"Stress testing" units' string parsing about units

[phlptp]

I hope to make a new release with these changes within this coming week.

[phlptp]

Deal with these one by one

Metres

I think I missed a capital letter in the check for this spelling, will be fixed

cars/hour/meter^2

interpreted as centi- ares =m^2
a little odd but I would consider it correct

degrees C

0.0174532925199432955s*A*rad | to be fixed right now reading as degrees (angle) * coulomb

cm+cm

+ is usually non-sensical in terms of units so interpreted as * in non numeric context. Could consider alternate interpretations if the units have the same base then it could be additive

kM

kmol/L correct, but may be surprising (M= molarity = mol/L, so this seems correct) correct,

Others

Still exploring these
| gramm | 6.47989100000000217e-23gm | this would be a typo, or German spelling
| kilogramm | 6.47989100000000188e-20gm | ?

| fps | ft/s | not sure if this is the most common interpretation? Isn't frames per second more common?

| inchy | 4830819306288062m^3/[c] | would expect an error instead of bad parsing
| fln_oz | 9.7132811577355196e-08g*cd/m^2 | would expect an error instead of bad parsing

[phlptp]

inchy my first thought was that this should read inch*year. Whether even that would be good is probably questionable.
What seems to be happening is that it interprets it as inch*chain*year which oddly dividing by the speed of light ends up producing the most compact unit string since chain isn't one of the standard output strings.

Probably some tweaking on the partitioning that should be done

[phlptp]

fln_oz = foot-lambert newton ounce

I think I am going to disable the partitioning when there is an underscore, and be a little less permissive. That will come in the next PR.

[phlptp]

After the recent updates several of the issues have been resolved

Input	old Output	new Output	Comments
gramm	6.47989100000000217e-23g*m	g	if the non-english units are disabled would translate to gram * m
kilogramm	6.47989100000000188e-20g*m	kg	same as above
cars/hour/meter^2	1/h	1/h	centi-ares/hr/meter^2 odd but correct
cm+cm	cm^2	2cm	the addition symbol for units that are convertible to eachother now operates
fps	ft/s	ft/s	leaving as is, there is a future update coming which will better handle commodities but for now this is correct
Metres	Unsupported unit.	m	fixed the missing variant
degrees C	0.0174532925199432955s*A*rad	degC	this was an overlooked variant
inchy	4830819306288062m^3/[c]	not valid	the parsing was getting overly aggressive on unit partitioning
fln_oz	9.7132811577355196e-08g*cd/m^2	not valid	partitioning now takes the '_' in account better
kM	kmol/L	kmol/L	kiloMolarity -this is correct

[phlptp]

This was pretty interesting to work through.
There is now a defaultflag property that can be set and gets used if no flags are entered in the api call. A minimum partition size can be specified as part of those flags. This allows better control of the conversion process for programs using the conversion.

[SimonHeybrock]

Thanks! Do you plan to make a release soon?

[phlptp]

new release made