I get this message from VC++ 6.0 Compiling... cPickleCache.c C:\Python\devel\Zope\lib\python\ZODB\cPickleCache.c(323) : warning C4700: local variable 'dt' used without having been initialized line 323 is the second UNLESS inside static PyObject * cc_full_sweep(ccobject *self, PyObject *args) { int dt=0; UNLESS(PyArg_ParseTuple(args, "|i", &dt)) return NULL; UNLESS(-1 != fullgc(self,dt)) return NULL; Py_INCREF(Py_None); return Py_None; } -- Robin Becker
On Fri, 8 Oct 1999, Robin Becker wrote:
I get this message from VC++ 6.0 [snip] C4700: local variable 'dt' used without having been initialized [snip] int dt=0; ^^^ Why doesn't it pick this up?
Run debug on it and see what the value is. Is this the latest CVS version? Cheers, Anthony Pfrunder
In article <Pine.OSF.4.10.9910082251270.5532-100000@student.uq.edu.au>, Anthony Pfrunder <s341625@student.uq.edu.au> writes
On Fri, 8 Oct 1999, Robin Becker wrote:
I get this message from VC++ 6.0 [snip] C4700: local variable 'dt' used without having been initialized [snip] int dt=0; ^^^ Why doesn't it pick this up?
Run debug on it and see what the value is. Is this the latest CVS version?
Cheers,
Anthony Pfrunder
My point exactly! Why doesn't it see this. This is the latest CVS. -- Robin Becker
In article <pqnLtBAU5e$3Ewa3@jessikat.demon.co.uk>, Robin Becker <robin@jessikat.demon.co.uk> writes
In article <Pine.OSF.4.10.9910082251270.5532-100000@student.uq.edu.au>, Anthony Pfrunder <s341625@student.uq.edu.au> writes
On Fri, 8 Oct 1999, Robin Becker wrote:
I get this message from VC++ 6.0 [snip] C4700: local variable 'dt' used without having been initialized [snip] int dt=0; ^^^ Why doesn't it pick this up?
Run debug on it and see what the value is. Is this the latest CVS version?
Cheers,
Anthony Pfrunder
My point exactly! Why doesn't it see this. This is the latest CVS. OK I have found the solution to this.
VC++ 6.0 is apparently substituting out the call to fullgc when optimising and it seems that in my version of the code the fullgc routine has a argument called idt. Inside fullgc it seems as though dt is used. can anyone enlighten me as to what the argument should actually be called? is the argument name a typo? -- Robin Becker
In message <37FE4EF0.ABFB374F@campuspipeline.com>, Kevin Butler <kbutler@campuspipeline.com> writes
My bet is that UNLESS is expanding to something that declares dt, as well as using it.
Note that I haven't looked at the code at all, but that's what the error looks like...
You can try running just the preprocessor on the code (cl /v? /e?) to see preprocessor output.
kb
unfortunately I see static PyObject * cc_full_sweep(ccobject *self, PyObject *args) { int dt=0; if(!(PyArg_ParseTuple(args, "|i", &dt))) return ((void *)0); if(!(-1 != fullgc(self,dt))) return ((void *)0); (((&_Py_NoneStruct))->ob_refcnt++); return (&_Py_NoneStruct); } -- Robin Becker
When I do a search on something on the zope.org home page that is evidently unknown, I get this error message. Try "portal toolkit" or "portal" or "digicool" or "apache" and get: ---copy SiteIndex ! System Unavailable This site is currently experiencing technical difficulties. Please contact the site administrator for more information. For additional technical information, please refer to the HTML source for this page. Thank you for your patience. Error type: KeyError Error value: 520 with "portal" Error value: 673 with "digicool" Error value: 677 with "Apache" ---endcopy If I try "bananas" I get "There was no data..." this is expected. But the above is not. If I try "toolkit", I do get 7 results. Also "zope", "redhat", "amos", "brian" is ok The results are reproducable. Wierd. What is going on? I tested with IE5 and NS4.6. Logged on as guest. PS: When I click the "more information" button on the 520 error page, Nothing happens with IE5 but NS4.6 opens the source page. (Browser inconsistancy!) -Bob bob@rocnet.com
At 04:33 PM 10/9/99 -0400, you wrote:
When I do a search on something on the zope.org home page that is evidently unknown, I get this error message.
Thanks for the bug report. However, zope-dev is not the right place for zope.org site bug reports. You should send messages about zope.org to webmaster@zope.org. BTW, we are well aware of this problem. It has to do with indexes referring to objects which are no longer in the catalog. We *are* working on it. It is not a Zope bug, but an issue in the way we have implemented some ZClasses on Zope.org.
PS: When I click the "more information" button on the 520 error page, Nothing happens with IE5 but NS4.6 opens the source page.
This is a javascript issue. We should probably get rid of this button since it doesn't work on most browsers. Thanks for your input. -Amos
Thanks Amos for the details on this problem.
BTW, we are well aware of this problem. It has to do with indexes referring to objects which are no longer in the catalog. We *are* working on it. It is not a Zope bug, but an issue in the way we have implemented some ZClasses on Zope.org.
Once you solve this "issue" that's not a "bug". Please share the details with everyone so the average developer won't incorrectly implement ZClasses on their own Zope sites. -Bob OConnor ----- Original Message ----- From: Amos Latteier <amos@aracnet.com> To: Robert O'Connor <bob@rocnet.com> Cc: <zope-dev@zope.org> Sent: Saturday, October 09, 1999 7:36 PM Subject: Re: [Zope-dev] Zope Search Broken with certain words.
At 04:33 PM 10/9/99 -0400, you wrote:
When I do a search on something on the zope.org home page that is evidently unknown, I get this error message.
Thanks for the bug report. However, zope-dev is not the right place for zope.org site bug reports. You should send messages about zope.org to webmaster@zope.org. ....
Robert O'Connor wrote:
Once you solve this "issue" that's not a "bug". Please share the details with everyone so the average developer won't incorrectly implement ZClasses on their own Zope sites.
That's the $100,000 question. We have not been able to reproduce the problem reliably yet. It does not apear to be a catalog bug because under no other circumstances other than on the zope site have been able to get this error to happen. -Michel
participants (5)
-
Amos Latteier -
Anthony Pfrunder -
Michel Pelletier -
Robert O'Connor -
Robin Becker