Mercurial > hg-stable
changeset 8661:883f14fcd1df
convert/cvsps: wrap long lines
author | Martin Geisler <mg@lazybytes.net> |
---|---|
date | Sun, 31 May 2009 01:28:18 +0200 |
parents | b9ec04eb2aa3 |
children | eaee3491ce11 |
files | hgext/convert/cvsps.py |
diffstat | 1 files changed, 24 insertions(+), 12 deletions(-) [+] |
line wrap: on
line diff
--- a/hgext/convert/cvsps.py Sun May 31 01:26:23 2009 +0200 +++ b/hgext/convert/cvsps.py Sun May 31 01:28:18 2009 +0200 @@ -34,7 +34,8 @@ .revision - revision number as tuple .tags - list of tags on the file .synthetic - is this a synthetic "file ... added on ..." revision? - .mergepoint- the branch that has been merged from (if present in rlog output) + .mergepoint- the branch that has been merged from + (if present in rlog output) ''' def __init__(self, **entries): self.__dict__.update(entries) @@ -105,14 +106,18 @@ re_00 = re.compile('RCS file: (.+)$') re_01 = re.compile('cvs \\[r?log aborted\\]: (.+)$') re_02 = re.compile('cvs (r?log|server): (.+)\n$') - re_03 = re.compile("(Cannot access.+CVSROOT)|(can't create temporary directory.+)$") + re_03 = re.compile("(Cannot access.+CVSROOT)|" + "(can't create temporary directory.+)$") re_10 = re.compile('Working file: (.+)$') re_20 = re.compile('symbolic names:') re_30 = re.compile('\t(.+): ([\\d.]+)$') re_31 = re.compile('----------------------------$') - re_32 = re.compile('=============================================================================$') + re_32 = re.compile('=======================================' + '======================================$') re_50 = re.compile('revision ([\\d.]+)(\s+locked by:\s+.+;)?$') - re_60 = re.compile(r'date:\s+(.+);\s+author:\s+(.+);\s+state:\s+(.+?);(\s+lines:\s+(\+\d+)?\s+(-\d+)?;)?(.*mergepoint:\s+([^;]+);)?') + re_60 = re.compile(r'date:\s+(.+);\s+author:\s+(.+);\s+state:\s+(.+?);' + r'(\s+lines:\s+(\+\d+)?\s+(-\d+)?;)?' + r'(.*mergepoint:\s+([^;]+);)?') re_70 = re.compile('branches: (.+);$') file_added_re = re.compile(r'file [^/]+ was (initially )?added on branch') @@ -282,7 +287,8 @@ if re_31.match(line): state = 5 else: - assert not re_32.match(line), _('must have at least some revisions') + assert not re_32.match(line), _('must have at least ' + 'some revisions') elif state == 5: # expecting revision number and possibly (ignored) lock indication @@ -308,7 +314,9 @@ if len(d.split()) != 3: # cvs log dates always in GMT 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.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' @@ -443,7 +451,8 @@ .parents - list of one or two parent changesets .tags - list of tags on this changeset .synthetic - from synthetic revision "file ... added on branch ..." - .mergepoint- the branch that has been merged from (if present in rlog output) + .mergepoint- the branch that has been merged from + (if present in rlog output) ''' def __init__(self, **entries): self.__dict__.update(entries) @@ -689,9 +698,10 @@ def debugcvsps(ui, *args, **opts): - '''Read CVS rlog for current directory or named path in repository, and - convert the log to changesets based on matching commit log entries and dates.''' - + '''Read CVS rlog for current directory or named path in + repository, and convert the log to changesets based on matching + commit log entries and dates. + ''' if opts["new_cache"]: cache = "write" elif opts["update_cache"]: @@ -724,7 +734,8 @@ if opts["ancestors"]: if cs.branch not in branches and cs.parents and cs.parents[0].id: - ancestors[cs.branch] = changesets[cs.parents[0].id-1].branch, cs.parents[0].id + ancestors[cs.branch] = (changesets[cs.parents[0].id-1].branch, + cs.parents[0].id) branches[cs.branch] = cs.id # limit by branches @@ -736,7 +747,8 @@ # bug-for-bug compatibility with cvsps. ui.write('---------------------\n') ui.write('PatchSet %d \n' % cs.id) - ui.write('Date: %s\n' % util.datestr(cs.date, '%Y/%m/%d %H:%M:%S %1%2')) + ui.write('Date: %s\n' % util.datestr(cs.date, + '%Y/%m/%d %H:%M:%S %1%2')) ui.write('Author: %s\n' % cs.author) ui.write('Branch: %s\n' % (cs.branch or 'HEAD')) ui.write('Tag%s: %s \n' % (['', 's'][len(cs.tags)>1],