From: Timo Heister <timo.heister@gmail.com>
Date: Thu, 13 Feb 2014 02:10:35 +0000 (+0000)
Subject: add python script to reformat and rewrap doxygen comments
X-Git-Tag: v8.2.0-rc1~833
X-Git-Url: https://gitweb.dealii.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=05d853b9778d44678f6102032f9b7f466f1efb67;p=dealii.git

add python script to reformat and rewrap doxygen comments

git-svn-id: https://svn.dealii.org/trunk@32474 0785d39b-7218-0410-832d-ea1e28bc413d
---

diff --git a/deal.II/contrib/utilities/wrapcomments.py b/deal.II/contrib/utilities/wrapcomments.py
new file mode 100755
index 0000000000..97dc9dc3a2
--- /dev/null
+++ b/deal.II/contrib/utilities/wrapcomments.py
@@ -0,0 +1,293 @@
+#!/usr/bin/python
+
+# run this script on all headers of deal.II to fix comment line wrapping for
+# doxygen comments.
+# Example:
+# cd include
+# find . -name "*h" -print | while read file;do ../contrib/utilities/wrapcomments.py $file >temp;mv temp $file;done
+
+from __future__ import print_function
+import textwrap
+import sys
+import string
+wrapper = textwrap.TextWrapper()
+
+# take an array of lines and wrap them to 78 columns and let each lines start
+# with @p startwith
+def wrap_block(lines, startwith):
+    longline = " ".join(lines)
+    wrapper.initial_indent = startwith
+    wrapper.subsequent_indent = startwith
+    wrapper.width = 78
+    return wrapper.wrap(longline)
+
+# strips whitespace and leading "*" from each lines
+def remove_junk(lines):
+    out = []
+    for line in lines:
+        line = line.strip()
+        if line.startswith("*"):
+            line = line[1:].strip()
+        out.append(line)
+    return out
+
+# returns True if at least one entry in @p list is contained in @p str
+def one_in(list, str):
+    for li in list:
+        if li in str:
+            return True
+    return False
+
+# returns True if @p str starts with one of the entries in @p list
+def starts_with_one(list, str):
+    for li in list:
+        if str.startswith(li):
+            return True
+    return False
+
+# take a doxygen comment block "/** bla ... */" given as a list of lines and
+# format it in a pretty way rewrapping lines while taking care to keep certain
+# structure.
+def format_block(lines, infostr=""):
+    if len(lines)==1 and "/*" in lines[0] and "*/" in lines[0]:
+        return lines
+
+    if (not "/**" in lines[0]
+          or not "*/" in lines[-1]):
+        print("%s not a code block"%infostr, file=sys.stderr)
+        return lines
+
+    for line in lines[1:-1]:
+        assert(not "/*" in line)
+        assert(not "*/" in line)
+
+    lines[-1]=lines[-1].replace("**/","*/")
+
+    if not lines[0].strip().startswith("/**"):
+        print ("%s error, ignoring code block with junk in same line before"%infostr, file=sys.stderr)
+        return lines
+    if not lines[-1].strip().endswith("*/"):
+        print ("%s error, ignoring code block not ending at end of line"%infostr, file=sys.stderr)
+        return lines
+
+    if lines[0].strip()!="/**":
+        #print ("%s warning code block not starting in separate line"%infostr, file=sys.stderr)
+        idx = string.find(lines[0],"/**")
+        temp = [lines[0][0:idx+3], lines[0][idx+3:]]
+        temp.extend(lines[1:])
+        lines = temp
+    if lines[-1].strip()!="*/":
+        #print ("%s warning code block not ending in separate line"%infostr, file=sys.stderr)
+        idx = string.find(lines[-1],"*/")
+        temp = lines[0:-1]
+        temp.append(lines[-1][0:idx])
+        temp.append(lines[-1][idx:])
+        lines = temp
+
+    idx = string.find(lines[0],"/**")
+    start = lines[0][:idx]+" * "
+    
+    out = [lines[0]]
+    idx = 1
+    endidx = len(lines)-1
+    curlines = []
+
+    ops_startline = ["<li>", "@param", "@returns", "@warning", "@ingroup", "@author", "@date", "@related", "@deprecated"]
+
+    ops_separate_line = ["<ul>", "</ul>", "@{", "@}"]
+
+    while idx<endidx:
+        if one_in(ops_separate_line, lines[idx]):
+            if curlines!=[]:
+                out.extend(wrap_block(remove_junk(curlines), start))
+                curlines=[]
+            thisline = remove_junk([lines[idx]])[0]
+            for it in ops_separate_line:
+                if it in thisline and thisline!=it:
+                    print ("%s warning %s not in separate line"%(infostr, it), file=sys.stderr)
+            out.append(start + thisline)
+        elif one_in(ops_startline, lines[idx]):
+            if curlines!=[]:
+                out.extend(wrap_block(remove_junk(curlines), start))
+                curlines=[]
+            thisline = remove_junk([lines[idx]])[0]
+            if not starts_with_one(ops_startline, thisline):
+                for it in ops_startline:
+                    if it in thisline:
