XBimP21Scanner. '1e-05' is not recognized as Tokens.Float.

[santiagoIT]

If you attempt to load the attached IFC file, nothing will be displayed. Other IFC Viewers display the content just fine (BimVision, etc..).

If you edit the following line:

#19=IFCGEOMETRICREPRESENTATIONCONTEXT('Building Model','Model',3,1e-05,#18,$);

to

#19=IFCGEOMETRICREPRESENTATIONCONTEXT('Building Model','Model',3,1.E-05,#18,$);

it then works.

The lexer/tokenizer is splitting 1e-05 into two tokens, when it should be only one token.

'1.E-05' is tokenized properly

XbimScannerIssue.zip

[santiagoIT]

The problem is with StepP21Lex.lex.

The regular expression expects a mandatory dot:

The first test string is not recognized. It should.

[andyward]

Looks like the exponent regex needs a bit of a rethink!

This seems to be a more robust implementation from StackOverflow

^(?:-?\d*)\.?\d?[eE][-\+]?\d+$

[andyward]

@martin1cerny Had a look at fixing this but it seems the GP Lexer needs a fix so we don't over-write the hand edited large file fix: commit 6517bc16042b3cfd820dd7eb45f72bbab92d13ad