fix: Improve the accuracy of the coverage report.

- Ignore trailing '()' characters in function names when reporting coverage.
- Only track functions that match a function in the header file of interest as implemented.
- Mark the uenum::uloc_open_keywords as deprecated, copy the implementation into the uloc crate and copy the test that was in the uloc crate into the uenum crate so that it is still protected by a test even though deprecated.

build/showprogress.sh

@@ -29,86 +29,99 @@ C_API_HEADER_NAMES=(
 
 ICU_INCLUDE_PATH="$(icu-config --cppflags-searchpath | sed -e 's/-I//' | sed -e 's/ //g')"
 
-for header_basename in ${C_API_HEADER_NAMES[@]}; do
-  header_fullname="${ICU_INCLUDE_PATH}/unicode/${header_basename}.h"
-  echo $header_basename: $header_fullname
-  ctags -x --c-kinds=fp $header_fullname | sed -e 's/ .*$//' \
-    | grep -v U_DEFINE | sort | uniq \
-    > "${TOP_DIR}/coverage/${header_basename}_all.txt"
-
-  # Extracts all mentions of functions such as "utext_close" for example from
-  # rust docs of the form "/// Implements `utext_close` ... ".  This is
-  # simplistic but quite enough with a little bit of care.
-  find . -path "*rust_icu_${header_basename}/src/*.rs" | \
-    xargs grep "/// Implements \`" | sed -e 's/.*`\(.*\)`.*$/\1/' | sed -e 's/\(.*\)()$/\1/' | \
-    sort | uniq > "${TOP_DIR}/coverage/${header_basename}_implemented.txt"
-done
-
-# Now, write a report out. Again, simplistic, but gets the job done.
+# Write a report out as we read the data.
 
 REPORT_FILE="${TOP_DIR}/coverage/report.md"
-cat <<EOF > "${REPORT_FILE}"
+REPORT_FILE_HEADER="${TOP_DIR}/coverage/report_header.md"
+REPORT_FILE_DETAIL="${TOP_DIR}/coverage/report_detail.md"
+cat <<EOF > "${REPORT_FILE_HEADER}"
 # Implementation coverage report
 
 | Header | Implemented |
 | ------ | ----------- |
 EOF
-for header_basename in ${C_API_HEADER_NAMES[@]}; do
-  total_functions="$(cat "${TOP_DIR}"/coverage/${header_basename}_all.txt | wc -l)"
-  implemented_functions="$(cat "${TOP_DIR}"/coverage/${header_basename}_implemented.txt | wc -l)"
-  printf "| \`%s.h\` | %s / %s | \n" ${header_basename} ${implemented_functions} ${total_functions} >> "${REPORT_FILE}"
-done
 
-cat <<EOF >>"${REPORT_FILE}"
+  cat <<EOF >>"${REPORT_FILE_DETAIL}"
 # Unimplemented functions per header
 
 EOF
-for header_basename in ${C_API_HEADER_NAMES[@]}; do
-  cat <<EOF >>"${REPORT_FILE}"
-
-# Header: \`${header_basename}.h\`
-
-| Unimplemented | Implemented |
-| ------------- | ----------- |
-EOF
-  all=`cat ${TOP_DIR}/coverage/${header_basename}_all.txt | sort -fs`
-  implemented=`cat "${TOP_DIR}/coverage/${header_basename}_implemented.txt" | sort -fs`
-  for fun in ${implemented}; do
-    echo "| | \`${fun}\` |" >>"${REPORT_FILE}"
-  done
 
-  unimplemented="$(comm -23 \
-    "${TOP_DIR}/coverage/${header_basename}_all.txt" \
-    "${TOP_DIR}/coverage/${header_basename}_implemented.txt" | sort -fs | uniq)"
-  for fun in ${unimplemented}; do
-    echo "| \`${fun}\` | |" >>"${REPORT_FILE}"
+for header_basename in ${C_API_HEADER_NAMES[@]}; do
+  : > "${TOP_DIR}/coverage/${header_basename}_all.txt"
+  : > "${TOP_DIR}/coverage/${header_basename}_implemented.txt"
+  header_fullname="${ICU_INCLUDE_PATH}/unicode/${header_basename}.h"
+  all=$(ctags -x --c-kinds=fp $header_fullname | sed -e 's/ .*$//' \
+    | grep -v U_DEFINE | sort -fs | uniq)
+  for fn in ${all}; do
+    printf "%s\n" ${fn} >> "${TOP_DIR}/coverage/${header_basename}_all.txt"
   done
 
