Fix possible errors introduced while generating Xcode project.

Under some unknown circumstances two added source files are concatenated into one faulty entry in an Xcode project (this probably occurs either at the Xcode or XMLTools2 level). Try to detect these errors in the Python script and fix them by patching up the project.pbxproj file.


git-svn-id: https://svn.wxwidgets.org/svn/wx/wxWidgets/trunk@65554 c3d73ce0-8a6f-49c7-b76d-6d57e0e08775

build/osx/fix_xcode_ids.py

@@ -9,26 +9,242 @@
 # Licence:      wxWindows licence
 ###############################################################################
 
+testFixStage = False
+
+import os
 import sys
 import re
 
 USAGE = """fix_xcode_ids - Modifies an Xcode project in-place to use the same identifiers (based on name) instead of being different on each regeneration"
 Usage: fix_xcode_ids xcode_proj_dir"""
 
-if len(sys.argv) < 2:
+if not testFixStage:
-    print USAGE
+    if len(sys.argv) < 2:
-    sys.exit(1)
+        print USAGE
+        sys.exit(1)
 
-projectFile = sys.argv[1] + "/project.pbxproj"
+    projectFile = sys.argv[1] + "/project.pbxproj"
+    fin = open(projectFile, "r")
+    strIn = fin.read()
+    fin.close()
 
-fin = open(projectFile, "r")
-strIn = fin.read()
-fin.close()
 
