--- a/hgext/convert/cvsps.py Sun Jun 15 15:59:27 2008 +0100
+++ b/hgext/convert/cvsps.py Tue Jun 17 12:10:19 2008 -0500
@@ -18,7 +18,7 @@
try:
list.sort(key=key)
except TypeError:
- list.sort(lambda l, r:cmp(key(l), key(r)))
+ list.sort(lambda l, r: cmp(key(l), key(r)))
class logentry(object):
'''Class logentry has the following attributes:
@@ -78,13 +78,13 @@
try:
prefix = file(os.path.join('CVS','Repository')).read().strip()
if prefix == ".":
- prefix=""
+ prefix = ""
directory = prefix
except IOError:
raise logerror('Not a CVS sandbox')
if prefix and not prefix.endswith('/'):
- prefix+='/'
+ prefix += '/'
# Use the Root file in the sandbox, if it exists
try:
@@ -113,9 +113,10 @@
# and
# /pserver/user/server/path
# are mapped to different cache file names.
- cachefile = root.split(":")+[directory, "cache"]
+ cachefile = root.split(":") + [directory, "cache"]
cachefile = ['-'.join(re.findall(r'\w+', s)) for s in cachefile if s]
- cachefile = os.path.join(cachedir, '.'.join([s for s in cachefile if s]))
+ cachefile = os.path.join(cachedir,
+ '.'.join([s for s in cachefile if s]))
if cache == 'update':
try:
@@ -135,8 +136,8 @@
cmd.append('-d%s' % root)
p = root.split(':')[-1]
if not p.endswith('/'):
- p+='/'
- prefix = p+prefix
+ p += '/'
+ prefix = p + prefix
cmd.append(['log', 'rlog'][rlog])
if date:
# no space between option and date string
@@ -206,8 +207,8 @@
# Convert magic branch number to an odd-numbered one
revn = len(rev)
- if revn>3 and (revn%2) == 0 and rev[-2] == 0:
- rev = rev[:-2]+rev[-1:]
+ if revn > 3 and (revn % 2) == 0 and rev[-2] == 0:
+ rev = rev[:-2] + rev[-1:]
rev = tuple(rev)
if rev not in tags:
@@ -244,11 +245,11 @@
d = match.group(1)
if d[2] == '/':
# Y2K
- d = '19'+d
+ d = '19' + d
if len(d.split()) != 3:
# cvs log dates always in GMT
- d = d+' UTC'
+ d = d + ' UTC'
e.date = util.parsedate(d, ['%y/%m/%d %H:%M:%S', '%Y/%m/%d %H:%M:%S', '%Y-%m-%d %H:%M:%S'])
e.author = scache(match.group(2))
e.dead = match.group(3).lower() == 'dead'
@@ -266,7 +267,7 @@
state = 7
elif state == 7:
- # read the revision numbers of branches that start at this revision,
+ # read the revision numbers of branches that start at this revision
# or store the commit log message otherwise
m = re_70.match(line)
if m:
@@ -301,15 +302,15 @@
e.comment = scache('\n'.join(e.comment))
revn = len(e.revision)
- if revn>3 and (revn%2) == 0:
+ if revn > 3 and (revn % 2) == 0:
e.branch = tags.get(e.revision[:-1], [None])[0]
else:
e.branch = None
log.append(e)
- if len(log)%100 == 0:
- ui.status(util.ellipsis('%d %s'%(len(log), e.file), 80)+'\n')
+ if len(log) % 100 == 0:
+ ui.status(util.ellipsis('%d %s' % (len(log), e.file), 80)+'\n')
listsort(log, key=lambda x:(x.rcs, x.revision))
@@ -330,9 +331,10 @@
listsort(log, key=lambda x:x.date)
if oldlog and oldlog[-1].date >= log[0].date:
- raise logerror('Log cache overlaps with new log entries, re-run without cache.')
+ raise logerror('Log cache overlaps with new log entries,'
+ ' re-run without cache.')
- log = oldlog+log
+ log = oldlog + log
# write the new cachefile
ui.note(_('writing cvs log cache %s\n') % cachefile)
@@ -377,14 +379,17 @@
e.comment == c.comment and
e.author == c.author and
e.branch == c.branch and
- (c.date[0]+c.date[1]) <= (e.date[0]+e.date[1]) <= (c.date[0]+c.date[1])+fuzz and
+ ((c.date[0] + c.date[1]) <=
+ (e.date[0] + e.date[1]) <=
+ (c.date[0] + c.date[1]) + fuzz) and
e.file not in files):
c = changeset(comment=e.comment, author=e.author,
branch=e.branch, date=e.date, entries=[])
changesets.append(c)
files = {}
- if len(changesets)%100 == 0:
- ui.status(util.ellipsis('%d %s'%(len(changesets), repr(e.comment)[1:-1]), 80)+'\n')
+ if len(changesets) % 100 == 0:
+ t = '%d %s' % (len(changesets), repr(e.comment)[1:-1])
+ ui.status(util.ellipsis(t, 80) + '\n')
e.Changeset = c
c.entries.append(e)
@@ -402,13 +407,13 @@
nr = len(r)
n = min(nl, nr)
for i in range(n):
- if i+1 == nl and nl<nr:
+ if i + 1 == nl and nl < nr:
return -1
- elif i+1 == nr and nl>nr:
+ elif i + 1 == nr and nl > nr:
return +1
- elif l[i]<r[i]:
+ elif l[i] < r[i]:
return -1
- elif l[i]>r[i]:
+ elif l[i] > r[i]:
return +1
return 0
def entitycompare(l, r):
@@ -419,7 +424,7 @@
# Sort changesets by date
def cscmp(l, r):
- d = sum(l.date)-sum(r.date)
+ d = sum(l.date) - sum(r.date)
if d:
return d
@@ -526,8 +531,8 @@
cc = changeset(author=c.author, branch=m, date=c.date,
comment='convert-repo: CVS merge from branch %s' % c.branch,
entries=[], tags=[], parents=[changesets[branches[m]], c])
- changesets.insert(i+1, cc)
- branches[m] = i+1
+ changesets.insert(i + 1, cc)
+ branches[m] = i + 1
# adjust our loop counters now we have inserted a new entry
n += 1
@@ -540,7 +545,7 @@
# Number changesets
for i, c in enumerate(changesets):
- c.id = i+1
+ c.id = i + 1
ui.status(_('%d changeset entries\n') % len(changesets))