-  total_cnt=`echo ${all} | wc -w`
-  implemented_cnt=`echo ${implemented} | wc -w`
-  unimplemented_cnt=`echo ${unimplemented} | wc -w`
-  computed_total_cnt=`expr ${unimplemented_cnt} + ${implemented_cnt}`
-  if [ ${computed_total_cnt} -ne ${total_cnt} ]
-  then
-    printf "Warning: Implemented + Unimplemented != Total\n" >> ${REPORT_FILE}
-    printf "Total: %s Implemented: %s Unimplemented: %s\n" ${total_cnt} ${implemented_cnt} ${unimplemented_cnt} >> "${REPORT_FILE}"
-    for impl_fn in ${implemented}; do
-      found="false"
+  # Extracts all mentions of functions such as "utext_close" for example from
+  # rust docs of the form "/// Implements `utext_close` ... ".  This is
+  # simplistic but quite enough with a little bit of care.
+  files=`find . -path "*rust_icu_${header_basename}/src/*.rs"`
+  impl_fns=""
+  unimpl_fns=""
+  for file in ${files}; do
+    echo $header_basename: $header_fullname ${file}
+    found_fns="$(grep "/// Implements \`" ${file} | sed -e 's/.*`\(.*\)`.*$/\1/' | sed -e 's/\(.*\)()$/\1/' | sort -fs | uniq)"
+    # Match the extracted function to a function in the header being processed
+    # (in case the "/// Implements `" comment is used in another context)
+    for impl_fn in ${found_fns}; do
       for all_fn in ${all}; do
         if [ ${impl_fn} = ${all_fn} ]
         then
-          found="true"
+          impl_fns="${impl_fns} ${impl_fn}"
           break
         fi
       done
-      if [ "false" = ${found} ]
+    done
+  done
+  # Sort again in case we process multiple source files
+  impl_fns="$(echo ${impl_fns} | sort -fs | uniq)"
+  for impl_fn in ${impl_fns}; do
+    printf "%s\n" ${impl_fn} >> "${TOP_DIR}/coverage/${header_basename}_implemented.txt"
+  done
+  unimpl_fns=""
+  for all_fn in ${all}; do
+    found="false"
+    for impl_fn in ${impl_fns}; do
+      if [ ${impl_fn} = ${all_fn} ]
       then
-        printf "Function %s was designated as implemented but not found in the API\n" ${impl_fn} >> "${REPORT_FILE}"
+        found="true"
+        break
       fi
     done
-  fi
+    if [ "false" = ${found} ]
+    then
+      unimpl_fns="${unimpl_fns} ${all_fn}"
+    fi
+  done
+
+  total_functions=$(echo ${all} | wc -w)
+  implemented_functions=$(echo ${impl_fns} | wc -w)
+  printf "| \`%s.h\` | %s / %s | \n" ${header_basename} ${implemented_functions} ${total_functions} >> "${REPORT_FILE_HEADER}"
+
+  cat <<EOF >>"${REPORT_FILE_DETAIL}"
+
+# Header: \`${header_basename}.h\`
+
+| Unimplemented | Implemented |
+| ------------- | ----------- |
+EOF
+  for impl_fn in ${impl_fns}; do
+    printf "| | \`%s\` |\n" ${impl_fn} >>"${REPORT_FILE_DETAIL}"
+  done
+
+  for fun in ${unimpl_fns}; do
+    printf "| \`%s\` | |\n"  ${fun} >>"${REPORT_FILE_DETAIL}"
+  done
 
   sort -fs -o "${TOP_DIR}/coverage/${header_basename}_all.txt" "${TOP_DIR}/coverage/${header_basename}_all.txt"
   sort -fs -o "${TOP_DIR}/coverage/${header_basename}_implemented.txt" "${TOP_DIR}/coverage/${header_basename}_implemented.txt"
 done
 
