Class P4

Description

Main interface to the Python client API.

This module provides an object-oriented interface to Helix Core Server, the Perforce version control system. Data is returned in Python arrays and dictionaries (hashes) and input can also be supplied in these formats.

Each P4 object represents a connection to the Helix Core Server, and multiple commands may be executed (serially) over a single connection (which of itself can result in substantially improved performance if executing long sequences of Helix Core Server commands).

1. Instantiate your P4 object.

2. Specify your Helix Core Server client environment:

   • client
   • host
   • password
   • port
   • user

3. Set any options to control output or error handling:

   • exception_level

4. Connect to the Perforce service.

   The Helix Core Server protocol is not designed to support multiple concurrent queries over the same connection. Multithreaded applications that use the C++ API or derived APIs (including P4Python) should ensure that a separate connection is used for each thread, or that only one thread may use a shared connection at a time.

5. Run your Helix Core Server commands.
6. Disconnect from the Perforce service.

Instance Attributes

p4.api_level -> int

Contains the API compatibility level desired. This is useful when writing scripts using Helix Core Server commands that do not yet support tagged output. In these cases, upgrading to a later server that supports tagged output for the commands in question can break your script. Using this method allows you to lock your script to the output format of an older Helix Core Server release and facilitate seamless upgrades. Must be called before calling P4.connect().

from P4 import P4
p4 = P4()
p4.api_level = 67 # Lock to 2010.1 format
p4.connect()
...
p4.disconnect

For more information about the API integer levels, see Protocol levels: server/client in the Helix Core Server Administrator Guide.

p4.charset -> string

Contains the character set to use when connecting to a Unicode enabled server. Do not use when working with non-Unicode-enabled servers. By default, the character set is the value of the P4CHARSET environment variable. If the character set is invalid, this method raises a P4Exception.

from P4 import P4
p4 = P4()
p4.client = "www"
p4.charset = "iso8859-1"
p4.connect()
p4.run_sync()
p4.disconnect()

p4.client -> string

Contains the name of your client workspace. By default, this is the value of the P4CLIENT taken from any P4CONFIG file present, or from the environment according to the normal Helix Core Server conventions.

p4.cwd -> string

Contains the current working directly. Can be set prior to executing any Helix Core Server command. Sometimes necessary if your script executes a chdir() as part of its processing.

from P4 import P4
p4 = P4()
p4.cwd = "/home/bruno"

p4.disable_tmp_cleanup -> string

Invoke this prior to connecting if you need to use multiple P4 connections in parallel in a multi-threaded Python application. This applies to platforms running Linux and macOS where P4Python establishes multiple connections to the same Helix Core Server.

from P4 import P4
p4 = P4()
p4.disable_tmp_cleanup()
p4.connect()
...
p4.disconnect()

p4.encoding -> string

When decoding strings from a non-Unicode server, strings are assumed to be encoded in UTF8. To use another encoding, set p4.encoding to a legal Python encoding, or raw to receive Python bytes instead of a Unicode string. Available only when compiled with Python 3.

p4.errors -> list (read-only)

Returns an array containing the error messages received during execution of the last command.

from P4 import P4, P4Exceptionp4 = P4()

try:
  p4.connect()
  p4.exception_level = 1
  # ignore "File(s) up-to-date"s
  files = p4.run_sync()

except P4Exception:
  for e in p4.errors:
    print e

finally:
  p4.disconnect()

p4.exception_level -> int

Configures the events which give rise to exceptions. The following three levels are supported:

• 0 : disables all exception handling and makes the interface completely procedural; you are responsible for checking the p4.errors and p4.warnings arrays.
• 1 : causes exceptions to be raised only when errors are encountered.
• 2 : causes exceptions to be raised for both errors and warnings. This is the default.

