From: Brian G. <bri...@ea...> - 2017-06-23 14:24:04
|
On Jun 23, 2017, at 1:50 AM, Jan Nijtmans <jan...@gm...<mailto:jan...@gm...>> wrote: 2017-06-22 20:26 GMT+02:00 Donal K. Fellows: And now for the complicated one! (In fact, this is probably the most complex vote result we've seen.) • TIP #472: Add Support for 0d Radix Prefix to Integer Literals For: AF, BG, DKF (in 8.7) Present: JN with %d, SL with %d, KBK with %d Against: DKF (in 8.6), DGP (in 8.6) I think there's a reasonable consensus for putting this into 8.7, but not into 8.6 — i.e., there's not sufficient votes to go into 8.6 with at least some configuratinos — and if it is going to 8.7 then it can have the changes to %#d be done as part of it (as we'll have time to warn people and debug any weird issues that come up). I've heard that Jan's suggestions for improving the handling of %#d (and related) magic in general are a good thing. Agreed, I'll formulate my suggestion in a new TIP (taking Kevin's %od in consideration as well), there's still time for that. Current implementation merged to trunk now (open for widespread testing!). Another thing to be done is in test-cases. I would expect that the test outcome for format is independent on longIs32bit/longIs64bit, but the test outcome shows something different: <http://core.tcl.tk/tcl/artifact?ln=81-98&name=96269eab4cf851cf> The longIs32bit output looks correct, but the lingIs64bit output is truncated. I guess there's a bug hidden somewhere. Brian? I'll take a look. -Brian |