 
 # Xcode identifiers (IDs) consist of 24 hexadecimal digits
 idMask = "[A-Fa-f0-9]{24}"
 
+idDict = {}
+
+# convert a name to an identifier for Xcode
+def toUuid(name):
+    from uuid import uuid3, UUID
+    id = uuid3(UUID("349f853c-91f8-4eba-b9b9-5e9f882e693c"), name).hex[:24].upper()
+
+    # Some names can appear twice or even more (depending on number of
+    # targets), make them unique
+    while id in idDict.values() :
+        id = "%024X" % (int(id, 16) + 1)
+    return id
+
+def insertBuildFileEntry(filePath, fileRefId):
+    global strIn
+    print "\tInsert PBXBuildFile for '%s'..." % filePath,
+
+    matchBuildFileSection = re.search("/\* Begin PBXBuildFile section \*/\n", strIn)
+    dirName, fileName = os.path.split(filePath)
+
+    fileInSources = fileName + " in Sources"
+    id = toUuid(fileInSources)
+    idDict[id] = id
+    insert = "\t\t%s /* %s */ = {isa = PBXBuildFile; fileRef = %s /* %s */; };\n" % (id, fileInSources, fileRefId, fileName)
+
+    strIn = strIn[:matchBuildFileSection.end()] + insert + strIn[matchBuildFileSection.end():]
+
+    print "OK"
+    return id
+
+
+def insertFileRefEntry(filePath, id = 0):
+    global strIn
+    print "\tInsert PBXFileReference for '%s'..." % filePath,
+
+    matchFileRefSection = re.search("/\* Begin PBXFileReference section \*/\n", strIn)
+    dirName, fileName = os.path.split(filePath)
+    if id == 0:
+        id = toUuid(fileName)
+        idDict[id] = id
+
+    insert = "\t\t%s /* %s */ = {isa = PBXFileReference; lastKnownFileType = file; name = %s; path = %s; sourceTree = \"<group>\"; };\n" % (id, fileName, fileName, filePath)
+    strIn = strIn[:matchFileRefSection.end()] + insert + strIn[matchFileRefSection.end():]
+
+    print "OK"
+    return id
+
+
+def insertSourcesBuildPhaseEntry(id, fileName, insertBeforeFileName, startSearchPos = 0):
+    global strIn
+    print "\tInsert PBXSourcesBuildPhase for '%s'..." % fileName,
+
+    matchBuildPhase = re.compile(".+ /\* " + insertBeforeFileName + " in Sources \*/,") \
+        .search(strIn, startSearchPos)
+    insert = "\t\t\t\t%s /* %s in Sources */,\n" % (id, fileName)
+    strIn = strIn[:matchBuildPhase.start()] \
+        + insert \
+        + strIn[matchBuildPhase.start():]
+
+    print "OK"
+    return matchBuildPhase.start() + len(insert) + len(matchBuildPhase.group(0))
+
+# Detect and fix errors in the project file that might have been introduced.
+# Sometimes two source files are concatenated. These are spottable by
+# looking for patterns such as "filename.cppsrc/html/"
+# Following is a stripped Xcode project containing several problems that
+# are solved after finding the error.
+strTest = \
+"""/* Begin PBXBuildFile section */
+        95DE8BAB1238EE1800B43069 /* m_fonts.cpp in Sources */ = {isa = PBXBuildFile; fileRef = 95DE8BAA1238EE1700B43069 /* m_fonts.cpp */; };
+        95DE8BAC1238EE1800B43069 /* m_fonts.cpp in Sources */ = {isa = PBXBuildFile; fileRef = 95DE8BAA1238EE1700B43069 /* m_fonts.cpp */; };
+/* End PBXBuildFile section */
+
+/* Begin PBXFileReference section */
+        95DE8BAA1238EE1700B43069 /* m_fonts.cpp */ = {isa = PBXFileReference; lastKnownFileType = file; name = m_fonts.cpp; path = ../../src/html/m_dflist.cppsrc/html/m_fonts.cpp; sourceTree = "<group>"; };
+/* End PBXFileReference section */
+
+/* Begin PBXGroup section */
+        95DE8B831238EE1000B43069 /* html */ = {
+            isa = PBXGroup;
+            children = (
+                95DE8B841238EE1000B43069 /* src/html */,
+                95DE8BA91238EE1700B43069 /* src/html/m_dflist.cppsrc/html */,
+                95DE8BCE1238EE1F00B43069 /* src/generic */,
+            );
+            name = html;
+            sourceTree = "<group>";
+        };
+        95DE8B841238EE1000B43069 /* src/html */ = {
+            isa = PBXGroup;
+            children = (
+                95DE8B851238EE1000B43069 /* chm.cpp */,
+                95DE8BAD1238EE1800B43069 /* m_hline.cpp */,
+            );
+            name = src/html;
+            sourceTree = "<group>";
+        };
+
+        95DE8BA91238EE1700B43069 /* src/html/m_dflist.cppsrc/html */ = {
+            isa = PBXGroup;
+            children = (
+                95DE8BAA1238EE1700B43069 /* m_fonts.cpp */,
+            );
+            name = src/html/m_dflist.cppsrc/html;
+            sourceTree = "<group>";
+        };
+/* End PBXGroup section */
+
+
+/* Begin PBXSourcesBuildPhase section */
+        404BEE5E10EC83280080E2B8 /* Sources */ = {
+            files = (
+                95DE8BAC1238EE1800B43069 /* m_fonts.cpp in Sources */,
+            );
+            runOnlyForDeploymentPostprocessing = 0;
+        };
+        D2AAC0C405546C1D00DB518D /* Sources */ = {
+            files = (
