Ticket UUID: | 483500 | |||
Title: | Changes for TIP #73 | |||
Type: | Patch | Version: | None | |
Submitter: | kennykb | Created on: | 2001-11-19 17:28:35 | |
Subsystem: | 06. Time Measurement | Assigned To: | kennykb | |
Priority: | 5 Medium | Severity: | ||
Status: | Closed | Last Modified: | 2001-11-21 09:29:17 | |
Resolution: | Fixed | Closed By: | hobbs | |
Closed on: | 2001-11-21 02:29:17 | |||
Description: |
This patch implements the change described in TIP #73. It replaces all uses of TclpGetTime with Tcl_GetTime, and adds the latter to the Stubs table and the library documentation. It also corrects bug #478847 (the change overlaps with the other changes). | |||
User Comments: |
hobbs added on 2001-11-21 09:29:17:
Logged In: YES user_id=72656 commited to 8.4a4cvs. Successfully tested on AIX4.3, HP11, Solaris, Linux and Win2K. kennykb added on 2001-11-20 22:24:55: File Deleted - 13539: File Added - 13578: tip73.patch kennykb added on 2001-11-20 22:24:54: Logged In: YES user_id=99768 Donal's suggestion is incorporated. dkf added on 2001-11-20 17:28:24: Logged In: YES user_id=79902 Stick an extra pair of parentheses around the casting of the difference between seconds (in the changes to the [time] implementation in tclCmdMZ.c) so that it is clear that this operation is being done before the multiply by 1e6, and the change to tclCmdMZ.c will be OK with me (assuming all other changes are OKed too.) I *hate* relying on the obscurer C precedence rules (is the monadic cast operator more tightly binding than the diadic multiply?) especially when doing maintenance. Adding some extra parens removes the ambiguity. :^) kennykb added on 2001-11-20 00:29:05: File Added - 13539: tip73.patch |
Attachments:
- tip73.patch [download] added by kennykb on 2001-11-20 22:24:54. [details]