+                        print ("%s warning %s not at start of line"%(infostr, it), file=sys.stderr)
+            curlines.append(lines[idx])
+        elif one_in(["@code", "@verbatim", "@f["], lines[idx]):
+            if curlines!=[]:
+                out.extend(wrap_block(remove_junk(curlines), start))
+                curlines=[]
+            while True:
+                if lines[idx][0:len(start)].strip()!="*":
+                    print ("%s Error: wrong formatting inside @code block"%infostr, file=sys.stderr)
+                    thisline = lines[idx].strip()
+                    if thisline.startswith("* "):
+                        thisline = thisline[2:]
+                    elif thisline.startswith("*"):
+                        thisline = thisline[1:]
+                    thisline = start + thisline
+                else:
+                    thisline = start + lines[idx][len(start):]
+                out.append(thisline.rstrip())
+                if one_in(["@endcode", "@verbatim", "@f]"], lines[idx]):
+                    break
+                idx += 1
+        elif lines[idx].strip()=="*":
+            if curlines!=[]:
+                out.extend(wrap_block(remove_junk(curlines), start))
+                curlines=[]
+            out.append(start[:-1]) #skip whitespace at the end
+        else:
+            curlines.append(lines[idx])
+        idx += 1
+
+    if curlines!=[]:
+        out.extend(wrap_block(remove_junk(curlines), start))        
+
+    out.append(start[0:-2] + lines[-1].strip())
+
+    return out
+
+
+# test the routines:
+lineI = ["   * blub", \
+         "   * two three ", \
+         "   * four"]
+lineO = ["blub", \
+         "two three", \
+         "four"]
+assert(remove_junk(lineI)==lineO)
+
+lineI = ["blub", \
+         "two three", \
+         "four"]
+lineO = ["   * blub two three four"]
+assert(wrap_block(lineI,"   * ")==lineO)
+
+lineI = [" * 1 2 3 4 5 6 7 9 0 1 2 3 4 5 6 7 9 0 1 2 3 4 5 6 7 9 0 1 2 3 4 5 6 7 9 0 1 2",\
+         " * A 4 5 6 7 9 0 1 2 3 4 5 6 7 9 0"]
+assert(wrap_block(remove_junk(lineI)," * ")==lineI)
+
+lineI = [" * Structure which is passed to Triangulation::create_triangulation. It",\
+         " * contains all data needed to construct a cell, namely the indices of the",\
+         " * vertices and the material indicator."]
+assert(wrap_block(remove_junk(lineI)," * ")==lineI)
+
+lineI = ["  /**", \
+         "   * blub", \
+         "   * two three", \
+         "   * four", \
+         "   */"]
+lineO = ["  /**", \
+         "   * blub two three four", \
+         "   */"]
+assert(format_block(lineI)==lineO)
+
+lineI = ["  /**", \
+         "   * blub", \
+         "   * two three", \
+         "   * ", \
+         "   * four", \
+         "   */"]
+lineO = ["  /**", \
+         "   * blub two three", \
+         "   *", \
+         "   * four", \
+         "   */"]
+assert(format_block(lineI)==lineO)
+
+lineI = ["  /**", \
+         "   * blub", \
+         "   * @code", \
+         "   *   two three", \
+         "   * @endcode", \
+         "   * four", \
+         "   */"]
+assert(format_block(lineI)==lineI)
+
+lineI = ["  /**", \
+         "   * blub", \
+         "   * @code ", \
+         "   *   two three", \
+         "   *   two three  ", \
+         "   * ", \
+         "   *   two three", \
+         "   * @endcode ", \
+         "   * four", \
+         "   */"]
+lineO = ["  /**", \
+         "   * blub", \
+         "   * @code", \
+         "   *   two three", \
+         "   *   two three", \
+         "   *", \
+         "   *   two three", \
+         "   * @endcode", \
+         "   * four", \
+         "   */"]
+assert(format_block(lineI)==lineO)
+ 
+
+lineI = ["  /**", \
+         "   * blub", \
+         "   * <ul>", \
+         "   * <li> bla", \
+         "   * <li> blub", \
+         "   * </ul>", \
+         "   */"]
+assert(format_block(lineI)==lineI)
+ 
+lineI = ["    /** @addtogroup Exceptions", \
+         "     * @{ */"]
+lineO = ["    /**", \
+         "     * @addtogroup Exceptions", \
+         "     * @{", \
+         "     */"]
+assert(format_block(lineI)==lineO)
+
+
+# now open the file and do the work
+
+
+args=sys.argv
+args.pop(0)
+
+if len(args)!=1:
+    print("Usage: wrapcomments.py infile >outfile")
+    exit(0)
+
+f = open(args[0])
+lines = f.readlines()
+f.close()
+
+out = []
+cur = []
+inblock = False
+lineidx = 0
+for line in lines:
+    lineidx += 1
+    line = line.replace("\n","")
+    if not inblock and "/**" in line:
+        inblock = True
+        cur = []
+    if inblock:
+        cur.append(line)
+    else:
+        out.append(line)
+    if inblock and "*/" in line:
+        out.extend(format_block(cur, args[0]+":%d"%lineidx))
+        cur = []
+        inblock = False
+
+assert(cur==[])
+
+for line in out:
+    print (line)
+
+
+
+
+