Another rewrite of scan4dups to fix things still being missed. Added NABLA and INTEGRAL: how did we ever manage without them??

2 files changed, 33 insertions, 18 deletions

diff --git a/dotXCompose b/dotXCompose
index ce67240..73f0df2 100644
--- a/dotXCompose
+++ b/dotXCompose
@@ -52,6 +52,10 @@ include "%L"
 <Multi_key> <Up> <Down>			: "↕"	U2195           # UP DOWN ARROW (kragen's)
 <Multi_key> <Down> <Left>		: "↵"	U21B5		# DOWNWARDS ARROW WITH CORNER LEFTWARDS
 
+# Arrow keys don't always work: some apps trap them for cursor control and
+# other boring things.  The arrow symbols have alternate keystrokes.  Do
+# we need others for these printer's fists?  If so, what?  The -= and =-
+# we had before are not necessarily the best choices.
 <Multi_key> <F> <Left>		: "☚"	U261A		# BLACK LEFT POINTING INDEX
 <Multi_key> <F> <Right>		: "☛"	U261B		# BLACK RIGHT POINTING INDEX
 <Multi_key> <f> <Left>		: "☜"	U261C		# WHITE LEFT POINTING INDEX
@@ -150,6 +154,9 @@ include "%L"
 <Multi_key> <period> <quotedbl>	   	: "∴"	U2234  		# THEREFORE
 <Multi_key> <quotedbl> <period>	   	: "∵"	U2235  		# BECAUSE
 <Multi_key> <percent> <percent>		: "‱"	U2031	# PER TEN THOUSAND (basis points)
+# OK, absolutely cannot believe we made it this long without NABLA or INTEGRAL
+<Multi_key> <ampersand> <d> <e> <l>	: "∇"	U2207	        # NABLA
+<Multi_key> <ampersand> <i> <n> <t>	: "∫"	U222B		# INTEGRAL
 <Multi_key> <asciicircum> <greater>     : "⃗"   U20D7           # COMBINING RIGHT ARROW ABOVE (vector)
 # There's a whole passel of these guys starting at U+1D538 but I have no fonts for those.
 <Multi_key> <bar> <C>                   : "ℂ"   U2102           # DOUBLE-STRUCK CAPITAL C (set of complex numbers)
diff --git a/scan4dups.py b/scan4dups.py
index f504623..00aef78 100644
--- a/scan4dups.py
+++ b/scan4dups.py
@@ -19,33 +19,41 @@ try:
                 break
             word=m.group(1)
             name+=' '+word
-            if listing.has_key(name):
-                dupsfound.append(name)
             startpos+=m.end()
         if startpos<=0:
             continue
         m=re.match(r'[^"]*"(.+)"',line)
         if not m:
             # shouldn't happen, but just in case
-            listing[name]='???'
+            val='???'
             print "couldn't make sense of line: "+line
         else:
-            listing[name]=m.group(1)
-        # THIS IS STILL FAULTY.
-        # What if a long one comes through first, and then a short one?
-        #
-        # Probably have to do two passes: record them all in the hash
-        # (and check for exact duplicates), then go though all the
-        # keys and check all their prefixes.
-        for i in dupsfound:
-            if listing[name]==listing[i]:
-                msg="Redundant definition: "
+            val=m.group(1)
+        if listing.has_key(name):
+            if val != listing[name]:
+                print "Exact conflict found: (%s )[%s][%s]"%(name, 
+                                                             listing[name], val)
             else:
-                msg="Prefix conflict found: "
-            print msg+"(%s )[%s] vs (%s )[%s]"% \
-                (name, listing[name], i, listing[i])
+                print "Redundant definition: (%s )[%s]"%(name, val)
+        else:
+            listing[name]=val
 except StopIteration:
     print "hit end"
-    pass
-print "Done."
+# NOW check for prefix conflicts:
+print "Checking prefixes."
+for key in listing.keys():
+    # print "Key: (%s)"%key
+    pref=''
+    # Careful when splitting.  They key always starts with a space.
+    for word in key.split(" ")[:-1]: # chop the last one; that'll always match.
+        # Skip the empty first entry
+        if not word:
+            continue
+        pref+=" "+word
+        # print "checking (%s)"%pref
+        if listing.has_key(pref):
+            print "Prefix conflict found: " \
+                "(%s )[%s] vs (%s )[%s]"%(pref, listing[pref],
+                                          key, listing[key])
+