Bug 157110

Summary: bmptopnm does not convert valid bitmaps -- reports error instead and segfaults
Product: [Fedora] Fedora Reporter: David Costanzo <david_costanzo>
Component: netpbmAssignee: Jindrich Novy <jnovy>
Status: CLOSED RAWHIDE QA Contact: Ben Levenson <benl>
Severity: high Docs Contact:
Priority: medium    
Version: 4CC: pknirsch
Target Milestone: ---   
Target Release: ---   
Hardware: i686   
OS: Linux   
Whiteboard:
Fixed In Version: 10.27-2 Doc Type: Bug Fix
Doc Text:
Story Points: ---
Clone Of: Environment:
Last Closed: 2005-05-09 10:27:24 UTC Type: ---
Regression: --- Mount Type: ---
Documentation: --- CRM:
Verified Versions: Category: ---
oVirt Team: --- RHEL 7.3 requirements from Atomic Host:
Cloudforms Team: --- Target Upstream Version:
Embargoed:
Attachments:
Description Flags
24bpp-320x240.bmp -- a valid bitmap
none
Proposed fix none

Description David Costanzo 2005-05-06 21:45:08 UTC 
From Bugzilla Helper:
User-Agent: Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.7) Gecko/20050416 Fedora/1.7.7-2

Description of problem:
bmptopnm does not convert valid bitmaps.  Instead, it reports that the header is bogus.


Version-Release number of selected component (if applicable):
netpbm-progs-10.27-1

How reproducible:
Always

Steps to Reproduce:
1. Execute "bmptopnm 24bpp-320x240.bmp"
  

Actual Results:  bmptopnm prints out:

  bmptopnm: Standard Input: unknown Info Header size: 0 bytes

Expected Results:  bmptopnm outputs the bitmap as a pnm file to stdout.  (At least, I think *think* that's what should happen.  I'm new to this toolkit).

Additional info:

I have given this a High severity, even though it's not a crash, memory leak, or loss of data, because it makes the program is useless.
Comment 1 David Costanzo 2005-05-06 21:46:41 UTC 
Created attachment 114103 [details]
24bpp-320x240.bmp -- a valid bitmap

This is the bitmap that I used to reproduce the bug.  Any valid bitmap should
do.
Comment 2 David Costanzo 2005-05-06 22:01:28 UTC 
Created attachment 114106 [details]
Proposed fix

The problem is more severe than I thought.  The bug is in pm_readlittleshort()
and pm_readlittlelong(), which are called by many other programs within the
toolkit (not just bmptopbm).  The bug is that pm_readlittlelong() only called
getch() twice and pm_readlittleshort() only called getch() once.

I checked that similar bugs are NOT present in the big-endian version of these
functions.
Comment 3 Jindrich Novy 2005-05-09 08:36:27 UTC 
Hello David,

the high severity is pretty suitable for this as bmptopnm doesn't work at all in
certain cases. I found a memory corruption in the code where it segfaults
because bmptopnm uses uinitialized pointer for colormaps what it tries to free()
at the end in case BMPheader.cmapsize == 0.
Comment 4 Jindrich Novy 2005-05-09 10:27:24 UTC 
Fixed & rebuilt.