bugPSPP - Bugs: bug #26861, REGRESSION Crashes on alpha...

 
 
Show feedback again

bug #26861: REGRESSION Crashes on alpha variables.

Submitted by:  John Darrington <jmd>
Submitted on:  Sun 21 Jun 2009 09:21:01 AM UTC  
 
Category: Syntax ParserSeverity: 5 - Average
Status: ConfirmedAssigned to: Jason H Stover <jstover>
Open/Closed: OpenRelease: None
Effort: 0.00

Add a New Comment (Rich MarkupRich Markup):
   

You are not logged in

Please log in, so followups can be emailed to you.

 

(Jump to the original submission Jump to the original submission)

Sun 25 Oct 2009 08:29:53 AM UTC, comment #8:

What's the status of this item? Has it been fixed or is there still a problem?

John Darrington <jmd>
Project Administrator
Mon 20 Jul 2009 08:04:45 PM UTC, comment #7:

Jason: thanks for the fix.

When I apply this to stable (without pushing it), however, I get the following results. I think that these are still wrong. Is there a missing patch?

Ben Pfaff <blp>
Project Administrator
Fri 17 Jul 2009 08:07:21 PM UTC, comment #6:

I just pushed a fix for the numerical errors in the output.
Please test to make sure it's OK, and I'll close this bug.
You should see these tables as part of the output:

2.1 REGRESSION. Model Summary
#=============#========#=================#==========================#
# R #R Square|Adjusted R Square|Std. Error of the Estimate#
#========#====#========#=================#==========================#
# |1.00# 1.00| 1.00| .00#
#========#====#========#=================#==========================#

2.2 REGRESSION. ANOVA
#===================#==============#==#===========#========#============#
# #Sum of Squares|df|Mean Square| F |Significance#
#========#==========#==============#==#===========#========#============#
# |Regression# 2.00| 1| 2.00|+Infinit| .00#
# |Residual # .00| 1| .00| | #
# |Total # 2.00| 2| | | #
#========#==========#==============#==#===========#========#============#

Jason H Stover <jstover>
Project MemberIn charge of this item.
Fri 17 Jul 2009 03:38:21 PM UTC, comment #5:

The residual sum of squares should be 0.0 here, instead of -5.5. It's not the same bug that was causing the segfault, though. I'll look into it.

Jason H Stover <jstover>
Project MemberIn charge of this item.
Fri 17 Jul 2009 03:29:51 AM UTC, comment #4:

>Please test the attached patch.


That fixes the valgrind complaints.

I pushed a slightly modified version of this patch to the pre0.6.2 patch (there is no design_matrix_get_n_cols function on that branch).

The output then looks like this, is that correct? I see some NaNs so I'm not sure:

Ben Pfaff <blp>
Project Administrator
Thu 16 Jul 2009 06:39:22 PM UTC, comment #3:

Please test the attached patch.

(file #18425)

Jason H Stover <jstover>
Project MemberIn charge of this item.
Tue 14 Jul 2009 03:01:56 AM UTC, comment #2:

Running valgrind shows that the memory allocated on line 944 of regression.q:
is being accessed after the end of the allocated space in this loop spanning line 974:
and in this loop on lines 324 to 326 of linreg.c:
My guess is that more space should be allocated, but I am hoping that you will know for sure.

The same valgrind run also shows uninitialized data being passed to tab_text() on line 240 of regression.q:
but sometimes that sort of thing can be a side effect of earlier errors and not a separate bug.

Ben Pfaff <blp>
Project Administrator
Sat 11 Jul 2009 07:47:24 PM UTC, comment #1:

I was unable to reproduce the segfault, but I did get NAN's as the
estimated coefficients. I just pushed a fix for the NAN's, let me know if this also fixes your segfaults.

Jason H Stover <jstover>
Project MemberIn charge of this item.
Sun 21 Jun 2009 09:21:01 AM UTC, original submission:

This syntax causes a seg fault:

DATA LIST LIST /x * y * a (a23).
BEGIN DATA.
1 11 One
2 22 Two
3 33 Three
END DATA.

REGRESSION
/VARIABLES= a
/DEPENDENT= x y
/STATISTICS=COEFF R ANOVA.

John Darrington <jmd>
Project Administrator

 

(Note: upload size limit is set to 512 kB, after insertion of the required escape characters.)

Attach File(s):
   
   
Comment:
   

Attached Files
file #18425:  regression.patch added by jstover (958B - text/x-diff)

 

Depends on the following items: None found

Items that depend on this one: None found

 

Carbon-Copy List
  • -unavailable- added by blp (Posted a comment)
  • -unavailable- added by jstover (Posted a comment)
  • -unavailable- added by jmd (Submitted the item)
  •  

    Do you think this task is very important?
    If so, you can click here to add your encouragement to it.
    This task has 0 encouragements so far.

    Only logged-in users can vote.

     

    Please enter 421 here (basic anti-spam test):

     

     

    Follow 3 latest changes.

    Date Changed By Updated Field Previous Value => Replaced By
    Thu 16 Jul 2009 06:39:22 PM UTCjstoverAttached File-=>Added regression.patch, #18425
    Tue 14 Jul 2009 03:01:56 AM UTCblpAssigned toNone=>jstover
    Tue 14 Jul 2009 03:01:55 AM UTCblpStatusNone=>Confirmed
    Show feedback again

    Back to the top


    Powered by Savane 3.1-cleanup