Change type of column-names option

Change type of column-names map option to map(xs:string, xs:integer)
from map(xs:integer, xs:string). It turns out I had used this format for
all the examples, and the more I thought about it, the more it seemed
unhelpful to have a map like `map { 1: "a", 2: "b" }` produce a
`csv-columns-record` whose `names` entry was `map { "a": 1, "b": 2 }`.

Not least, that prevented the `names` entry from a `csv-columns-record`
being used as the `column-names` option in another invocation, which
might be desirable if a user has one CSV with headers and subsequent
CSVs of the same schema but without the header line (transaction logs
split for size, perhaps).

specifications/xpath-functions-40/src/function-catalog.xml

@@ -21524,7 +21524,7 @@ return $M(collation-key("a", $C))</eg></fos:expression>
                <fos:meaning>Determines whether the first row of the CSV should be treated as a list
                   of column names and returned as a <code>csv-columns-record</code> in the
                   <code>columns</code> entry of the returned map. Permitted values are a map of type
-                  <code>map(xs:integer, xs:string)</code> or an <code>xs:boolean</code>.
+                  <code>map(xs:string, xs:integer)</code> or an <code>xs:boolean</code>.
                </fos:meaning>
                <fos:type>item()</fos:type>
                <fos:default>false</fos:default>
@@ -21544,7 +21544,7 @@ return $M(collation-key("a", $C))</eg></fos:expression>
                      the returned <code>parsed-csv-structure-record</code>. Implementations
                      <rfc2119>must not</rfc2119> exclude the first row from the <code>rows</code>
                      entry of the <code>parsed-csv-structure-record</code>.</fos:value>
-                  <fos:value value="map(xs:integer, xs:string)">A <code>csv-columns-record</code> is
+                  <fos:value value="map(xs:string, xs:integer)">A <code>csv-columns-record</code> is
                      constructed using the supplied map and returned as the <code>header</code>
                      entry of the <code>parsed-csv-structure-record</code>. The supplied map is used
                      as the <code>names</code> entry, and a sequence of strings for the
@@ -21758,7 +21758,7 @@ return $M(collation-key("a", $C))</eg></fos:expression>
             </fos:test>
          </fos:example>
          <fos:variable name="csv-string" id="csv-string-no-headers">`Alice,Aachen{$crlf}Bob,Berlin{$crlf}`</fos:variable>
-         <fos:variable name="options" id="explicit-csv-column-opts">map { "column-names": map { 1: "Person", 2: "Location" } }</fos:variable>
+         <fos:variable name="options" id="explicit-csv-column-opts">map { "column-names": map { "Person": 1, "Location": 2 } }</fos:variable>
          <fos:example>
             <p>Specifying column names explicitly:</p>
             <fos:test use="csv-string-no-headers escaped-crlf-2 explicit-csv-column-opts">
@@ -21807,7 +21807,7 @@ return $M(collation-key("a", $C))</eg></fos:expression>
          <fos:example>
             <p>Filtering columns, with <code>column-names: map { ... }</code></p>
             <fos:test use="escaped-crlf-2 uneven-cols-csv-string">
-               <fos:expression>csv-to-xdm($csv-uneven-cols, map { "column-names": map { 1: "Person", 3: "Amount" }, "filter-columns": (2,1,4) })?columns</fos:expression>
+               <fos:expression>csv-to-xdm($csv-uneven-cols, map { "column-names": map { "Person": 1, "Amount": 3 }, "filter-columns": (2,1,4) })?columns</fos:expression>
                <fos:result>map {
    "names": map { "Person": 1, "Amount": 3 },
    "fields": ("Person", "", "Amount")
@@ -21934,7 +21934,7 @@ return $M(collation-key("a", $C))</eg></fos:expression>
                <fos:meaning>Determines whether the first row of the CSV should be treated as a list
                   of column headers and returned as <code><![CDATA[<column>]]></code> elements in
                   the <code><![CDATA[<header>]]></code> element. Permitted values are a map of type
-                  <code>map(xs:integer, xs:string)</code> or an <code>xs:boolean</code>.
+                  <code>map(xs:string, xs:integer)</code> or an <code>xs:boolean</code>.
                </fos:meaning>
                <fos:type>item()</fos:type>
                <fos:default>false</fos:default>
@@ -21946,7 +21946,7 @@ return $M(collation-key("a", $C))</eg></fos:expression>
                      element.</fos:value>
                   <fos:value value="false">Implementations <rfc2119>must not</rfc2119> include a
                      <code><![CDATA[<header>]]></code> element in the output.</fos:value>
-                  <fos:value value="map(xs:integer, xs:string)">The supplied map is used to
+                  <fos:value value="map(xs:string, xs:integer)">The supplied map is used to
                      construct a sequence of <code><![CDATA[<column>]]></code> elements to populate
                      the <code><![CDATA[<header>]]></code> element. The <code>xs:integer</code>
                      denotes the column number, and the <code>xs:string</code> the column name. Gaps
@@ -22027,7 +22027,7 @@ return $M(collation-key("a", $C))</eg></fos:expression>
          <fos:example>
             <p>An empty CSV with explicit column names:</p>
             <fos:test>
-               <fos:expression>csv-to-xml("", map { "column-names": map { 1: "name", 3: "city"})</fos:expression>
+               <fos:expression>csv-to-xml("", map { "column-names": map { "name": 1, "city": 3 })</fos:expression>
                <fos:result><![CDATA[
 <fn:csv>
    <fn:header>