The whitelist does not work when words have mixed case

[xsmq]

1. Unused whitelist

$ codespell -q 7 mindspore/mindspore/lite/tools/converter/parser/tf
mindspore/mindspore/lite/tools/converter/parser/tf/tf_merge_parser.cc:39: MergeT ==> merge
mindspore/mindspore/lite/tools/converter/parser/tf/tf_activation_parser.cc:66: shoud ==> should

2. Add whitelist

(1) MergeT

$ cat codespell.txt
MergeT
shoud

$ codespell -q 7 -I codespell.txt mindspore/mindspore/lite/tools/converter/parser/tf
mindspore/mindspore/lite/tools/converter/parser/tf/tf_merge_parser.cc:39: MergeT ==> merge

(2) nNumber

$ cat codespell.allow
MergeT
nNumber

$ codespell -q 7 -I codespell.allow mindspore/mindspore/ccsrc/minddata/dataset/engine/connector.h
mindspore/mindspore/ccsrc/minddata/dataset/engine/connector.h:151: nNumber ==> number
mindspore/mindspore/ccsrc/minddata/dataset/engine/connector.h:152: nNumber ==> number

(3) REALEASE

$ cat codespell.allow
pyhton
REALEASE

$ codespell -q 7 -I codespell.allow mindspore/mindspore/lite/examples/train_lenet/README.md
mindspore/mindspore/lite/examples/train_lenet/README.md:67: REALEASE ==> RELEASE
mindspore/mindspore/lite/examples/train_lenet/README.md:68: betweeen ==> between
mindspore/mindspore/lite/examples/train_lenet/README.md:72: followings ==> following
mindspore/mindspore/lite/examples/train_lenet/README.md:78: paramaters ==> parameters

[r3econ]

I'm experiencing the same problem. Makes using this tool in production impossible

[peternewman]

Hi @xsmq ,

As mentioned to @r3econ in codespell-project/actions-codespell#29 , from the main codespell help ( https://github.com/codespell-project/codespell#readme ):

Important note: The list passed to -I is case-sensitive based on how it is listed in the codespell dictionaries.

2. Add whitelist

(1) MergeT

grep -iIR merget codespell_lib/data/
codespell_lib/data/dictionary.txt:merget->merge

(2) nNumber

$ cat codespell.allow
MergeT
nNumber

grep -iIR nnumber codespell_lib/data/
codespell_lib/data/dictionary.txt:nnumber->number

(3) REALEASE

$ cat codespell.allow
pyhton
REALEASE

grep -iIR ^realease codespell_lib/data/
codespell_lib/data/dictionary.txt:realease->release
codespell_lib/data/dictionary.txt:realeased->released
codespell_lib/data/dictionary.txt:realeases->releases

So you want these in your codespell.allow:

merget
nnumber
realease

[xsmq]

OK，thanks.

[peterjc]

I just struggled with this with an all capital term, having initially expected this to mean case sensitive to match the input file. That does seem a more intuitive behaviour - although a change to the tool.

[peternewman]

I just struggled with this with an all capital term, having initially expected this to mean case sensitive to match the input file.

@peterjc we'd welcome suggestions for how to make the existing help file more clear/less confusing:

Important note: The list passed to -I is case-sensitive based on how it is listed in the codespell dictionaries.

That does seem a more intuitive behaviour - although a change to the tool.

If it had to match the input file, then if I had this input:

SPELING TOOLS
There are lots of tools to catch spelings available

Then I'd need to add two ignore entries for the two cases that appear in the input.

I'm not entirely sure of the history of why we don't just do a case-insensitive comparison, but it would prevent some classes of typos being entered and therefore detected (e.g. names being in lower case).