#1995 _tcsclen missing from tchar.h

[Earnie Boyd]

Per recent mingw-users conversation "Trouble with specs".

See MSDN reference for details.

[Keith Marshall]

Hah! You beat me to it; I was just about to open a (more comprehensive) ticket.

The issue goes (much) further than just _tcsclen missing; there is no handling whatsoever, (as there should be), specified for the case where _MBCS is defined. This additional MSDN reference identifies what should be defined for each of _UNICODE defined, _MBCS defined, and neither defined. (It is confusingly ambiguous for the anomalous case where both _UNICODE and _MBCS might be defined concurrently; I've challenged them to clarify it, but I won't hold my breath). This legacy reference, and intervening editions may also be of interest.

I guess a complete audit of tchar.h is called for; at the very least, _MBCS handling should be specified, in addition to furnishing missing function defines such as _tcsclen.

[Earnie Boyd]

I guess a complete audit of tchar.h is called for; at the very least, _MBCS handling should be specified, in addition to furnishing missing function defines such as _tcsclen.

[#1996]

[Keith Marshall]

As a starting point, I've performed a (significant) format clean-up on tchar.h

[Earnie Boyd]

Can you point out the non white space changes? Attaching a diff ignoring the white space changes would be good.

[Keith Marshall]

Sure; output from git diff -w include/tchar.h is attached.

Technically, there are no significant non-whitespace changes, (unless you consider the coalescence of two consecutive and logically related comments into a single comment block, and the extension of one other comment block by one line. Other than those two, I removed a few blank lines which felt were superfluous, added a few more, where I felt they were needed, and predominantly, tidied up the mish-mash of inconsistently used hard tab characters.

[Earnie Boyd]

Okay, I understand you now. I'll wait until I draft a WSL coding spec for these changes. This way we can remain consistent. This review will take place during the 5.0 development phase.

[Earnie Boyd]

I've added the define for _tcsclen to tchar.h.

[Keith Marshall]

I see you ducked the issue of the third required mapping for tcsclen() -- the _MBCS defined case -- so strictly this can't be classed as completely "fixed"; I guess you plan to pick this up again, under [#1996].

[Earnie Boyd]

Yes, that is the plan. This isn't the only _MCSB (I mean _MBSC (really I mean _MBCS)) issue.