[PATCH 2 of 8] debugrevlog: track empty revlog entries

[Boris Feld]

# HG changeset patch
# User Boris Feld <boris.feld at octobus.net>
# Date 1532711992 -7200
#      Fri Jul 27 19:19:52 2018 +0200
# Node ID 1ec1b66c5db9247ce3666f2b0fcecdd6ef57315f
# Parent  370c3f5021dedb80ded496d39f76f73facd6b7d3
# EXP-Topic sparse-snapshot
# Available At https://bitbucket.org/octobus/mercurial-devel/
#              hg pull https://bitbucket.org/octobus/mercurial-devel/ -r 1ec1b66c5db9
debugrevlog: track empty revlog entries

A revlog entry can be "empty" for two reasons:

* the stored text is empty
* we found another entry with the exact same content

Tracking them separately make revlog content clearer.

diff --git a/mercurial/debugcommands.py b/mercurial/debugcommands.py
--- a/mercurial/debugcommands.py
+++ b/mercurial/debugcommands.py
@@ -2087,6 +2087,8 @@ def debugrevlog(ui, repo, file_=None, **
     nummerges = 0
 
     ### tracks ways the "delta" are build
+    # nodelta
+    numempty = 0
     # full file content
     numfull = 0
     # delta against previous revision
@@ -2133,27 +2135,33 @@ def debugrevlog(ui, repo, file_=None, **
             chainlengths.append(0)
             chainbases.append(r.start(rev))
             chainspans.append(size)
-            numfull += 1
-            addsize(size, fullsize)
+            if size == 0:
+                numempty += 1
+            else:
+                numfull += 1
+                addsize(size, fullsize)
         else:
             chainlengths.append(chainlengths[delta] + 1)
             baseaddr = chainbases[delta]
             revaddr = r.start(rev)
             chainbases.append(baseaddr)
             chainspans.append((revaddr - baseaddr) + size)
-            addsize(size, deltasize)
-            if delta == rev - 1:
-                numprev += 1
-                if delta == p1:
-                    nump1prev += 1
+            if size == 0:
+                numempty += 1
+            else:
+                addsize(size, deltasize)
+                if delta == rev - 1:
+                    numprev += 1
+                    if delta == p1:
+                        nump1prev += 1
+                    elif delta == p2:
+                        nump2prev += 1
+                elif delta == p1:
+                    nump1 += 1
                 elif delta == p2:
-                    nump2prev += 1
-            elif delta == p1:
-                nump1 += 1
-            elif delta == p2:
-                nump2 += 1
-            elif delta != nullrev:
-                numother += 1
+                    nump2 += 1
+                elif delta != nullrev:
+                    numother += 1
 
         # Obtain data on the raw chunks in the revlog.
         segment = r._getsegmentforrevs(rev, rev)[1]
@@ -2174,15 +2182,15 @@ def debugrevlog(ui, repo, file_=None, **
         if size[0] is None:
             size[0] = 0
 
-    numdeltas = numrevs - numfull
+    numdeltas = numrevs - numfull - numempty
     numoprev = numprev - nump1prev - nump2prev
     totalrawsize = datasize[2]
     datasize[2] /= numrevs
     fulltotal = fullsize[2]
     fullsize[2] /= numfull
     deltatotal = deltasize[2]
-    if numrevs - numfull > 0:
-        deltasize[2] /= numrevs - numfull
+    if numdeltas > 0:
+        deltasize[2] /= numdeltas
     totalsize = fulltotal + deltatotal
     avgchainlen = sum(chainlengths) / numrevs
     maxchainlen = max(chainlengths)
@@ -2215,6 +2223,7 @@ def debugrevlog(ui, repo, file_=None, **
     ui.write(('    merges    : ') + fmt % pcfmt(nummerges, numrevs))
     ui.write(('    normal    : ') + fmt % pcfmt(numrevs - nummerges, numrevs))
     ui.write(('revisions     : ') + fmt2 % numrevs)
+    ui.write(('    empty     : ') + fmt % pcfmt(numempty, numrevs))
     ui.write(('    full      : ') + fmt % pcfmt(numfull, numrevs))
     ui.write(('    deltas    : ') + fmt % pcfmt(numdeltas, numrevs))
     ui.write(('revision size : ') + fmt2 % totalsize)
diff --git a/tests/test-debugcommands.t b/tests/test-debugcommands.t
--- a/tests/test-debugcommands.t
+++ b/tests/test-debugcommands.t
@@ -23,7 +23,8 @@
       merges    :  0 ( 0.00%)
       normal    :  3 (100.00%)
   revisions     :  3
-      full      :  3 (100.00%)
+      empty     :  1 (33.33%)
+      full      :  2 (66.67%)
       deltas    :  0 ( 0.00%)
   revision size : 88
       full      : 88 (100.00%)
@@ -42,7 +43,7 @@
   compression ratio :  0
   
   uncompressed data size (min/max/avg) : 0 / 43 / 28
-  full revision size (min/max/avg)     : 0 / 44 / 29
+  full revision size (min/max/avg)     : 44 / 44 / 44
   delta size (min/max/avg)             : 0 / 0 / 0
 #endif