For example:

from P4 import P4
p4 = P4()
p4.exception_level = 1
p4.connect()   # P4Exception on failure
p4.run_sync()  # File(s) up-to-date is a warning - no exception raised
p4.disconnect()

p4.handler -> handler

Set the output handler to a subclass of P4.OutputHandler.

p4.host -> string

Contains the name of the current host. It defaults to the value of P4HOST taken from any P4CONFIG file present, or from the environment as per the usual Helix Core Server convention. Must be called before connecting to the Helix Core Server.

from P4 import P4
p4 = P4()
p4.host = "workstation123.perforce.com"
p4.connect()
...
p4.disconnect()

p4.ignore_file -> string

Contains the path of the ignore file. It defaults to the value of P4IGNORE. Set P4.ignore_file prior to calling P4.is_ignored().

from P4 import P4
p4 = P4()
p4.connect()
p4.ignore_file = "/home/bruno/workspace/.ignore"
p4.disconnect()

p4.input -> string | dict | list

Contains input for the next command.

Set this attribute prior to running a command that requires input from the user. When the command requests input, the specified data is supplied to the command. Typically, commands of the form p4 cmd -i are invoked using the P4.save_<spectype>() methods, which retrieve the value from p4.input internally; there is no need to set p4.input when using the P4.save_<spectype>() shortcuts.

You may pass a string, a hash, or (for commands that take multiple inputs from the user) an array of strings or hashes. If you pass an array, note that the first element of the array will be popped each time Helix Core Server asks the user for input.

For example, the following code supplies a description for the default changelist and then submits it to the depot:

from P4 import P4
p4 = P4()
p4.connect()
change = p4.run_change( "-o" )[0]
change[ "Description" ] = "Autosubmitted changelist"
p4.input = change
p4.run_submit( "-i" )
p4.disconnect()

p4.maxlocktime -> int

Limit the amount of time (in milliseconds) spent during data scans to prevent the server from locking tables for too long. Commands that take longer than the limit will be aborted. The limit remains in force until you disable it by setting it to zero. See p4 help maxlocktime for information on the commands that support this limit.

p4.maxresults -> int

Limit the number of results Helix Core Server permits for subsequent commands. Commands that produce more than this number of results will be aborted. The limit remains in force until you disable it by setting it to zero. See p4 help maxresults for information on the commands that support this limit.

p4.maxscanrows -> int

Limit the number of database records Helix Core Server scans for subsequent commands. Commands that attempt to scan more than this number of records will be aborted. The limit remains in force until you disable it by setting it to zero. See p4 help maxscanrows for information on the commands that support this limit.

p4.messages -> list (read-only)

Returns a list of P4.Message objects, one for each message (info, warning or error) sent by the server.

p4.p4config_file -> string (read-only)

Contains the name of the current P4CONFIG file, if any. This attribute cannot be set.

p4.password -> string

Contains your Helix Core Server password or login ticket. If not used, takes the value of P4PASSWD from any P4CONFIG file in effect, or from the environment according to the normal Helix Core Server conventions.

This password is also used if you later call p4.run_login() to log in using the 2003.2 and later ticket system. After running p4.run_login(), the attribute contains the ticket allocated by the server.

from P4 import P4
p4 = P4()
p4.password = "mypass"
p4.connect()
p4.run_login()

p4.port -> string

Contains the host and port of the Helix Core Server to which you want to connect. It defaults to the value of P4PORT in any P4CONFIG file in effect, and then to the value of P4PORT taken from the environment.

from P4 import P4
p4 = P4()
p4.port = "localhost:1666"
p4.connect()
...

p4.prog -> string

Contains the name of the program, as reported to Helix Core Server system administrators running p4 monitor show -e. The default is unnamed p4-python script.

from P4 import P4
p4 = P4()
p4.prog = "sync-script"
print p4.prog
p4.connect
...

p4.progress -> progress

Set the progress indicator to a subclass of P4.Progress.

p4.server_case_insensitive -> boolean

Detects whether or not the server is case-sensitive.

p4.server_level -> int (read-only)

Returns the current Helix Core Server level. Each iteration of the Helix Core Server is given a level number. As part of the initial communication this value is passed between the client application and the Helix Core Server. This value is used to determine the communication that the Helix Core Server will understand. All subsequent requests can therefore be tailored to meet the requirements of this server level.

This attribute is 0 before the first command is run, and is set automatically after the first communication with the server.

