This is the mail archive of the cygwin@cygwin.com mailing list for the Cygwin project.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: Samba Client compilation on the latest DLL


No program is ever bug free! ;)

I have located where the problem is within samba... It is within the
string handling library that they custom write so they can handle
different character sets/code maps.

And yes, I have been caught out by OS changes before. I'm a windows
developer by trade and have been caught out my Microsoft actually fixing
their APIs (odd, they do actually fix stuff).

The reason this is annoying me a little though, is that the same version
of gcc/glibc on my linux workstation runs this version of smbclient
absolutely fine. So that implies that gcc/glibc/stupid windows memory
handling is at fault somewhere?

I ripped out the custom string stuff in samba and replaced a load of
higher up calls (lib/debug.c mainly... which is where ALL printing to the
screen is done... another silly thing to do) and it worked a treat.

I know it may have sounded that I came on here mouthing that Cygwin was at
fault. That was not my intention and I am sorry if it came across that
way.

I have done these tests on multiple installations of cygwin on different
machines/OS's (2k and XP to be exact) and the problem still exists. Both
times, they were fresh Cygwin installs. I even reinstalled the 2k one to
make sure!

As I said in an earlier post, i've been maintaining (the now, very popular
with sysadmins) smbclient for Win32 for nearly a year and have NEVER had a
problem. Like, EVER. The binary on the page with the dlls that are
included in the zip file runs absolutely fine. Take the binary and run it
with the new DLLs it works fine. This implies something is different with
compilation as samba doesn't link anything statically...

So... /me reinstalled cygwin to use the old 2.95 compiler... No joy. Same
problem. Hence why I came here... hence why i'm stuck.

> Lee,
>
> Surely all but the most inexperienced programmers using C or C++ know
> that  a program that executes without overt failure cannot be considered
>  bug-free? That a perfectly valid change in the compiler, the linker,
> the  libraries or a switch to a different platform (which usually means
> all of  these change) can make formerly latent bugs become manifest,
> right?
>
> Considering that your program is not able to get off the ground, as it
> were, then you may have an unholy combination of libraries, run-times or
>  compiler options. Complex software with big configuration scripts and
> large  (and / or multiple) makefiles can sometimes cause such unsound
> mixtures.
>
> It's also possible that there is some corruption in a library with which
>  you're linking. If this is so and unless it's a fairly subtle
> corruption it  might be visible in the debugger by viewing the assembly
> code at or near  the point of failure.
>
>
> By the way, I'm certainly not offended, but I take it you were. That was
>  not my intent. I'm sorry.
>
> Randall Schulz
> Mountain View, CA USA
>
>
> At 10:49 2002-11-29, Lee Packham wrote:
>>Then, I start to wonder why smbclient fails on a line where the example
>> I  give, fails.
>>
>>/me goes back to drawing board.
>
>
> --
> Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
> Bug reporting:         http://cygwin.com/bugs.html
> Documentation:         http://cygwin.com/docs.html
> FAQ:                   http://cygwin.com/faq/




--
Unsubscribe info:      http://cygwin.com/ml/#unsubscribe-simple
Bug reporting:         http://cygwin.com/bugs.html
Documentation:         http://cygwin.com/docs.html
FAQ:                   http://cygwin.com/faq/


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]