Species names with a colon (:) can cause problems in parseCompString

[rwest]

Cantera version

2.3.0 (via Anaconda)

Operating System

MacOS X
Darwin-16.7.0-x86_64-i386-64bit

Python version

2.7.11 |Anaconda custom (x86_64)| (default, Dec 6 2015, 18:57:58)
[GCC 4.2.1 (Apple Inc. build 5577)]

Preamble

In [1]: import cantera as ct

In [2]: ct.__version__
Out[2]: '2.3.0'

Expected Behavior

In [3]: ct.Solution('cantera-working.cti')
< some warnings and then...>
Out[3]: <cantera.composite.Solution at 0x109ffbb90>

Actual Behavior

In [4]: ct.Solution('cantera-broken.cti')
< some warnings and then...>
---------------------------------------------------------------------------
CanteraError                              Traceback (most recent call last)
<ipython-input-12-556309c8f232> in <module>()
----> 1 ct.Solution('cantera-broken.cti')

interfaces/cython/cantera/base.pyx in cantera._cantera._SolutionBase.__cinit__ (interfaces/cython/cantera/_cantera.cpp:7873)()

interfaces/cython/cantera/base.pyx in cantera._cantera._SolutionBase._init_cti_xml (interfaces/cython/cantera/_cantera.cpp:8556)()

CanteraError: 
***********************************************************************
CanteraError thrown by parseCompString:
Duplicate key: 'CF2'.
***********************************************************************

The difference between the two .cti files is that the -broken.cti one contains some species with : in their names, such as CH2:CHF and CH2:CF2, generated using ck2cti on the attached chemkin files, and the working one had all the : replaced with # in the chemkin files (also attached) before they were converted to the -working.cti file.

I think the offending line is
https://github.com/Cantera/cantera/blob/master/src/base/stringUtils.cpp#L85
but I wasn't able to figure out on what occasion it was being called to cause this problem.

My guess, from just staring at the code, is that the checks and work-arounds for colons occurring around line 91 of that file (stringUtils.cpp) should happen before the duplicate key check on line 83, which is currently using the uncorrected name that is only what appears before the first colon.
But I can't easily test my theory, so am just reporting it here.

The chemkin and cantera input files are in this Archive.zip

[speth]

Thanks, Richard. Your diagnosis was exactly correct.