Fails on comment after closing brace in Table spec

[marktaff]

The error:
pyparsing.exceptions.ParseException: Expected {LineEnd | StringEnd}, found '/' (at char 1112), (line:31, col:3)

Expected behavior:
Personally, I'd like if the comments were put in the <Database> object attached to the table or column as appropriate, as I need to extend dbml for mariadb features they don't support. But at a minimum, it seems like having the parser ignoring comments totally is better than failing.

This dbml snippet will reproduce:

Table nameSuffixOptions {
    id int(11) [primary key, increment]
    suffix varchar(7) [not null]
    sortOrder int(11) [not null]
    createdAt timestamp [default: `CURRENT_TIMESTAMP`]
    updatedAt timestamp [default: `CURRENT_TIMESTAMP`] // maria: [ onupdate: `CURRENT_TIMESTAMP`]
    note: 'Options for name suffixes: Sr., Jr., etc'
} // maria: [engine: innodb, charset: utf8]

[mjfii]

The comments are evaluated prior to the object being defined. So this should evaluate:

// maria: [engine: innodb, charset: utf8]
Table nameSuffixOptions {
    id int(11) [primary key, increment]
    suffix varchar(7) [not null]
    sortOrder int(11) [not null]
    createdAt timestamp [default: `CURRENT_TIMESTAMP`]
    // maria: [ onupdate: `CURRENT_TIMESTAMP`]
    updatedAt timestamp [default: `CURRENT_TIMESTAMP`]
    note: 'Options for name suffixes: Sr., Jr., etc'
}

Additionally, you can use the Project DBML object to store metadata:

Project newProject {
  name: 'New Project'
  dbms: 'maria'
  engine: 'innodb'
  charset: 'utf8'
  author: 'markaff'
  Note: '''
  a customized model
  '''
}

Table nameSuffixOptions {
    id int(11) [primary key, increment]
    suffix varchar(7) [not null]
    sortOrder int(11) [not null]
    createdAt timestamp [default: `CURRENT_TIMESTAMP`]
    // maria: [ onupdate: `CURRENT_TIMESTAMP`]
    updatedAt timestamp [default: `CURRENT_TIMESTAMP`]
    note: 'Options for name suffixes: Sr., Jr., etc'
}

You can access it via the project.items dict.

[Vanderhoof]

Sorry for a very late reply: I was afk most of the previous weeks due to moving.

I've fixed the error in 1.0.8, thanks a lot for reporting! As @mjfii pointed out, currently PyDBML attaches the comments, which are put directly before the Table, into the parsed Table object. Same behavior is true for any other entity: a comment before the entity gets into the parsed object. To illustrate it on @mjfii 's example:

>>> from pydbml import PyDBML
>>> source = '''
... // maria: [engine: innodb, charset: utf8]
... Table nameSuffixOptions {
...     id int(11) [primary key, increment]
...     suffix varchar(7) [not null]
...     sortOrder int(11) [not null]
...     createdAt timestamp [default: `CURRENT_TIMESTAMP`]
...     // maria: [ onupdate: `CURRENT_TIMESTAMP`]
...     updatedAt timestamp [default: `CURRENT_TIMESTAMP`]
...     note: 'Options for name suffixes: Sr., Jr., etc'
... }
... '''
>>> parsed = PyDBML(source)
>>> parsed.tables[0].comment  # the table object has the comment attribute
'maria: [engine: innodb, charset: utf8]'
>>> parsed.tables[0]['updatedAt']
<Column 'updatedAt', 'timestamp'>
>>> parsed.tables[0]['updatedAt'].comment  # the column `updatedAt` also has the comment, defined right before it
'maria: [ onupdate: `CURRENT_TIMESTAMP`]'