# hg.py - repository classes for mercurial

#

# Copyright 2005 Matt Mackall <mpm@selenic.com>

#

# This software may be used and distributed according to the terms

# of the GNU General Public License, incorporated herein by reference.

import sys, struct, sha, socket, os, time, base64, re, urllib2

import urllib

from mercurial import byterange

from mercurial.transaction import *

from mercurial.revlog import *

from difflib import SequenceMatcher

class filelog(revlog):

    def __init__(self, opener, path):

        s = self.encodepath(path)

        revlog.__init__(self, opener, os.path.join("data", s + "i"),

                        os.path.join("data", s))

    def encodepath(self, path):

        s = sha.sha(path).digest()

        s = base64.encodestring(s)[:-3]

        s = re.sub("\+", "%", s)

        s = re.sub("/", "_", s)

        return s

    def read(self, node):

        return self.revision(node)

    def add(self, text, transaction, link, p1=None, p2=None):

        return self.addrevision(text, transaction, link, p1, p2)

    def annotate(self, node):

        revs = []

        while node != nullid:

            revs.append(node)

            node = self.parents(node)[0]

        revs.reverse()

        prev = []

        annotate = []

        for node in revs:

            curr = self.read(node).splitlines(1)

            linkrev = self.linkrev(node)

            sm = SequenceMatcher(None, prev, curr)

            offset = 0

            for o, m, n, s, t in sm.get_opcodes():

                if o in ('insert','replace'):

                    annotate[m+offset:n+offset] = \

                        [ (linkrev, l) for l in curr[s:t]]

                    if o == 'insert':

                        offset += m-n

                elif o == 'delete':

                    del annotate[m+offset:n+offset]

                    offset -= m-n

            assert len(annotate) == len(curr)

            prev = curr

        return annotate

class manifest(revlog):

    def __init__(self, opener):

        self.mapcache = None

        self.listcache = None

        self.addlist = None

        revlog.__init__(self, opener, "00manifest.i", "00manifest.d")

    def read(self, node):

        if self.mapcache and self.mapcache[0] == node:

            return self.mapcache[1].copy()

        text = self.revision(node)

        map = {}

        self.listcache = (text, text.splitlines(1))

        for l in self.listcache[1]:

            (f, n) = l.split('\0')

            map[f] = bin(n[:40])

        self.mapcache = (node, map)

        return map

    def diff(self, a, b):

        # this is sneaky, as we're not actually using a and b

        if self.listcache and len(self.listcache[0]) == len(a):

            return mdiff.diff(self.listcache[1], self.addlist, 1)

        else:

            return mdiff.textdiff(a, b)

    def add(self, map, transaction, link, p1=None, p2=None):

        files = map.keys()

        files.sort()

        self.addlist = ["%s\000%s\n" % (f, hex(map[f])) for f in files]

        text = "".join(self.addlist)

        n = self.addrevision(text, transaction, link, p1, p2)

        self.mapcache = (n, map)

        self.listcache = (text, self.addlist)

        return n

class changelog(revlog):

    def __init__(self, opener):

        revlog.__init__(self, opener, "00changelog.i", "00changelog.d")

    def extract(self, text):

        if not text:

            return (nullid, "", "0", [], "")

        last = text.index("\n\n")

        desc = text[last + 2:]

        l = text[:last].splitlines()

        manifest = bin(l[0])

        user = l[1]

        date = l[2]

        files = l[3:]

        return (manifest, user, date, files, desc)

    def read(self, node):

        return self.extract(self.revision(node))

    def add(self, manifest, list, desc, transaction, p1=None, p2=None):

        user = (os.environ.get("HGUSER") or

                os.environ.get("EMAIL") or

                os.environ.get("LOGNAME", "unknown") + '@' + socket.getfqdn())

        date = "%d %d" % (time.time(), time.timezone)

        list.sort()

        l = [hex(manifest), user, date] + list + ["", desc]

        text = "\n".join(l)

        return self.addrevision(text, transaction, self.count(), p1, p2)

class dircache:

    def __init__(self, opener, ui):

        self.opener = opener

        self.dirty = 0

        self.ui = ui

        self.map = None

    def __del__(self):

        if self.dirty: self.write()

    def __getitem__(self, key):

        try:

            return self.map[key]

        except TypeError:

            self.read()

            return self[key]

    def read(self):