For more information about the Helix Core Server version levels, see Protocol levels: server/client in the Helix Core Server Administrator Guide.

p4.server_unicode -> boolean

Detects whether or not the server is in Unicode mode.

p4.streams -> int

If 1 or True, p4.streams enables support for streams. By default, streams support is enabled at 2011.1 or higher (api_level >= 70). Raises a P4Exception if you attempt to enable streams on a pre-2011.1 server. You can enable or disable support for streams both before and after connecting to the server.

from P4 import P4
p4 = P4()
p4.streams = False
print p4.streams

p4.tagged -> int

If 1 or True, p4.tagged enables tagged output. By default, tagged output is on.

from P4 import P4
p4 = P4()
p4.tagged = False
print p4.tagged

p4.ticket_file -> string

Contains the location of the P4TICKETS file.

p4.track -> boolean

If set to 1 or True, p4.track indicates that server performance tracking is enabled for this connection. By default, performance tracking is disabled.

p4.track_output -> list (read-only)

If performance tracking is enabled with p4.track, returns an array containing the performance data received during execution of the last command.

from P4 import P4
p4 = P4()
p4.track = 1
p4.run_info()
print p4.track_output

p4.user -> string

Contains the Helix Core Server username. It defaults to the value of P4USER taken from any P4CONFIG file present, or from the environment as per the usual Helix Core Server convention.

from P4 import P4
p4 = P4()
p4.user = "bruno"
p4.connect()
...
p4.disconnect()

p4.version -> string

Contains the version of the program, as reported to Helix Core Server system administrators in the server log.

from P4 import P4
p4 = P4()
p4.version = "123"
print p4.version
p4.connect()
...
p4.disconnect()

p4.warnings -> list (read-only)

Contains the array of warnings that arose during execution of the last command.

from P4 import P4, P4Exception
p4 = P4()

try:
  p4.connect()
  p4.exception_level = 2 # File(s) up-to-date is a warning
  files = p4.run_sync()

except P4Exception, ex:
  for w in p4.warnings:
      print w

finally:
  p4.disconnect()

Class Methods

P4.P4()

Construct a new P4 object. For example:

import P4
p4 = P4.P4()

P4.clone( arguments…​ )

Clone from another Perforce service into a local Helix Core Server, and returns a new P4 object.

P4.clone() requires specification of the port of the source Perforce service from which files and version history should be cloned from, and either a remotespec or a filespec that specify which files and history to clone. For example, to clone using a remotespec:

import P4
p4 = P4.clone( "-p", "port", "-r", "remotespec" )

or to clone using a filespec:

import P4
p4 = P4.clone( "-p", "port", "-f", "filespec" )

The cloned instance inherits the case sensitivity and Unicode settings from the source Perforce service.

P4.identify()

Return the version of P4Python that you are using.

python -c "from P4 import P4; print P4.identify()"

The read-only string attributes PATCHLEVEL and OS are also available to test an installation of P4Python without having to parse the output of P4.identify().

If applicable, P4.identify() also reports the version of the OpenSSL library used for building the underlying Helix C/C++ API with which P4Python was built.

P4.init( [arguments] )

Initializes a new, personal (local) Helix Core Server, and returns a new P4 object.

Without any arguments, P4.init() creates a new DVCS server in the current working directory, using the settings for case sensitivity and Unicode support from current environment variables.

P4.init() accepts the following keyword arguments:

Keyword	Explanation	Example
client	Workspace and server name	client="sknop-dvcs"
user	Helix Core Server username used for pushing	user="sven_erik_knop"
directory	local path of the root directory for the new server	directory="/tmp/test-dvcs"
casesensitive	specify case sensitivity	casesensitive=False
unicode	specify whether Unicode is enabled	unicode=True

import P4
p4 = P4.init( directory="/Users/sknop/dvcs/" )
p4.connect()
# ...
p4.disconnect()

The P4 instance returned by P4.init() has the port, user, and client workspace already set; all that is required for you is to connect to the server to perform any commands. Connection is not automatic, to give you an opportunity to set any protocol parameters; these parameters can only be set once before a connection is established.

p4.iterate_<spectype>( arguments ) -> P4.Spec

The iterate_<spectype>() methods are shortcut methods that allow you to quickly iterate through clients, labels, branches, etc. Valid <spectypes> are clients, labels, branches, changes, streams, jobs, users, groups, depots and servers. Valid arguments are any arguments that would be valid for the corresponding run_<spectype>() command.

