Clarification on the association of a language mode with an mvs data set

[FALLAI-Denis]

Hi,

Many features of VS Code in general, and of Z Open Editor are based on the association of a language mode with a file.
Basically, the language mode is determined by the extension carried by the file.

When a mvs data set is transferred to the workstation, an extension must be associated with it, which will then make it possible to determine a language mode, and to activate the corresponding functionalities in VS Code and Z Open Editor.

I would like the mechanism for associating a language mode to a mvs data set to be more detailed than what is now in the documentation, and for each step of this mechanism to be described and the order of operations to be specified.

There is the VS Code mechanism based on the files.associations declarations in the settings.
A priori this makes it possible to associate a file name pattern (and not a file extension) with a language mode.

 // Configure file associations to languages (e.g. `"*.extension": "html"`). These have precedence over the default associations of the languages installed.
 "files.associations": {},

There are Z Open Editor settings of type zopeneditor.datasets.<language-id>Datasets.
These parameters do not appear to be directly involved in the process of determining the language mode of a file. I understood that they were mainly used to complement the files.associations settings which are used for the determination of the language mode.
The values provided in zopeneditor.datasets.<language-id>Datasets are transformed into a pattern in files.associations.
For example :
"zopeneditor.datasets.cobolDatasets": [".COBOL"]
will be transformed into:
"files.associations": {"*.COBOL*": "cobol"}

Because files.associations is a pattern, especially with a "*" selector at the start and end of the pattern, it can capture any filename having this pattern, and it is problematic in handling priority in determining the language mode of a mvs data set .

On these mvs data sets (real case):
<hlq>.COBOL.SOURCE: contains sources (and not necessarily only COBOL)
<hlq>.COBOL.JCL: contains JCLs for compiling and executing programs
<hlq>.COBOL.REXX: contains REXX tools for analyzing and processing COBOL sources
and others...
and the settings:
"zopeneditor.datasets.cobolDatasets": [".COBOL"]
"zopeneditor.datasets.jclDatasets": [".JCL"]
"zopeneditor.datasets.rexxDatasets": [".REXX"] (currently does not exist, given for illustration)
what will generate the settings:
"files.associations": {"*.COBOL*": "cobol", "*.JCL*": "jcl", "*.REXX*": "rexx"}
What will be the language mode associated with each of the mvs data sets?
In fact, I see that they all inherit the "cobol" language mode, which is not appropriate for JCL or REXX PDS

I think the pattern generated by the settings zopeneditor.datasets.<language-id>Datasets should instead be of the form:
"zopeneditor.datasets.cobolDatasets": [".COBOL"]
"zopeneditor.datasets.jclDatasets": [".JCL"]
"zopeneditor.datasets.rexxDatasets": [".REXX"] (currently does not exist, given for illustration)
should generate something like:
"files.associations": {"*.COBOL": "cobol", "*.COBOL(*": "cobol", "*.JCL": "jcl", "*.JCL(*": "jcl", "*.REXX": "rexx", "*.REXX(*": "rexx"}
thus indicating that it can be a sequential data set (for low level qualifier given) or members of PDS (there will remain an ambience with the VSAM GDG).

In the process of determining the language mode of an mvs data set, is it Z Open Editor that is involved, Zowe Explorer, or VS Code itself?