        if self.map is not None: return self.map

        self.map = {}

        try:

            st = self.opener("dircache").read()

        except: return

        pos = 0

        while pos < len(st):

            e = struct.unpack(">llll", st[pos:pos+16])

            l = e[3]

            pos += 16

            f = st[pos:pos + l]

            self.map[f] = e[:3]

            pos += l

    def update(self, files):

        if not files: return

        self.read()

        self.dirty = 1

        for f in files:

            try:

                s = os.stat(f)

                self.map[f] = (s.st_mode, s.st_size, s.st_mtime)

            except IOError:

                self.remove(f)

    def taint(self, files):

        if not files: return

        self.read()

        self.dirty = 1

        for f in files:

            self.map[f] = (0, -1, 0)

    def remove(self, files):

        if not files: return

        self.read()

        self.dirty = 1

        for f in files:

            try:

                del self.map[f]

            except KeyError:

                self.ui.warn("Not in dircache: %s\n" % f)

                pass

    def clear(self):

        self.map = {}

        self.dirty = 1

    def write(self):

        st = self.opener("dircache", "w")

        for f, e in self.map.items():

            e = struct.pack(">llll", e[0], e[1], e[2], len(f))

            st.write(e + f)

        self.dirty = 0

    def copy(self):

        self.read()

        return self.map.copy()

# used to avoid circular references so destructors work

def opener(base):

    p = base

    def o(path, mode="r"):

        if p[:7] == "http://":

            f = os.path.join(p, urllib.quote(path))

            return httprangereader(f)

        f = os.path.join(p, path)

        if mode != "r" and os.path.isfile(f):

            s = os.stat(f)

            if s.st_nlink > 1:

                file(f + ".tmp", "w").write(file(f).read())

                os.rename(f+".tmp", f)

        return file(f, mode)

    return o

class localrepository:

    def __init__(self, ui, path=None, create=0):

        self.remote = 0

        if path and path[:7] == "http://":

            self.remote = 1

            self.path = path

        else:

            if not path:

                p = os.getcwd()

                while not os.path.isdir(os.path.join(p, ".hg")):

                    p = os.path.dirname(p)

                    if p == "/": raise "No repo found"

                path = p

            self.path = os.path.join(path, ".hg")

        self.root = path

        self.ui = ui

        if create:

            os.mkdir(self.path)  

            os.mkdir(self.join("data"))

        self.opener = opener(self.path)

        self.manifest = manifest(self.opener)

        self.changelog = changelog(self.opener)

        self.ignorelist = None

        self.tags = None

        if not self.remote:

            self.dircache = dircache(self.opener, ui)

            try:

                self.current = bin(self.opener("current").read())

            except IOError:

                self.current = None

    def setcurrent(self, node):

        self.current = node

        self.opener("current", "w").write(hex(node))

    def ignore(self, f):

        if self.ignorelist is None:

            self.ignorelist = []

            try:

                l = open(os.path.join(self.root, ".hgignore"))

                for pat in l:

                    if pat != "\n":

                        self.ignorelist.append(re.compile(pat[:-1]))

            except IOError: pass

        for pat in self.ignorelist:

            if pat.search(f): return True

        return False

    def lookup(self, key):

        if self.tags is None:

            self.tags = {}

            try:

                fl = self.file(".hgtags")

                for l in fl.revision(fl.tip()).splitlines():

                    if l:

                        n, k = l.split(" ")

                        self.tags[k] = bin(n)

            except KeyError: pass

        try:

            return self.tags[key]

        except KeyError:

            return self.changelog.lookup(key)

    def join(self, f):

        return os.path.join(self.path, f)

    def file(self, f):

        return filelog(self.opener, f)

    def transaction(self):

        return transaction(self.opener, self.join("journal"))

    def commit(self, parent, update = None, text = ""):

        tr = self.transaction()

        try:

            remove = [ l[:-1] for l in self.opener("to-remove") ]

            os.unlink(self.join("to-remove"))

        except IOError:

            remove = []

        if update == None:

            update = self.diffdir(self.root, parent)[0]

        # check in files

        new = {}

        linkrev = self.changelog.count()

        for f in update:

            self.ui.note(f + "\n")

            try:

                t = file(f).read()

            except IOError:

                remove.append(f)

                continue

            r = self.file(f)

            new[f] = r.add(t, tr, linkrev)

