| Ticket UUID: | 21e3c44ac94dc5cbe2946fe66f7357588c16efe9 | ||
| Title: | Table encoding limitations? | ||
| Type: | Bug | Created on: | 2020-04-23 16:39:33 |
| Submitter: | dgp | Assigned to: | nobody |
| Subsystem: | 10. Objects | Severity: | Minor |
| Priority: | 5 Medium | Last modified: | 2020-04-26 18:11:37 |
| Status: | Open | Closed by: | nobody |
| Resolution: | None | Closed on: | |
| Version: | |||
| Description: | ||||
The code in tclEncoding.c that implements table encodings appears to be limited to be able to work with only code points in UCS-2. This kind of encoding includes most of them provided as *.enc files. It's possible this is fine, but I wanted to enter a ticket mentioning it so it could be examined. | ||||
| User Comments: | ||||
dgp added on 2020-04-26 18:11:37:
I think the implication is that if any third party wants to create an encoding that supports the entire Unicode character set, they cannot accomplish that by the creation of any *.enc file. They must use the C routine Tcl_CreateEncoding(), which means they must create a binary extension for [load]ing. If this is correct, it seems to be a constraint worthy of noting in the docs. | ||||
