[EPIC] Improve feedback to user on XML errors

[MelReyCG]

Typical XML user syntax, path and naming errors  (additional commas or spaces, misplaced slash, missing fields) produce messages in the standard output which consistency and readibility by a user can be improved.

error improvement exemples.pdf general presentation.pdf

---

• [x] General need: To indicate from where an error comes from, we could output it to the user in the message GEOS PRs:
• #2404
• #2447
• #2448
• #2449
• #2450

---

• [x] Case 1 - When adding a comma at start, end or between the entries of a {materialList, cellBlocks, cellBlockNames, phaseName, targetRegions}.
• Typical current output: Controlling expression (should be false): inputStream.peek() != '}'
• Suggested output: "wellPressureCollection": Empty entry in line "{ water, ,rock }". Please remove a comma or add an entry.
• GEOS PR #2357
• LvArray PR #280

---

• [x] Case 2.1 - Enter a wrong objectPath in a PackCollection.
• Typical current output: std::bad_alloc / Segmentation fault
• Suggested output:

wellPressureCollection: Specified name (wellRegion1) did not find a match with a object in group (ElementRegions). 
Objects that are present in (ElementRegions) are:
reservoir, wellRegion

• GEOS PR #2341

---

• [x] Case 2.2 - Enter a wrong objectPath in general
• Typical current output: Group m/ElementRegion does not have a child wellRegion
• Suggested output: Error in initialPressure objectPath : Group ElementRegions doesn't have a child named wellRegion123. ElementRegions have the following children: wellRegion1,wellRegion2,meshRegion1
• GEOS PR #2341

---

• [x] Case 3 - Enter a wrong fieldName
• Typical current output: Group m/domain/MeshBodies/cartesianMesh/meshLevels/Level0/ElementRegions/elementRegionsGroup/reservoir/elementSubRegions/cellBlock doesn't have a child pressures
• Suggested output:

"initialPressure": couldn'd find a field "pressures" in "ElementRegions/reservoir1"
Could be one of : pressure, globalCompFraction
Error happened while checking the children of 
m/domain/MeshBodies/cartesianMesh/meshLevels/Level0/ElementRegions/elementRegionsGroup/reservoir/elementSubRegions/cellBlock.

• GEOS PR #2449

---

• [ ] Case 4 - Enter an empty CompositionalMultiphaseFVM::targetRegions, materialList, cellBlockNames
• Typical current output: Group m/domain/MeshBodies/cartesianMesh/meshLevels/Level0/ElementRegions/elementRegionsGroup doesn't have a child -1
• Suggested output: "reservoir" must have cellBlocks.

---

• [x] Case 5.1 - Enter a forbiden characters (like , or or /) in a name field
• Typical current output:  Group m/Solvers/compositionalMultiphaseWell doesn't have a child wellControls (the error happens when we reference the object which name ends with a space)
• Suggested output: "cartesianMesh " name field cannot contain forbidden characters.
• GEOS PR #2764

---

• [x] Case 5.2 - Enter a forbiden characters (like , or or /) in a name field reference
• Typical current output: Error detected while parsing string: "wellControls "
• Suggested output: "wellInjector1" wellControlsName field must not contain forbidden characters.
• GEOS PR #2764

---

• [x] Case 5.3 - Enter an objectPath with forbiden special characters in it (like , or or /) 
• Typical current outputs:  -> Group m/domain/MeshBodies/cartesianMesh/meshLevels/Level0/ElementRegions/elementRegionsGroup/wellRegion/elementSubRegions doesn't have a child wellRegionUniqueSubRegion_ -> Error detected while parsing string: "ElementRegions/wellRegion/wellRegionUniqueSubRegion "
• Suggested output: Group "wellPressureCollection" objectPath "wellRegionUniqueSubRegion_" cannot contain forbidden characters.
• GEOS PR #2764

---

Side note

These suggested messages could change, they must be followed by the stacktrace, and their goal is to contain all needed information for a developper to debug and for a user to understand his mistake in his xml.

[TotoGaz]

Can you put the actual error message close to the desired error message? It's complicated to follow...

[TotoGaz]

Hurray @MelReyCG , this is done. I let you the pleasure to tick the boxes!

EDIT: what about case 4?

[MelReyCG]

I re-run the tests I made for this EPIC, and this point is not solved yet (the error is still the same). I'll submit a last PR to solve it, we need to be able to specify if an X_array is allowed to be empty.