Stability of CreateSampleShape GUI

[Nick Draper]

The gui can cause Mantid to crash when insufficient information is present to display the sample.

e.g.

1. Create a sphere (enter no details)
2. Create a union above
3. Add a cylinder to the union (no details)
4. Click update 3d view -> freeze

[Nick Draper]

Moved to iteration 33 at iteration 32 code freeze

[Nick Draper]

Moved at end of release 2.1

[Nick Draper]

Moved at the end of release 2.2

[Nick Draper]

Moved to milestone 2.4

[Nick Draper]

Moved at the code freeze for release 2.4

[Nick Draper]

Moved to r2.6 at the end of r2.5

[Martyn Gigg]

This should also be able to load a previously created shape

[Keith Brown]

There seems to be two parts to this:

• Fix the crash
• Make the custom GUI operate more sensibly, populating itself if possible.

[Keith Brown]

Fixed crash by means of extra validation

The program should no longer freeze as it is no longer possible to update the 3D view or create the shape without entering in all the details for a component shape. If any fields don't contain a number that's convertable to double, the program will show an error message. The error doesn't show what's wrong specifically, it just tells the user there is something to check.

Refs #3737

Changeset: 549e777f0ada41034a8ca9ffb92e677b4a07d40b

[Keith Brown]

Stashing in order to fix another ticket

Validation has had a fix as in the shape valid() methods I had negated some conditions when I shouldn't have.

Otherwise, work is starting on loading a stored shape

Refs #3737

Changeset: f1fd374d9033a3bee7855be5fda62bbd91815e03

[Keith Brown]

Started writing the shape XML parser

The XML parser has been started and currently only gets the algebra string so that the tree can be constructed.

Next job is to actually build a data structure to hold the tree info in

Refs #3737

Changeset: d1f6d5840d0cc4f3b48909ec8241dd292b44ad90

[Keith Brown]

Started writing equation parser

Committing to save progress before looking at another branch the equasion parser will be written as a recursive function, finding the most central opperator and splitting the string there and passing to a lower function until only a shape remains

Refs #3737

Changeset: 4dfa9c4e4dce5ad636de17c71584fe12d94e4446

[Keith Brown]

Equation parser taking shape

the recursive equation parsing algorithm is taking shape and the framework is nealry complete

Refs #3737

Changeset: ac1b131307d98e92fb822e25ffbf9d1c9e1b7d84

[Keith Brown]

Parsing fucntion practically complete.

The parsing function is as complete as it can be for now, as it can't create anything before i supply the storage framework for storing the dummy tree The new constructors for the shapes need written too

Refs #3737

Changeset: 1bdfcc733770a20635e15182bdec53ed5758a96a

[Keith Brown]

Discovered another bug where if a shape has it's complement flag set (the text in the widget has a # in it) it won't be able to be edited as setupDetailsBox() looks directly to see if item->text(0) is in m_setup_map, which it won't be as the # is in the way. so it continues without adding the editing widget, leaving a blank space.

Will have to change either the map or strip out the # before looking it up.

[Keith Brown]

Stashing to switch tickets

Now creating the functionality to populate the tree. Committing in order to switch branches and fix a reopened ticket

Refs #3737

Changeset: fc16513906be06c008663a3c8965a0405f9c7170

[Keith Brown]

Instanciation methods taking shape

The instanciation methods that build the QTreeWidget are taking shape.

I've added new constructors to ShapeDetails and derived classes that take a Poco::XML::Element and can fill in details from the XML contained within that element

Refs #3737

Changeset: 61cb9aba7a44b62a1c72fe5fe532b03db3db11de

[Keith Brown]

Tree is being built, XML being parsed

The tree is being built as expected and the XML is being parsed, but i've revealed that the new way i've done something seems to mean that widgets aren't being loaded, but i'll fix that by making it get the widget a more sensible way.

Refs #3737

Changeset: 8301b67af959e1169d5da84299040675fe8b5839

[Keith Brown]

Some shapes confirmed to work. Complement identification works

At least spheres and finite cylinders import properly, the code is there for sliceofcylinderring and infinitecylinder but they've not been tested. The other shapes are awaiting parseing code.

I had made some mistakes when it came to complements of shapes. The first was that even if they were identified they weren't getting set as them, this was because I'd accidentally redeclared the flag as local and it was going out of scope. the second was that the displayed text was wrong due to the way i was constructing the QString.

Fixed a possible memory leak (or at least objects were kept in memory and not deleted) as when i was parsing XML in the shapes, some lists weren't being released.

Refs #3737

Changeset: 95686b7279f15e92e744e44537277f4cd9f79ecd

[Keith Brown]

All shapes now Parse XML given to them

All shapes now have a construtor that will intialise them from XML.

I still need to inlcude parts for the spherical coordinate system but once that's in all i need to do is test it out and make sure it works fully

Refs #3737

Changeset: bc1149c8e330b7f9d04e7fed8f68086d545761ca

[Keith Brown]

This is blocked as some shapes don't render properly

[Keith Brown]

not going to get this done before i leave, leaving to anyone

There is work against this ticket on branch bugfix/3737_CreateSampleShape_freeze_and_load_fix

[Stuart Campbell]

This ticket has been transferred to github issue ​4584