Thank you very much for very clear and detailed instructions, Eli!
I got the following output:

(gdb) print *ppi2
$1 = {pServerName = 0x0, pPrinterName = 0x38a0136
<__register_frame_info+59375926> "RICOH Aficio 2018D PCL 6",
  pShareName = 0x38a0134 <__register_frame_info+59375924> "", pPortName =
0x38a0118 <__register_frame_info+59375896> "82.97.201.165",
  pDriverName = 0x38a00e6 <__register_frame_info+59375846> "RICOH Aficio
2018D PCL 6", pComment = 0x38a00e4 <__register_frame_info+59375844> "",
  pLocation = 0x38a00e2 <__register_frame_info+59375842> "", pDevMode =
0x389f1c0 <__register_frame_info+59371968>,
  pSepFile = 0x38a00e0 <__register_frame_info+59375840> "", pPrintProcessor
= 0x38a00ce <__register_frame_info+59375822> "winprint",
  pDatatype = 0x38a00c6 <__register_frame_info+59375814> "RAW", pParameters
= 0x38a00c4 <__register_frame_info+59375812> "",
  pSecurityDescriptor = 0x389f088 <__register_frame_info+59371656>,
Attributes = 576, Priority = 1, DefaultPriority = 0, StartTime = 0,
  UntilTime = 0, Status = 0, cJobs = 0, AveragePPM = 0}

I see that pServerName, pShareName are not set, which confuses
Fdefault_printer_name big time (e.g. it thinks the printer is local).
It seems that in my case the problem lies is a very exotic way of
configuration. I'll try to setup printer in the orthodox way when I'm in
the office next time. Current configuration was done by IT guy; I'm not
sure why he decided to use numeric IP addess and port instead of simple
share name.

Thank you for your support,
Andrey Paramonov


2013/12/6 Eli Zaretskii <eliz@gnu.org>

> > From: Андрей Парамонов <cmr.pent@gmail.com>
> > Date: Fri, 6 Dec 2013 16:00:34 +0400
> > Cc: 16068@debbugs.gnu.org
> >
> > > Is it possible for you to step through default-printer-name with GDB
> > > and see what happens there?  I can send you detailed instructions, if
> > > you need them.
> > >
> >
> > Please do.
>
> See below, and thanks in advance.
>
> > Is it correct that default-printer-name is implemented in C (which is
> why I
> > need GDB)?
>
> Yes, this is correct.
>
> I see that you are using Emacs 24.3, which is an official release.  I
> suspect that this binary has been stripped of debugging symbols, so
> you will be unable to debug it on the C source level.  So the first
> thing I suggest to do after installing GDB (let me know if you need a
> link for a Windows port of GDB) is type this from the shell prompt:
>
>   gdb /path/to/emacs.exe
>
> If GDB says something like this:
>
>  Reading symbols from D:\foo\bar\emacs.exe...(no debugging symbols
> found)...done.
>                                               ^^^^^^^^^^^^^^^^^^^^^^^^^^
> then indeed the binary is stripped.  In that case, please download a
> recent snapshot from here:
>
>   https://sourceforge.net/projects/emacs-bin/files/snapshots/
>
> and start by verifying that the same problem still exists there.  (It
> could be that it was somehow fixed since 24.3 was released, although I
> doubt that.)
>
> Once you establish that the problem still exists in the snapshot,
> here's the sequence of commands I'd like you to perform:
>
>   gdb /path/to/emacs.exe
>
> After GDB starts and displays its startup blurb, do this:
>
>   (gdb) break Fdefault_printer_name
>   (gdb) run -Q
>
> The last command runs Emacs.  When Emacs starts up, it will show the
> *scratch* buffer.  Type into that buffer:
>
>   (default-printer-name)
>
> Then go to the right parenthesis and type C-j.  This evaluates the
> expression.  (Please forgive me if I'm telling you things you know
> very well.)
>
> When you evaluate that expression, the breakpoint you have set will
> break, Emacs will stop, and GDB will take control.  Switch to the
> window where you run GDB, you should see there something like this:
>
>   Breakpoint 1, Fdefault_printer_name () at w32fns.c:7217
>   7217      PRINTER_INFO_2 *ppi2 = NULL;
>   (gdb)
>
> (Since you don't have the sources, you will probably not see the
> second line.)
>
> If you want to see the source code of default-printer-name, you can
> find it here:
>
>   http://bzr.savannah.gnu.org/lh/emacs/trunk/annotate/head:/src/w32fns.c
>
> (The function's source starts around line 7210 in that file.)
>
> We want to get to line #7252, which is marked in the snippet below:
>
>   err = GetPrinter (hPrn, 2, (LPBYTE)ppi2, dwNeeded, &dwReturned);
>   ClosePrinter (hPrn);
>   if (!err)
>     {
>       xfree (ppi2);
>       return Qnil;
>     }
>
>   if (ppi2)  <<<<<<<<<<<<<<<<<<<<<<<<<<<<<
>     {
>       if (ppi2->Attributes & PRINTER_ATTRIBUTE_SHARED && ppi2->pServerName)
>         {
>
> The GDB command to do that is:
>
>   (gdb) until 7252
>
> (The "(gdb)" part is GDB's prompt, you don't need to type it.)
>
> GDB will step Emacs till that line, and respond like this:
>
>   Fdefault_printer_name () at w32fns.c:7252
>   7252      if (ppi2)
>   (gdb)
>
> Now please tell GDB to display the contents of the ppi2 structure,
> which holds the printer information Windows returned due to the call
> to GetPrinter API:
>
>   (gdb) print *ppi2
>
> And post here everything you see.
>
> If you will do this in Emacs 24.3 (in case it does have the debugging
> symbols), then please use "until 6943" instead of 7252 above, because
> the source line numbers of the snippet I show above are different in
> that version.
>
> Thanks again for your help.
>