=== TaskJuggler 2.x Migration ===

This section will cover changes between TaskJuggler 2.x and 3.x.

* The syntax for macros has changed slightly. The terminating
''''<nowiki>]</nowiki>'''' must be the last character before the
line break. No spaces or comments are allowed here. Parameters of
macro calls must always be enclosed by double quotes. In contrast to
regular strings, single quotes are not allowed here. The parameter
may not span multiple lines.

* The ''''projection'''' attribute has been removed. The
is now provided by [[trackingscenario]].

* The default working hours have been changed to 9:00 - 17:00.

* IDs for properties such as tasks, resources and reports are now
optional. If you don't need to reference a property, you can omit the
ID. TaskJuggler will automatically assign an ID then.

* Top-level accounts no longer need a ''''cost'''' or ''''revenue''''
attribute. Any two top level accounts can now be balanced against
each other using the [[balance]] attribute in the report.

* The ''''shift'''' attribute for tasks and resources has been
renamed to ''''shifts'''' to allow support for multiple shifts.

* The global ''''limits'''' attribute has been removed. Since both
tasks and resources have a ''''limits'''' attribute, a global
attribute was inconsistent as only resources inherited this
attribute. Use a parent resource to emulate the old behaviour.

* Shifts and limits for allocations have been deprecated. The concept
was bogus and not compatible with bookings. The functionality is
now provided by [[shifts.task|shifts]] and [[limits.task|limits]] on the
task level. Limits for a task can be [[resources.limit|selectively
applied]] to certain resources.

* The ''''startbuffer'''' and ''''endbuffer'''' attributes have been
deprecated. They have not been widely used and had no impact on
scheduling.

* The project attribute ''''allowredifinitions'''' has been dropped.
It was an ugly workaround for a rare corner case. Using [[supplement]] is
the clean way to do this.

* Camel case names for function names in logical expressions have
been deprecated. Function names need to be all lower case now. Some
functions have been removed as all attributes can now be accessed by
scenario.attribute_id notation.

* The format for report has been changed considerably.
The old format was not very flexible and had some design flaws.
TaskJuggler 3.x now supports report nesting and composition. A report
definition can be used to generated multiple output [[formats]]. The
name of a report must now be specified without the file name
extension. It will be automatically added depending on the output
format.

* The sorting modes have been extended to include the scenario. Also,
the sorting direction is no longer mangled with the attribute name.
What used to be ''''startup'''' is now ''''plan.start.up''''. See
[[sorttasks]] or [[sortresources]] for details.

* The attribute ''''properties'''' for ''''export'''' reports is no
longer supported. The naming was inconsistent with TaskJuggler lingo
and did not handle tasks and resources separately. It has been
replaced with [[taskattributes]] and [[resourceattributes]].

* The ''''barlabels'''' attribute for reports is no longer needed.
HTML reports have always empty Gantt-chart bars and the calendar
reports always have values.

* Support for reading and writing XML files is no longer available.
The content was redundant with the TJP file format and it was not
widely used. Keeping it in sync was too much of an effort to be worth
it. There is nothing in the TJ3 design that would prevent this
feature from being added again, but there are no plans for this right
now.

=== Using TaskJuggler 2.x and TaskJuggler 3.x in parallel ===

While TaskJuggler 3.x has many new features over TaskJuggler 2.x like
the much improved HTML reports, many 2.x users will miss the
graphical user interface.

To ease the migration, you can continue to use the TaskJuggler 2.x
front-end while using TaskJuggler 3.x for report generation. This is
possible because TaskJuggler 3.x can read-in the TaskJuggler 2.x
export files. Export files are fully scheduled projects that include
start and end dates for all tasks and bookings for resource
allocations. 

To export all tasks and resources into a TJP file that can be read by
TaskJuggler 3.x include the following export report definition in your
TaskJuggler 2.x project plan. The necessary patches to support this
only made it into TaskJuggler 2.x after the 2.4.3 release. So be sure
to use a recent version from the Git repository to try this.

 export "FullProject.tjp" {
   taskattributes all
   resourceattributes all
   hideresource 0
 }

The resulting ''''FullProject.tjp'''' file is a valid self-contained
project file that can be read with TaskJuggler 2.x or TaskJuggler
3.x. The file does not contain any report definitions. To generate
reports with TaskJuggler 3.x you need to create an additional file
that contains the TaskJugler 3.x report definitions.

Let's assume the file is called ''''tj3reports.tji''''. Start
TaskJuggler 3.x with the following command:

 tj3 FullProject.tjp tj3reports.tji

Now you have generated TaskJuggler 3.x reports from you TaskJuggler
2.x project.