+cat ${REPORT_FILE_HEADER} ${REPORT_FILE_DETAIL} > ${REPORT_FILE}
+rm ${REPORT_FILE_HEADER}
+rm ${REPORT_FILE_DETAIL}
+

coverage/report.md

@@ -5,18 +5,18 @@
 | `ubrk.h` | 19 / 23 | 
 | `ucal.h` | 15 / 47 | 
 | `ucol.h` | 8 / 51 | 
-| `udat.h` | 10 / 38 | 
-| `udata.h` | 4 / 8 | 
-| `uenum.h` | 7 / 8 | 
+| `udat.h` | 5 / 38 | 
+| `udata.h` | 2 / 8 | 
+| `uenum.h` | 3 / 8 | 
 | `uformattable.h` | 6 / 13 | 
 | `ulistformatter.h` | 2 / 8 | 
 | `uloc.h` | 28 / 42 | 
-| `umsg.h` | 6 / 20 | 
+| `umsg.h` | 5 / 20 | 
 | `unum.h` | 14 / 32 | 
-| `unumberformatter.h` | 7 / 6 | 
+| `unumberformatter.h` | 3 / 6 | 
 | `upluralrules.h` | 3 / 8 | 
-| `ustring.h` | 3 / 61 | 
-| `utext.h` | 3 / 28 | 
+| `ustring.h` | 2 / 61 | 
+| `utext.h` | 2 / 28 | 
 | `utrans.h` | 10 / 20 | 
 | `unorm2.h` | 8 / 22 | 
 # Unimplemented functions per header
@@ -162,11 +162,6 @@
 
 | Unimplemented | Implemented |
 | ------------- | ----------- |
-| | `UDateFormat` |
-| | `UDateTimePatternGenerator` |
-| | `udatpg_clone` |
-| | `udatpg_getBestPattern` |
-| | `udatpg_open` |
 | | `udat_close` |
 | | `udat_format` |
 | | `udat_open` |
@@ -205,41 +200,24 @@
 | `udat_toPatternRelativeDate` | |
 | `udat_toPatternRelativeTime` | |
 | `udat_unregisterOpener` | |
-Warning: Implemented + Unimplemented != Total
-Total: 38 Implemented: 10 Unimplemented: 33
-Function UDateFormat was designated as implemented but not found in the API
-Function UDateTimePatternGenerator was designated as implemented but not found in the API
-Function udatpg_clone was designated as implemented but not found in the API
-Function udatpg_getBestPattern was designated as implemented but not found in the API
-Function udatpg_open was designated as implemented but not found in the API
 
 # Header: `udata.h`
 
 | Unimplemented | Implemented |
 | ------------- | ----------- |
-| | `UDataMemory` |
 | | `udata_open` |
 | | `udata_setCommonData` |
-| | `u_setDataDirectory` |
 | `udata_close` | |
 | `udata_getInfo` | |
 | `udata_getMemory` | |
 | `udata_openChoice` | |
 | `udata_setAppData` | |
 | `udata_setFileAccess` | |
-Warning: Implemented + Unimplemented != Total
-Total: 8 Implemented: 4 Unimplemented: 6
-Function UDataMemory was designated as implemented but not found in the API
-Function u_setDataDirectory was designated as implemented but not found in the API
 
 # Header: `uenum.h`
 
 | Unimplemented | Implemented |
 | ------------- | ----------- |
-| | `ucal_openCountryTimeZones` |
-| | `ucal_openTimeZoneIDEnumeration` |
-| | `ucal_openTimeZones` |
-| | `UEnumeration` |
 | | `uenum_close` |
 | | `uenum_next` |
 | | `uenum_openCharStringsEnumeration` |
@@ -248,12 +226,6 @@ Function u_setDataDirectory was designated as implemented but not found in the A
 | `uenum_openUCharStringsEnumeration` | |
 | `uenum_reset` | |
 | `uenum_unext` | |
-Warning: Implemented + Unimplemented != Total
-Total: 8 Implemented: 7 Unimplemented: 5
-Function ucal_openCountryTimeZones was designated as implemented but not found in the API
-Function ucal_openTimeZoneIDEnumeration was designated as implemented but not found in the API
-Function ucal_openTimeZones was designated as implemented but not found in the API
-Function UEnumeration was designated as implemented but not found in the API
 
 # Header: `uformattable.h`
 
@@ -337,7 +309,6 @@ Function UEnumeration was designated as implemented but not found in the API
 
 | Unimplemented | Implemented |
 | ------------- | ----------- |
-| | `UMessageFormat` |
 | | `umsg_clone` |
 | | `umsg_close` |
 | | `umsg_format` |
@@ -358,9 +329,6 @@ Function UEnumeration was designated as implemented but not found in the API
 | `u_vformatMessageWithError` | |
 | `u_vparseMessage` | |
 | `u_vparseMessageWithError` | |
-Warning: Implemented + Unimplemented != Total
-Total: 20 Implemented: 6 Unimplemented: 15
-Function UMessageFormat was designated as implemented but not found in the API
 
