Ticket UUID: | 951220 | |||
Title: | append does not fire a read trace | |||
Type: | Bug | Version: | obsolete: 8.4.2 | |
Submitter: | trott13 | Created on: | 2004-05-10 12:48:26 | |
Subsystem: | 07. Variables | Assigned To: | hobbs | |
Priority: | 5 Medium | Severity: | ||
Status: | Closed | Last Modified: | 2004-09-21 03:57:52 | |
Resolution: | Invalid | Closed By: | hobbs | |
Closed on: | 2004-09-20 20:57:52 | |||
Description: |
The attached file (append.bug) registers a read trace for a variable, doing something (here a puts) as a side effect. I expect the "append" command to be executed as a "read/append/write" sequence, thus firing the read trace. I fact it does _not_ execute the traceProc. I think the "append" command should work (with respect to traces) exactly like the "incr" command (which does it right as a read/modify/write sequence). | |||
User Comments: |
hobbs added on 2004-09-21 03:57:51:
Logged In: YES user_id=72656 There may be an explanation that I can't recall know, but the best reason is because this has been the documented behavior for a long time, so we leave it that way for compatability reasons. msofer added on 2004-05-10 22:07:02: Logged In: YES user_id=148712 Fact is: [append] does not trigger read traces, [lappend] does . The behaviour is enforced by the testsuite (test appendComp-7.x) since forever. I don't know much about this, except that it struck me as weird a while back then. I seem to recall Jeff managed to explain why it was right the way it is, but I forget ... maybe he remembers? Maybe it wasn't Jeff? Let him comment ... dgp added on 2004-05-10 20:43:53: Logged In: YES user_id=80530 miguel, is this a known issue in the "append" operation on array vars? trott13 added on 2004-05-10 19:49:46: File Added - 86659: append.bug |
Attachments:
- append.bug [download] added by trott13 on 2004-05-10 19:49:46. [details]