feat: extract sex generator from name.gender to name.sex

[hankucz]

Fixes #353 Fixes #776

• extract sex generator from name.gender to name.sex
• deprecate name.gender(true), which generated binary gender, what we would consider now as sex
• rename locale definitions from bianry_gender to sex
• lowercase english values for sex
• extend english locale with non-binary

[codecov[bot]]

Codecov Report

Merging #1168 (ca21e20) into main (ca7cb41) will increase coverage by 0.00%. The diff coverage is 97.77%.

@@           Coverage Diff           @@
##             main    #1168   +/-   ##
=======================================
  Coverage   99.62%   99.62%           
=======================================
  Files        2158     2159    +1     
  Lines      240193   240217   +24     
  Branches     1003     1005    +2     
=======================================
+ Hits       239294   239325   +31     
+ Misses        878      871    -7     
  Partials       21       21           

Impacted Files	Coverage Δ
src/definitions/name.ts	0.00% <0.00%> (ø)
src/locales/fr/name/sex.ts	100.00% <ø> (ø)
src/locales/fr_CH/name/sex.ts	100.00% <ø> (ø)
src/locales/pt_BR/name/sex.ts	100.00% <ø> (ø)
src/locales/ur/name/sex.ts	100.00% <ø> (ø)
src/locales/de/name/index.ts	100.00% <100.00%> (ø)
src/locales/de/name/sex.ts	100.00% <100.00%> (ø)
src/locales/en/name/index.ts	100.00% <100.00%> (ø)
src/locales/en/name/sex.ts	100.00% <100.00%> (ø)
src/locales/fr/name/index.ts	100.00% <100.00%> (ø)
... and 7 more

[pkuczynski]

This is a breaking change as binary_gender could be used in faker.faker('{{binary_gender}}') without that we know it or users know that it had changed I mark this as blocked for now and would like to first discuss if this PR is needed anyway

If we are killing gender enum, it feels like this is a natural good step forward to make the differentiation between gender and sex.

In my opinion faker.fake should be deprecated, as the same result can be achieved with string templates, which additionally allow user to provide options to method calls, instead of just using locales.

So yes, technically it could be a breaking change, but at the same time locale formats are not officially published, so I have doubts if anybody would hook up to the internals so deep to find it. And if he did, it's kind of not our fault...

@ST-DDT @xDivisionByZerox wdyt?

[xDivisionByZerox]

In my opinion faker.fake should be deprecated, as the same result can be achieved with string templates, which additionally allow user to provide options to method calls, instead of just using locales.

TBH, I'm with you on this one. Never understood the use of fake at all. Gues its legacy from pre-v5.5.

So yes, technically it could be a breaking change, but at the same time locale formats are not officially published, so I have doubts if anybody would hook up to the internals so deep to find it. And if he did, it's kind of not our fault...

Not "technically". In this case, it is since the feature to use locale definition with fake is documented AND promoted by using it in our own codebase. Have a look at these deprecation statements:

https://github.com/faker-js/faker/blob/ca7cde59ec3747a93d1d47355a83287696628eb5/src/modules/address/index.ts#L103-L109

https://github.com/faker-js/faker/blob/ca7cde59ec3747a93d1d47355a83287696628eb5/src/modules/address/index.ts#L126-L132

So for me, there is no question whether renaming binary_gender is a breaking change. It clearly is.

[pkuczynski]

TBH, I'm with you on this one. Never understood the use of fake at all. Gues its legacy from pre-v5.5.

See my (same) comment https://github.com/faker-js/faker/pull/1161#issuecomment-1195852602

Not "technically". In this case, it is since the feature to use locale definition with fake is documented AND promoted by using it in our own codebase. Have a look at these deprecation statements:

https://github.com/faker-js/faker/blob/ca7cde59ec3747a93d1d47355a83287696628eb5/src/modules/address/index.ts#L103-L109

https://github.com/faker-js/faker/blob/ca7cde59ec3747a93d1d47355a83287696628eb5/src/modules/address/index.ts#L126-L132

So for me, there is no question whether renaming binary_gender is a breaking change. It clearly is.

These are two statements that do not mention binary_gender, so it's unlikely someone would be using it. Especially this data is provided in 5 locales only, so very few... So yes, this is technically breaking, but the impact is low...

[hankucz]

What are the next steps? Should I do anything about it?

[Shinigami92]

Should we duplicate the binary_gender locale definition for now and remove it in v8? That way we could merge this PR now and don't need to wait because of the breaking change for the fake function 🤔

[pkuczynski]

Should we duplicate the binary_gender locale definition for now and remove it in v8? That way we could merge this PR now and don't need to wait because of the breaking change for the fake function 🤔

We could if you think it's worth it. Hard to say if anybody was using it in fake, but I kind of doubt it... We should deprecate faker.helpers.fake as it does not allow us to do any kind of typings :(

[ST-DDT]

Docs-Preview (Click to expand)

[hankucz]

I believe I applied all of the feedback. Is there anything else I should do about it?

[Shinigami92]

Ping @ST-DDT, you need to approve to stale your requested changes