        # update manifest

        mmap = self.manifest.read(self.manifest.tip())

        mmap.update(new)

        for f in remove:

            del mmap[f]

        mnode = self.manifest.add(mmap, tr, linkrev)

        # add changeset

        new = new.keys()

        new.sort()

        edittext = text + "\n"+"".join(["HG: changed %s\n" % f for f in new])

        edittext += "".join(["HG: removed %s\n" % f for f in remove])

        edittext = self.ui.edit(edittext)

        n = self.changelog.add(mnode, new, edittext, tr)

        tr.close()

        self.setcurrent(n)

        self.dircache.update(new)

        self.dircache.remove(remove)

    def checkdir(self, path):

        d = os.path.dirname(path)

        if not d: return

        if not os.path.isdir(d):

            self.checkdir(d)

            os.mkdir(d)

    def checkout(self, node):

        # checkout is really dumb at the moment

        # it ought to basically merge

        change = self.changelog.read(node)

        mmap = self.manifest.read(change[0])

        l = mmap.keys()

        l.sort()

        stats = []

        for f in l:

            self.ui.note(f + "\n")

            r = self.file(f)

            t = r.revision(mmap[f])

            try:

                file(f, "w").write(t)

            except:

                self.checkdir(f)

                file(f, "w").write(t)

        self.setcurrent(node)

        self.dircache.clear()

        self.dircache.update(l)

    def diffdir(self, path, changeset):

        changed = []

        mf = {}

        added = []

        if changeset:

            change = self.changelog.read(changeset)

            mf = self.manifest.read(change[0])

        if changeset == self.current:

            dc = self.dircache.copy()

        else:

            dc = dict.fromkeys(mf)

        def fcmp(fn):

            t1 = file(os.path.join(self.root, fn)).read()

            t2 = self.file(fn).revision(mf[fn])

            return cmp(t1, t2)

        for dir, subdirs, files in os.walk(self.root):

            d = dir[len(self.root)+1:]

            if ".hg" in subdirs: subdirs.remove(".hg")

            for f in files:

                fn = os.path.join(d, f)

                try: s = os.stat(os.path.join(self.root, fn))

                except: continue

                if fn in dc:

                    c = dc[fn]

                    del dc[fn]

                    if not c:

                        if fcmp(fn):

                            changed.append(fn)

                    elif c[1] != s.st_size:

                        changed.append(fn)

                    elif c[0] != s.st_mode or c[2] != s.st_mtime:

                        if fcmp(fn):

                            changed.append(fn)

                else:

                    if self.ignore(fn): continue

                    added.append(fn)

        deleted = dc.keys()

        deleted.sort()

        return (changed, added, deleted)

    def diffrevs(self, node1, node2):

        changed, added = [], []

        change = self.changelog.read(node1)

        mf1 = self.manifest.read(change[0])

        change = self.changelog.read(node2)

        mf2 = self.manifest.read(change[0])

        for fn in mf2:

            if mf1.has_key(fn):

                if mf1[fn] != mf2[fn]:

                    changed.append(fn)

                del mf1[fn]

            else:

                added.append(fn)

        deleted = mf1.keys()

        deleted.sort()

        return (changed, added, deleted)

    def add(self, list):

        self.dircache.taint(list)

    def remove(self, list):

        dl = self.opener("to-remove", "a")

        for f in list:

            dl.write(f + "\n")

    def branches(self, nodes):

        if not nodes: nodes = [self.changelog.tip()]

        b = []

        for n in nodes:

            t = n

            while n:

                p = self.changelog.parents(n)

                if p[1] != nullid or p[0] == nullid:

                    b.append((t, n, p[0], p[1]))

                    break

                n = p[0]

        return b

    def between(self, pairs):

        r = []

        for top, bottom in pairs:

            n, l, i = top, [], 0

            f = 1

            while n != bottom:

                p = self.changelog.parents(n)[0]

                if i == f:

                    l.append(n) 

                    f = f * 2

                n = p

                i += 1

            r.append(l)

        return r

    def newer(self, nodes):

        m = {}

        nl = []

        cl = self.changelog

        t = l = cl.count()

        for n in nodes:

            l = min(l, cl.rev(n))

        for i in xrange(l, t):

            n = cl.node(i)

            for p in cl.parents(n):

                    nl.append(n)

        return nl