Handle git-rev-list and git-ls-tree with empty changesets (#132, #115).

If the changeset was empty or filtered with no matching files a bunch of
errors would be thrown from git with no proper results being returned
back to gitinspector.

We now pipe stderr (catching the output) and also check the return code
when running these commands.

gitinspector/blame.py

@@ -1,6 +1,6 @@
 # coding: utf-8
 #
-# Copyright © 2012-2015 Ejwa Software. All rights reserved.
+# Copyright © 2012-2017 Ejwa Software. All rights reserved.
 #
 # This file is part of gitinspector.
 #
@@ -123,12 +123,14 @@ PROGRESS_TEXT = N_("Checking how many rows belong to each author (2 of 2): {0:.0
 class Blame(object):
 	def __init__(self, repo, hard, useweeks, changes):
 		self.blames = {}
-		ls_tree_r = subprocess.Popen(["git", "ls-tree", "--name-only", "-r", interval.get_ref()], bufsize=1,
-		                             stdout=subprocess.PIPE).stdout
-		lines = ls_tree_r.readlines()
-		ls_tree_r.close()
+		ls_tree_p = subprocess.Popen(["git", "ls-tree", "--name-only", "-r", interval.get_ref()], bufsize=1,
+		                             stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
+		lines = ls_tree_p.communicate()[0].splitlines()
+		ls_tree_p.stdout.close()
 
+		if ls_tree_p.returncode == 0:
 			progress_text = _(PROGRESS_TEXT)
+
 			if repo != None:
 				progress_text = "[%s] " % repo.name + progress_text
 
@@ -137,8 +139,8 @@ class Blame(object):
 				row = row.encode("latin-1", "replace")
 				row = row.decode("utf-8", "replace").strip("\"").strip("'").strip()
 
-			if FileDiff.get_extension(row) in extensions.get_located() and FileDiff.is_valid_extension(row) and not \
-			   filtering.set_filtered(FileDiff.get_filename(row)):
+				if FileDiff.get_extension(row) in extensions.get_located() and \
+				   FileDiff.is_valid_extension(row) and not filtering.set_filtered(FileDiff.get_filename(row)):
 					blame_command = filter(None, ["git", "blame", "--line-porcelain", "-w"] + \
 							(["-C", "-C", "-M"] if hard else []) +
 					                [interval.get_since(), interval.get_ref(), "--", row])

gitinspector/changes.py

@@ -1,6 +1,6 @@
 # coding: utf-8
 #
-# Copyright © 2012-2015 Ejwa Software. All rights reserved.
+# Copyright © 2012-2017 Ejwa Software. All rights reserved.
 #
 # This file is part of gitinspector.
 #
@@ -183,13 +183,13 @@ class Changes(object):
 
 	def __init__(self, repo, hard):
 		self.commits = []
-		git_log_hashes_r = subprocess.Popen(filter(None, ["git", "rev-list", "--reverse", "--no-merges",
+		git_rev_list_p = subprocess.Popen(filter(None, ["git", "rev-list", "--reverse", "--no-merges",
 		                                  interval.get_since(), interval.get_until(), "HEAD"]), bufsize=1,
-		                                    stdout=subprocess.PIPE).stdout
-		lines = git_log_hashes_r.readlines()
-		git_log_hashes_r.close()
+		                                  stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
+		lines = git_rev_list_p.communicate()[0].splitlines()
+		git_rev_list_p.stdout.close()
 
-		if len(lines) > 0:
+		if git_rev_list_p.returncode == 0 and len(lines) > 0:
 			progress_text = _(PROGRESS_TEXT)
 			if repo != None:
 				progress_text = "[%s] " % repo.name + progress_text

gitinspector/metrics.py

@@ -1,6 +1,6 @@
 # coding: utf-8
 #
-# Copyright © 2012-2015 Ejwa Software. All rights reserved.
+# Copyright © 2012-2017 Ejwa Software. All rights reserved.
 #
 # This file is part of gitinspector.
 #
@@ -44,10 +44,13 @@ class MetricsLogic(object):
 		self.cyclomatic_complexity = {}
 		self.cyclomatic_complexity_density = {}
 
-		ls_tree_r = subprocess.Popen(["git", "ls-tree", "--name-only", "-r", interval.get_ref()], bufsize=1,
-		                             stdout=subprocess.PIPE).stdout
+		ls_tree_p = subprocess.Popen(["git", "ls-tree", "--name-only", "-r", interval.get_ref()], bufsize=1,
+		                             stdout=subprocess.PIPE, stderr=subprocess.STDOUT)
+		lines = ls_tree_p.communicate()[0].splitlines()
+		ls_tree_p.stdout.close()
 
-		for i in ls_tree_r.readlines():
+		if ls_tree_p.returncode == 0:
+			for i in lines:
 				i = i.strip().decode("unicode_escape", "ignore")
 				i = i.encode("latin-1", "replace")
 				i = i.decode("utf-8", "replace").strip("\"").strip("'").strip()