+                95DE8BAB1238EE1800B43069 /* m_fonts.cpp in Sources */,
+            );
+            runOnlyForDeploymentPostprocessing = 0;
+        };
+
+/* End PBXSourcesBuildPhase section */"""
+
+if testFixStage:
+    strIn = strTest
+
+rc = re.compile(".+ (?P<path1>[\w/.]+(\.cpp|\.cxx|\.c))(?P<path2>\w+/[\w/.]+).+")
+matchLine = rc.search(strIn)
+while matchLine:
+    line = matchLine.group(0)
+
+    # is it a line from the PBXFileReference section containing 2 mixed paths?
+    # example:
+    # FEDCBA9876543210FEDCBA98 /* file2.cpp */ = {isa = PBXFileReference; lastKnownFileType = file; name = file2.cpp; path = ../../src/html/file1.cppsrc/html/file2.cpp; sourceTree = "<group>"; };
+    if line.endswith("};") :
+        path1 = matchLine.group('path1')
+        path2 = matchLine.group('path2')
+        print "Correcting mixed paths '%s' and '%s' at '%s':" % (path1, path2, line)
+        # if so, make note of the ID used (belongs to path2), remove the line
+        # and split the 2 paths inserting 2 new entries inside PBXFileReference
+        fileRefId2 = re.search(idMask, line).group(0)
+
+        print "\tDelete the offending PBXFileReference line...",
+        # delete the PBXFileReference line that was found and which contains 2 mixed paths
+        strIn = strIn[:matchLine.start()] + strIn[matchLine.end()+1:]
+        print "OK"
+
+        # insert corrected path1 entry in PBXFileReference
+        fileRefId1 = insertFileRefEntry(path1)
+
+        # do the same for path2 (which already had a ID)
+        path2Corrected = path2
+        if path2Corrected.startswith('src') :
+            path2Corrected = '../../' + path2Corrected
+
+        insertFileRefEntry(path2Corrected, fileRefId2)
+
+
+        buildPhaseId = {}
+        # insert a PBXBuildFile entry, 1 for each target
+        # path2 already has correct PBXBuildFile entries
+        targetCount = strIn.count("isa = PBXSourcesBuildPhase")
+        for i in range(0, targetCount):
+            buildPhaseId[i] = insertBuildFileEntry(path1, fileRefId1)
+
+
+        fileName1 = os.path.split(path1)[1]
+        dir2, fileName2 = os.path.split(path2)
+
+        # refer to each PBXBuildFile in each PBXSourcesBuildPhase
+        startSearchIndex = 0
+        for i in range(0, targetCount):
+            startSearchIndex = insertSourcesBuildPhaseEntry(buildPhaseId[i], fileName1, fileName2, startSearchIndex)
+
+
+        # insert both paths in the group they belong to
+        matchGroupStart = re.search("/\* %s \*/ = {" % dir2, strIn)
+        endGroupIndex = strIn.find("};", matchGroupStart.start())
+
+        for matchGroupLine in re.compile(".+" + idMask + " /\* (.+) \*/,").finditer(strIn, matchGroupStart.start(), endGroupIndex) :
+            if matchGroupLine.group(1) > fileName1:
+                print "\tInsert paths in PBXGroup '%s', just before '%s'..." % (dir2, matchGroupLine.group(1)),
+                strIn = strIn[:matchGroupLine.start()] \
+                    + "\t\t\t\t%s /* %s */,\n" % (fileRefId1, fileName1) \
+                    + "\t\t\t\t%s /* %s */,\n" % (fileRefId2, fileName2) \
+                    + strIn[matchGroupLine.start():]
+                print "OK"
+
+                break
+
+    elif line.endswith("*/ = {") :
+        print "Delete invalid PBXGroup starting at '%s'..." % line,
+        find = "};\n"
+        endGroupIndex = strIn.find(find, matchLine.start()) + len(find)
+        strIn = strIn[:matchLine.start()] + strIn[endGroupIndex:]
+        print "OK"
+
+    elif line.endswith(" */,") :
+        print "Delete invalid PBXGroup child '%s'..." % line,
+        strIn = strIn[:matchLine.start()] + strIn[matchLine.end()+1:]
+        print "OK"
+
+    matchLine = rc.search(strIn)
+
+if testFixStage:
+    print "------------------------------------------"
+    print strIn
+    exit(1)
+
+
 # key = original ID found in project
 # value = ID it will be replaced by
 idDict = {}
@@ -48,22 +264,10 @@ idDict = {}
 rc = re.compile("\s+(" + idMask + ") /\* (.+) \*/ = {.*$", re.MULTILINE)
 dict = rc.findall(strIn)
 
-# convert a name to an identifier for Xcode
-def toUuid(name):
-    from uuid import uuid3, UUID
-    return uuid3(UUID("349f853c-91f8-4eba-b9b9-5e9f882e693c"), name).hex[:24].upper()
-
 for s in dict:
     # s[0] is the original ID, s[1] is the name
-    newId = toUuid(s[1])
-    # Some names can appear twice or even more (depending on number of
-    # targets), make them unique
-    while newId in idDict.values() :
-        # [2:-1] to skip prepended 0x and trailing L
-        newId = hex(int(newId, 16) + 1)[2:-1].upper()
- 
     assert(not s[0] in idDict)
-    idDict[s[0]] = newId
+    idDict[s[0]] = toUuid(s[1])
 
 
 # replace all found identifiers with the new ones