| Ticket UUID: | 726018 | |||
| Title: | cmdName ObjType either wrong or confusing | |||
| Type: | Support | Version: | None | |
| Submitter: | dgp | Created on: | 2003-04-23 04:02:08 | |
| Subsystem: | None | Assigned To: | nobody | |
| Priority: | 5 Medium | Severity: | ||
| Status: | Closed | Last Modified: | 2009-07-29 19:53:22 | |
| Resolution: | Closed By: | dgp | ||
| Closed on: | 2003-05-23 21:26:20 | |||
| Description: |
The tclCmdNameType (aka the "cmdName" Tcl_ObjType) appears to use the otherValuePtr field of the internalRep union to hold its (ResolvedCmdName *) value in some places, but appears to use the twoPtrValue field in other places. I suspect the code is inconsistent, but just happens to work because of the way the two interpretations of the internalRep union lie over each other. It's possible this is intentional, and due to some subtle trickery. If that's the case, then we really need some better comments in there explaining what's going on . Either way, there's a bug to fix. | |||
| User Comments: |
dkf added on 2009-07-29 19:53:22:
IP - Comment Removed: 130.88.1.31 dkf added on 2009-07-29 19:53:21: IP - Comment Removed: 130.88.1.31 dkf added on 2008-12-04 05:25:04: data_type - 210894 [email protected] added on 2008-12-04 02:17:24: M8XPvr dkg93jfbkSdLk496c [email protected] added on 2008-12-04 02:17:10: M8XPvr dkg93jfbkSdLk496c dgp added on 2003-05-24 04:26:20: Logged In: YES user_id=80530 ...and jyl reports the attached patch fixes things. Committing to core-8-4-branch and HEAD. Looks like we'll need an 8.4.4 soon. Sorry.... jyljyljyl added on 2003-05-24 04:25:04: Logged In: YES user_id=63723 Confirmed that the source build of 8.4.3 breaks e4Graph, and the patch submitted by dgp fixes it again. e4Graph's use of twoPtrValue.ptr2 is based on code snarfed from tclBlend, so no surprise. dgp added on 2003-05-24 04:14:38: Logged In: YES user_id=80530 Jacob Levy reports similar trouble with e4graph. mdejong added on 2003-05-24 01:04:25: Logged In: YES user_id=90858 There is indeed a poblem. I just ran the test cases with the current CVS of Tcl and it now deadlocks or crashes on exit. I have not had time to test any patches to see if they fix things. dgp added on 2003-05-23 22:09:26: File Added - 51106: 726018.patch dgp added on 2003-05-23 22:09:25: Logged In: YES user_id=80530 If someone confirms that Tcl 8.4.3 did in fact break TclBlend 1.3.0, can they also check whether this patch to Tcl 8.4.3 fixes the problem? dgp added on 2003-05-22 22:17:40: Logged In: YES user_id=80530 A report on CLT claims this change broke tclBlend 1.3.0. Can you confirm? dgp added on 2003-05-13 02:30:26: Logged In: YES user_id=80530 fixed for 8.4.3 dgp added on 2003-05-13 00:47:56: Logged In: YES user_id=80530 miguel notes on the chat that he believes this is an error, not cleverness. | |||
Attachments:
- 726018.patch [download] added by dgp on 2003-05-23 22:09:25. [details]