
 postingAsLines :: Bool -> [Posting] -> Posting -> [String]
 postingAsLines elideamount ps p =
     postinglines
-    ++ multilinecomment
+    ++ newlinecomments
   where
-    postinglines = map rstrip $ lines $ concatTopPadded [showacct p, "  ", amount, inlinecomment]
+    postinglines = map rstrip $ lines $ concatTopPadded [showacct p, "  ", amount, samelinecomment]
     amount = if elideamount then "" else showamt (pamount p)
-    (inlinecomment, multilinecomment) = commentLines $ pcomment p
+    (samelinecomment, newlinecomments) =
+      case renderCommentLines (pcomment p) of []   -> ("",[])
+                                              c:cs -> (c,cs)
     showacct p =
       indent $ showstatus p ++ printf (printf "%%-%ds" w) (showAccountName Nothing (ptype p) (paccount p))
         where
@@ -188,8 +197,7 @@ tests_postingAsLines = [
       }
      `gives` [
       "                $1.00",
-      "    * a          2.0h",
+      "    * a          2.0h    ; pcomment1",
-      "    ;pcomment1",
       "    ; pcomment2",
       "    ;   tag3: val3  "
       ]

hledger-lib/Hledger/Data/Types.hs

@@ -73,7 +73,7 @@ newtype MixedAmount = Mixed [Amount] deriving (Eq,Ord)
 data PostingType = RegularPosting | VirtualPosting | BalancedVirtualPosting
                    deriving (Eq,Show)
 
-type Tag = (String, String)
+type Tag = (String, String)  -- ^ A tag name and (possibly empty) value.
 
 data Posting = Posting {
       pdate :: Maybe Day,  -- ^ this posting's date, if different from the transaction's
@@ -81,9 +81,9 @@ data Posting = Posting {
       pstatus :: Bool,
       paccount :: AccountName,
       pamount :: MixedAmount,
-      pcomment :: String, -- ^ this posting's non-tag comment lines, as a single non-indented string
+      pcomment :: String, -- ^ this posting's comment lines, as a single non-indented multi-line string
       ptype :: PostingType,
-      ptags :: [Tag],
+      ptags :: [Tag], -- ^ tag names and values, extracted from the comment
       pbalanceassertion :: Maybe MixedAmount,  -- ^ optional: the expected balance in the account after this posting
       ptransaction :: Maybe Transaction    -- ^ this posting's parent transaction (co-recursive types).
                                            -- Tying this knot gets tedious, Maybe makes it easier/optional.
@@ -100,10 +100,10 @@ data Transaction = Transaction {
       tstatus :: Bool,  -- XXX tcleared ?
       tcode :: String,
       tdescription :: String,
-      tcomment :: String, -- ^ this transaction's non-tag comment lines, as a single non-indented string
+      tcomment :: String, -- ^ this transaction's comment lines, as a single non-indented multi-line string
-      ttags :: [Tag],
+      ttags :: [Tag], -- ^ tag names and values, extracted from the comment
       tpostings :: [Posting],            -- ^ this transaction's postings
-      tpreceding_comment_lines :: String
+      tpreceding_comment_lines :: String -- ^ any comment lines immediately preceding this transaction
     } deriving (Eq)
 
 data ModifierTransaction = ModifierTransaction {

hledger-lib/Hledger/Read/JournalReader.hs

@@ -817,19 +817,15 @@ emptyline = do many spacenonewline
 followingcomment :: GenParser Char JournalContext String
 followingcomment =
   -- ptrace "followingcomment"
-  (do first <- many spacenonewline >> followingcommentline
+  do samelinecomment <- many spacenonewline >> (try commentline <|> (newline >> return ""))
-      rest <- many (try (many1 spacenonewline >> followingcommentline))
+     newlinecomments <- many (try (many1 spacenonewline >> commentline))
-      return $ unlines $ first:rest
+     return $ unlines $ samelinecomment:newlinecomments
-  ) <|>
-  do
-    many spacenonewline >> newline
-    rest <- many (try (many1 spacenonewline >> followingcommentline))    
-    return $ unlines rest
 
-followingcommentline :: GenParser Char JournalContext String
+commentline :: GenParser Char JournalContext String
-followingcommentline = do
+commentline = do
-  -- ptrace "followingcommentline"
+  -- ptrace "commentline"
   char ';'
+  many spacenonewline
   l <- anyChar `manyTill` eolof
   optional newline
   return l

tests/comments.test

@@ -31,24 +31,28 @@ hledgerdev -f - print
 
 >>>=0
 
-# 3. print should preserve comments
+# 3. print should preserve transaction (entry) comments and which line they're on
 hledgerdev -f - print
 <<<
-; isolated journal comment
+; leading journal comment, not preserved
 
-; pre-transaction journal comment
+; transaction preceding comment, not preserved
-2009/1/1 x    ; transaction comment
+2009/1/1 x ; transaction same line comment
-    a  1    ; posting 1 comment
+    ; transaction new line comment
-    ; posting 1 comment 2
+    a  1    ; posting 1 same line comment
+    ; posting 1 new line comment
     b
-    ; posting 2 comment
+      ; posting 2 new line comment
-; post-transaction journal comment
+; journal comment right after the transaction, not preserved
+
+; trailing journal comment, not preserved
 >>>
-2009/01/01 x    ; transaction comment
+2009/01/01 x    ; transaction same line comment
-    a             1
+    ; transaction new line comment
-    ; posting 1 comment
+    a             1    ; posting 1 same line comment
-    ; posting 1 comment 2
+    ; posting 1 new line comment
-    b            -1    ; posting 2 comment
+    b            -1
+    ; posting 2 new line comment
 
 >>>2
 >>>=0

tests/tags.test

@@ -12,14 +12,12 @@ hledgerdev -f - print
   ; posting-2-tag-2:
 ; non-metadata:
 >>>
-2010/01/01
+2010/01/01    ; txntag1: txn val 1
-    ; txntag1: txn val 1
     ; txntag2: txn val 2
     a             1
     ; posting1tag1: posting 1 val 1
     ; posting1tag2: 
-    b            -1
+    b            -1    ; posting-2-tag-1: posting 2 val 1
-    ; posting-2-tag-1: posting 2 val 1
     ; posting-2-tag-2:
 
 >>>2