For example:

for client in p4.iterate_clients():
  # do something with the client spec

is equivalent to:

for c in p4.run_clients():
  client = p4.fetch_client( c['client'] )

Instance Methods

p4.at_exception_level()

In the context of a with statement, temporarily set the exception level for a block. For example:

from P4 import P4
p4 = P4()
p4.connect()
with p4.at_exception_level( P4.RAISE_ERRORS ):
  # no exceptions for warnings
  p4.run_sync( "//depot/main/..." )

# exceptions back to normal...

p4.connect()

Initializes the Helix Core Server client and connects to the server.

If the connection is successfully established, returns None. If the connection fails and P4.exception_level is 0, returns False, otherwise raises a P4Exception. If already connected, prints a message.

from P4 import P4
p4 = P4()
p4.connect()
...
p4.disconnect()

P4.connect() returns a context management object that is usable with a with statement within a block; after the block is finished, the connection is automatically disconnected:

import P4
p4 = P4.P4()
with p4.connect():
  # block in context of connection
  ...

# p4 is disconnected outside the block
...

p4.connected() -> boolean

Returns true if connected to the Helix Core Server and the connection is alive, otherwise false.

from P4 import P4
p4 = P4()

print p4.connected()
p4.connect()
print p4.connected()

p4.delete_<spectype>( [ options ], name) -> list

The delete_<spectype>() methods are shortcut methods that allow you to delete the definitions of clients, labels, branches, etc. These methods are equivalent to:

p4.run( "<spectype>", '-d', [options], "spec name" )

The following code uses P4.delete_client() to delete client workspaces that have not been accessed in more than 365 days:

# Important: this code sample assumes that the client time zone and server time zone are the same
from P4 import P4, P4Exception
from datetime import datetime, timedelta

now = datetime.now()
p4 = P4()

try:
  p4.connect()
  for client in p4.run_clients():
    atime = datetime.utcfromtimestamp( int( client[ "Access" ] ) )
    # If the client has not been accessed for a year, delete it
    if ( atime + timedelta( 365 ) ) < now :
      p4.delete_client( '-f', client[ "client" ] )

except P4Exception:
  for e in p4.errors:
    print e

finally:
  p4.disconnect()

p4.disconnect()

Disconnect from the Helix Core Server. Call this method before exiting your script.

from P4 import P4
p4 = P4()

p4.connect()
...
p4.disconnect()

p4.env( var )

Get the value of a Helix Core Server environment variable, taking into account P4CONFIG files and (on Windows or macOS) the registry or user preferences.

from P4 import P4
p4 = P4()

print p4.env( "P4PORT" )

p4.fetch_<spectype>() -> P4.Spec

The fetch_<spectype>() methods are shortcuts for running p4.run( "<spectype>", "-o" ).pop( 0 ). For example:

label      = p4.fetch_label( "labelname" )
change     = p4.fetch_change( changeno )
clientspec = p4.fetch_client( "clientname" )

are equivalent to:

label      = p4.run( "label", "-o", "labelname" )[0]
change     = p4.run( "change", "-o", changeno )[0]
clientspec = p4.run( "client", "-o", "clientname" )[0]

p4.format_spec( "<spectype>", dict ) -> string

Converts the fields in the dict containing the elements of a Helix Core Server form (spec) into the string representation familiar to users. The first argument is the type of spec to format: for example, client, branch, label, and so on. The second argument is the hash to parse.

