Re: [APD] Some Problems

On Sunday, September 1, 2002, at 12:47 PM, Christian Dickmann wrote:

Hi, Then I read that I need the CGI version, so maybe these crashes where expected behavior.

You don't need the cgi version (where did you read that?). This is not expected behavior. Are you using the last rolled version from the apd site, current cvs or the execute_wrapper branch? I'm going to assume it's one of the two former ones, as I could duplicate you error there. It's been fixed in cvs (just now).

But now I use the CGI version and apd_set_session_trace() crashes too ... apd_callstack() seems to work. Then I tried to use apd_set_pprof_trace() and used the "pprof" php file at the commandline to view to tracefile. I get a really nice overview, but:

  0.0  0.00  0.01   0.00  0.01   0.00  0.00    40   0.0000    0.00

mdb_manager_mysql_class::parseCurrentBlock

  0.0  0.00  0.00   0.00  0.00   0.00  0.00     1   0.0000    0.00

mdb_manager_mysql_class::setDatabase That can't be right, because parseCurrentBlock() is a method of HTML_Template_IT and no method of MDB ...

There are some issues with the class <-> method identification in the current HEAD. If what you're interested in is the pprof tracing (which imho is more than a full replacement for the session_trace stuff), try doing a cvs update -r execute_wrapper in the apd dir. That enables a profiling methodology which is hoped to be both faster and more accurate. apd_callstack doesn't work there, but if you're using head, then I think that Thies' backtrace function has been integrated into ZE anyway.

Christian Dickmann PS: Are there any Windows Frontends available to use the interactive-debugging? -- PEAR Development Mailing List (http://pear.php.net/) To unsubscribe, visit: http://www.php.net/unsub.php // George Schlossnagle

// Principal Consultant

// OmniTI, Inc          http://www.omniti.com

// (c) 240.460.5234 (e) george@omniti.com // 1024D/1100A5A0 1370 F70A 9365 96C9 2F5E 56C2 B2B9 262F 1100 A5A0