 # Header: `unum.h`
 
@@ -406,19 +374,9 @@ Function UMessageFormat was designated as implemented but not found in the API
 | | `unumf_formatDecimal` |
 | | `unumf_openForSkeletonAndLocale` |
 | | `unumf_openForSkeletonAndLocaleWithError` |
-| | `unumf_openResult` |
-| | `unumf_resultGetAllFieldPositions` |
-| | `unumf_resultNextFieldPosition` |
-| | `unumf_resultToString` |
 | `unumf_close` | |
 | `unumf_formatDouble` | |
 | `unumf_formatInt` | |
-Warning: Implemented + Unimplemented != Total
-Total: 6 Implemented: 7 Unimplemented: 3
-Function unumf_openResult was designated as implemented but not found in the API
-Function unumf_resultGetAllFieldPositions was designated as implemented but not found in the API
-Function unumf_resultNextFieldPosition was designated as implemented but not found in the API
-Function unumf_resultToString was designated as implemented but not found in the API
 
 # Header: `upluralrules.h`
 
@@ -437,7 +395,6 @@ Function unumf_resultToString was designated as implemented but not found in the
 
 | Unimplemented | Implemented |
 | ------------- | ----------- |
-| | `UChar*` |
 | | `u_strFromUTF8` |
 | | `u_strToUTF8` |
 | `u_austrcpy` | |
@@ -499,17 +456,13 @@ Function unumf_resultToString was designated as implemented but not found in the
 | `u_uastrncpy` | |
 | `u_unescape` | |
 | `u_unescapeAt` | |
-Warning: Implemented + Unimplemented != Total
-Total: 61 Implemented: 3 Unimplemented: 59
-Function UChar* was designated as implemented but not found in the API
 
 # Header: `utext.h`
 
 | Unimplemented | Implemented |
 | ------------- | ----------- |
 | | `utext_clone` |
 | | `utext_close` |
-| | `utext_open` |
 | `utext_char32At` | |
 | `utext_copy` | |
 | `utext_current32` | |
@@ -536,9 +489,6 @@ Function UChar* was designated as implemented but not found in the API
 | `utext_replace` | |
 | `utext_setNativeIndex` | |
 | `utext_setup` | |
-Warning: Implemented + Unimplemented != Total
-Total: 28 Implemented: 3 Unimplemented: 26
-Function utext_open was designated as implemented but not found in the API
 
 # Header: `utrans.h`
 

coverage/udat_implemented.txt

@@ -1,8 +1,3 @@
-UDateFormat
-UDateTimePatternGenerator
-udatpg_clone
-udatpg_getBestPattern
-udatpg_open
 udat_close
 udat_format
 udat_open

coverage/udata_implemented.txt

@@ -1,4 +1,2 @@
-UDataMemory
 udata_open
 udata_setCommonData
-u_setDataDirectory

coverage/uenum_implemented.txt

@@ -1,7 +1,3 @@
-ucal_openCountryTimeZones
-ucal_openTimeZoneIDEnumeration
-ucal_openTimeZones
-UEnumeration
 uenum_close
 uenum_next
 uenum_openCharStringsEnumeration

coverage/umsg_implemented.txt

@@ -1,4 +1,3 @@
-UMessageFormat
 umsg_clone
 umsg_close
 umsg_format

coverage/unumberformatter_implemented.txt

@@ -1,7 +1,3 @@
 unumf_formatDecimal
 unumf_openForSkeletonAndLocale
 unumf_openForSkeletonAndLocaleWithError
-unumf_openResult
-unumf_resultGetAllFieldPositions
-unumf_resultNextFieldPosition
-unumf_resultToString

coverage/ustring_implemented.txt

@@ -1,3 +1,2 @@
-UChar*
 u_strFromUTF8
 u_strToUTF8

coverage/utext_implemented.txt

@@ -1,3 +1,2 @@
 utext_clone
 utext_close
-utext_open

rust_icu_uloc/src/lib.rs

@@ -439,7 +439,7 @@ impl ULoc {
             .map(|result| result.unwrap())
     }
 
-    /// Implementation of `icu::Locale::getUnicodeKeywords()` from the C++ API.
+    /// Implements `icu::Locale::getUnicodeKeywords()` from the C++ API.
     pub fn unicode_keywords(&self) -> impl Iterator<Item = String> {
         self.keywords().filter_map(|s| to_unicode_locale_key(&s))
     }
@@ -463,7 +463,7 @@ impl ULoc {
         .map(|value| if value.is_empty() { None } else { Some(value) })
     }
 
-    /// Implementation of `icu::Locale::getUnicodeKeywordValue()` from the C++ API.
+    /// Implements `icu::Locale::getUnicodeKeywordValue()` from the C++ API.
     pub fn unicode_keyword_value(
         &self,
         unicode_keyword: &str,