There are shortcuts available for this method. You can use p4.format_<spectype>( dict ) instead of p4.format_spec( "<spectype>", dict), where <spectype> is the name of a Helix Core Server spec, such as client, label, etc.

p4.format_<spectype>( dict ) -> string

The format_<spectype>() methods are shortcut methods that allow you to quickly fetch the definitions of clients, labels, branches, etc. They’re equivalent to:

p4.format_spec( "<spectype>", dict )

p4.is_ignored( "<path>" ) -> boolean

Returns true if the <path> is ignored via the P4IGNORE feature. The <path> can be a local relative or absolute path.

from P4 import P4
p4 = P4()

p4.connect()
if ( p4.is_ignored( "/home/bruno/workspace/file.txt" ):
  print "Ignored."
else:
  print "Not ignored."

p4.disconnect()

p4.parse_spec( "<spectype>", string ) -> P4.Spec

Parses a Helix Core Server form (spec) in text form into a Python dict using the spec definition obtained from the server. The first argument is the type of spec to parse: client, branch, label, and so on. The second argument is the string buffer to parse.

There are shortcuts available for this method. You can use:

p4.parse_<spectype>( buf )

instead of:

p4.parse_spec( "<spectype>", buf )

where <spectype> is one of client, branch, label, and so on.

p4.parse_<spectype>( string ) -> P4.Spec

This is equivalent to:

p4.parse_spec( "<spectype>", string )

For example, parse_job( myJob ) converts the String representation of a job spec into a Spec object.

To parse a spec, P4 needs to have the spec available. When not connected to the Helix Core Server, P4 assumes the default format for the spec, which is hardcoded. This assumption can fail for jobs if the server’s jobspec has been modified. In this case, your script can load a job from the server first with the command p4.fetch_job( 'somename' ), and P4 will cache and use the spec format in subsequent p4.parse_job() calls.

p4.run( "<cmd>", [arg, …​] )

Base interface to all the run methods in this API. Runs the specified Helix Core Server command with the arguments supplied. Arguments may be in any form as long as they can be converted to strings by str(). However, each command's options should be passed as quoted and comma-separated strings, with no leading space. For example:

p4.run("print","-o","test-print","-q","//depot/Jam/MAIN/src/expand.c")

Failing to pass options in this way can result in confusing error messages.

The p4.run() method returns a list of results whether the command succeeds or fails; the list may, however, be empty. Whether the elements of the array are strings or dictionaries depends on:

1. server support for tagged output for the command, and
2. whether tagged output was disabled by calling p4.tagged = False.

In the event of errors or warnings, and depending on the exception level in force at the time, p4.run() raises a P4Exception. If the current exception level is below the threshold for the error/warning, p4.run() returns the output as normal and the caller must explicitly review p4.errors and p4.warnings to check for errors or warnings.

from P4 import P4
p4 = P4()
p4.connect()
spec = p4.run( "client", "-o" )[0]
p4.disconnect()

Shortcuts are available for p4.run(). For example:

p4.run_command(args)

is equivalent to:

p4.run( "command", args )

There are also some shortcuts for common commands such as editing Helix Core Server forms and submitting. For example, this:

from P4 import P4
p4 = P4()
p4.connect()
clientspec = p4.run_client( "-o" )[0]
clientspec[ "Description" ] = "Build client"
p4.input = clientspec
p4.run_client( "-i" )
p4.disconnect()

…​may be shortened to:

from P4 import P4
p4 = P4()
p4.connect()
clientspec = p4.fetch_client()
clientspec[ "Description" ] = "Build client"
p4.save_client( clientspec )
p4.disconnect()

The following are equivalent:

Shortcut	Equivalent to
p4.delete_<spectype>()	p4.run( "<spectype>", "-d ")
p4.fetch_<spectype>()	p4.run( "<spectype>", "-o ")[0]
p4.save_<spectype>( spec )	p4.input = spec ​p4.run( "<spectype>", "-i")

As the commands associated with p4.fetch_<spectype>() typically return only one item, these methods do not return an array, but instead return the first result element.

For convenience in submitting changelists, changes returned by p4.fetch_change() can be passed to p4.run_submit(). For example:

from P4 import P4
p4 = P4()
p4.connect()

spec = p4.fetch_change()
spec[ "Description" ] = "Automated change"
p4.run_submit( spec )
p4.disconnect()

p4.run_<cmd>()

Shorthand for:

p4.run( "<cmd>", arguments... )

p4.run_filelog( <fileSpec> ) -> list

Runs a p4 filelog on the fileSpec provided and returns an array of P4.DepotFile results (when executed in tagged mode), or an array of strings when executed in nontagged mode. By default, the raw output of p4 filelog is tagged; this method restructures the output into a more user-friendly (and object-oriented) form.

