bugs in sdb ?

[Karl Denninger]

In article <16899 at rpp386.Dallas.TX.US> jfh at rpp386.cactus.org (John F. Haugh II) writes:
>In article <142 at ssc.UUCP> fyl at ssc.UUCP (Phil Hughes) writes:
>>Sounds like you are doing nothing wrong.  Each new release of sdb
>>has a couple more commands that work.  Maybe someday more than
>>half will. :-)
>>
>>My favorite is single step 10 instructions:
>>s10
>>symbol not found
>
>And well it shouldn't!
>
>s10 is a valid C symbol.  As such, the line 's10' indicates to
>sdb it is to print the value of the symbol 's10' in its proper
>format.
>
>Try adding a space.  Perhaps the other half of the commands
>would work if you inserted a few spaces ;-)  They are free,
>you know!

Ok, how about this:

sdb akcs
*e read.c 		(Open one of the modules)
*101b			(stop here...)
*r
.........(execute the read command, etc.)
Breakpoint at .......
*confs[2].priv/		(This is a valid structure array, member, and request!)
Symbol not found
*

It doesn't work.  The correct results are, however, given by the 3b2 and 3b1 
sdb debuggers, as well as most others I have worked with (even Microport got
this one right!)  If I name the FIRST member in the structure, it DOES work.
Only second and subsequent member names fail.  This, of course, makes it
somewhat difficult to look at the rest of the structure.

If you're trying to debug a '286 program using sdb on a 386, you may as well
forget it.  That generally is as useful as using adb, as it can't even find 
externally-declared variables in large model programs!

Sdb on SCO Xenix is still useful, but it is definately brain-dead.

--
Karl Denninger (karl at ddsw1.MCS.COM, <well-connected>!ddsw1!karl)
Public Access Data Line: [+1 312 566-8911], Voice: [+1 312 566-8910]
Macro Computer Solutions, Inc.		"Quality Solutions at a Fair Price"