For example:

from P4 import P4, P4Exception
p4 = P4()

try:
  p4.connect()
  for r in p4.run_filelog( "index.html" )[0].revisions:
    for i in r.integrations:
      # Do something

except P4Exception:
  for e in p4.errors:
    print e

finally:
  p4.disconnect()

p4.run_login( <arg>…​ ) -> list

Runs p4 login using a password or ticket set by the user.

p4.run_password( oldpass, newpass ) -> list

A thin wrapper to make it easy to change your password. This method is (literally) equivalent to the following:

p4.input( [ oldpass, newpass, newpass ] )
p4.run( "password" )

For example:

from P4 import P4, P4Exception
p4 = P4()
p4.password = "myoldpass"

try:
  p4.connect()
  p4.run_password( "myoldpass", "mynewpass" )

except P4Exception:
  for e in p4.errors:
    print e

finally:
  p4.disconnect()

p4.run_resolve( [<resolver>], [arg…​] ) -> list

Run a p4 resolve command. Interactive resolves require the <resolver> parameter to be an object of a class derived from P4.Resolver. In these cases, the P4.Resolver.resolve() method is called to handle the resolve. For example:

p4.run_resolve ( resolver=MyResolver() )

To perform an automated merge that skips whenever conflicts are detected:

class MyResolver( P4.Resolver ):
  def resolve( self, mergeData ):
    if not mergeData.merge_hint == "e":
      return mergeData.merge_hint
    else:
      return "s" # skip the resolve, there is a conflict

In non-interactive resolves, no P4.Resolver object is required. For example:

p4.run_resolve ( "-at" )

p4.run_submit( [ hash ], [ arg…​ ] ) -> list

Submit a changelist to the server. To submit a changelist, set the fields of the changelist as required and supply any flags:

change = p4.fetch_change()
change._description = "Some description"
p4.run_submit( "-r", change )

You can also submit a changelist by supplying the arguments as you would on the command line:

p4.run_submit( "-d", "Some description", "somedir/..." )

p4.run_tickets( ) -> list

p4.run_tickets() returns an array of lists of the form (p4port, user, ticket) based on the contents of the local tickets file.

p4.save_<spectype>()>

The save_<spectype>() methods are shortcut methods that allow you to quickly update the definitions of clients, labels, branches, etc. They are equivalent to:

p4.input = dictOrString
p4.run( "<spectype>", "-i" )

For example:

from P4 import P4, P4Exception
p4 = P4()

try:
  p4.connect()
  client = p4.fetch_client()
  client[ "Owner" ] = p4.user
  p4.save_client( client )

except P4Exception:
  for e in p4.errors:
    print e

finally:
p4.disconnect()

p4.set_env( var, value )

On Windows or macOS, set a variable in the registry or user preferences. To unset a variable, pass an empty string as the second argument. On other platforms, an exception is raised.

p4.set_env = ( "P4CLIENT", "my_workspace" )
p4.set_env = ( "P4CLIENT", "" )

p4.temp_client( "<prefix>", "<template>" )

Creates a temporary client, using the prefix <prefix> and based upon a client template named <template>, then switches P4.client to the new client, and provides a temporary root directory. The prefix makes is easy to exclude the workspace from the spec depot.

This is intended to be used with a with statement within a block; after the block is finished, the temp client is automatically deleted and the temporary root is removed.

For example:

from P4 import P4
p4 = P4()
p4.connect()
with p4.temp_client( "temp", "my_template" ) as t:
  p4.run_sync()
  p4.run_edit( "foo" )
  p4.run_submit( "-dcomment" )

p4.while_tagged( boolean )

In the context of a with statement, enable or disable tagged behavior for the duration of a block. For example:

from P4 import P4
p4 = P4()
p4.connect()
with p4.while_tagged( False ):
  # tagged output disabled for this block
  print p4.run_info()

# tagged output back to normal
...