xcircuit-dev — XCircuit news and discussion

[Xcircuit-dev] Errors/Fixes

[R. T. E. <ti...@st...>]

Dear Massimo,

   Thanks for the bug reports.  I've been at a conference for several
days, so your bug reports have been piling up.  Here are the results:

---------------------------------------------
1) Start xcircuit
2) Load the attached bug?.lps library
3) Go to the bug? library ('l' command)
4) Edit the symbol bug? ('>' command)
5) Select the ? (only the '?') in the info label 'spice:R?' with the 
middle mouse key
6) Go to text -> unparameterize
7) segmentation fault
---------------------------------------------

I can't get this one to crash.  I *did* make a few changes to the
"unparameterize" function code between version 2.5.4 and version
3.0 (2.5.5), so my best guess is that the problem got fixed already.

---------------------------------------------
1) run xcircuit
2) "Import Xcircuit ps" the attached example (bug.ps)
3) Go to Netlist -> Make Matching Symbol
4) Input the name of the symbol (i.e. "All")
5) Now instead to draw the symbol press "<"
6) Segmentation fault
---------------------------------------------

Ouch.
I see I missed something I was supposed to remove.
"events.c" line 398:  Remove lines 398 to 401

#ifdef SCHEMA
   if (xobjs.stack->thisinst->thisobject->symschem == topobject)
      pop_stack(&xobjs.stack);
#endif

and the problem will go away.  However, the same change I made which
made those two lines necessary also produces the strange behavior that
when you go from a top page to the library, then edit (">") the library
object, the page appears drawn in gray as if you pushed from the page
and not the library.  I'll have to find a fix, although the problem is
minor.

---------------------------------------------
1) Load the attached library pippo1.lps
2) go to the library and ">" the nMesfet
3) insert an Info pin (I) and write "spice:Z? %pd %pg %ps ?"
4) Highlight (with the middle button) the first ? and parameterize it
5) Highlight the second ? and parameterize it
6) Segmentation fault. It is reproducible
---------------------------------------------

Ouch again!
"parameter.c", line 1105.  Change:

      begpart = splitstring(textend, &thislabel->string, NULL);
      endpart = splitstring(textpos, &thislabel->string, NULL);

to

      begpart = splitstring(textend, &thislabel->string, areastruct.topinstance);
      endpart = splitstring(textpos, &thislabel->string, areastruct.topinstance);

This fixes the problems.
The fixes will be reflected in the distribution as soon as the lab's ISP
allows me to connect to the university machine again. . .  (it amazes me
that the entire internet seems to be available except for the connection
between the university and its own research laboratory!)

						Regards,
						Tim

[Xcircuit-dev] Bug Report - #1

[Massimo G. <ga...@ti...>]

Dear Tim,

I think to have found another strange bug (starting from now I am 
numbering my bug reports, this is #1 :-) )

Case 1) Using: Xcircuit 3.0 dated 8 Jul 2002 + patch for Python 
(PyRun_StringFlags (...) ), compiled with
./configure --disable-focus-fix
make
make install

1) Load the attached library pippo1.lps
2) go to the library and ">" the nMesfet
3) insert an Info pin (I) and write "spice:Z? %pd %pg %ps ?"
4) Highlight (with the middle button) the first ? and parameterize it
5) Highlight the second ? and parameterize it
6) Segmentation fault. It is reproducible

Case 2) Using Xcircuit 2.5 from Slackware package downloaded from your site

This bug is not present.

Case 3) Compiling like in Case1 without --disable-focus-fix

exactly identical to Case 1)


Case 4)

Running the debugger (xxgdb) over the executable obtained in Case3)
I get


Program received signal SIGSEGV, Segmentation Fault
[switching to Thread 1024 (LWP 5182)]
linkstring (localinst=0x0, strstart=0x8268c50) at text.c:310

The line 310 of text.c is
310: if (localinst->thisobject->params[paramno]->type != XC_STRING) {
311:      fprintf(stderr, "linkstring() error: parameter %d is not a 
               substring!\n",
312:         paramno);
313      return strstart->nextpart;
314   }

I haven't the source code of the 2.5 version, so I cannot compare the 
differences. Anyway the error message (line 311:) is not printed on stderr.


I hope you will able to get the error (bug).

Hope this helps.

Massimo



-- 

                          ''~``
                         ( o o )
+------------------.oooO--(_)--Oooo.------------------+
|                                                     |
|          e-mail:  address ga...@ti...            |
|          PGP key available on keyserver             |
|PGP fingerprint16:                                   |
|  76 80 F2 F9 8D 70 F3 D1  42 2B CD 80 29 49 CB 25   |
|                                                     |
|                    .oooO                            |
|                    (   )   Oooo.                    |
+---------------------\ (----(   )--------------------+
                        \_)    ) /
                               (_/

[Xcircuit-dev] Bug report July 15 2002

[Massimo G. <ga...@ti...>]

Dear Tim and xcircuit users,

Try this

1) run xcircuit
2) "Import Xcircuit ps" the attached example (bug.ps)
3) Go to Netlist -> Make Matching Symbol
4) Input the name of the symbol (i.e. "All")
5) Now instead to draw the symbol press "<"
6) Segmentation fault


Hope this helps

Regards

Massimo

-- 

                          ''~``
                         ( o o )
+------------------.oooO--(_)--Oooo.------------------+
|                                                     |
|          e-mail:  address ga...@ti...            |
|          PGP key available on keyserver             |
|PGP fingerprint16:                                   |
|  76 80 F2 F9 8D 70 F3 D1  42 2B CD 80 29 49 CB 25   |
|                                                     |
|                    .oooO                            |
|                    (   )   Oooo.                    |
+---------------------\ (----(   )--------------------+
                        \_)    ) /
                               (_/

[Xcircuit-dev] Just another bug report (may be :-) )

[Massimo G. <ga...@ti...>]

Dear Tim and Xcircuit users,

Try this.

1) Start xcircuit
2) Load the attached bug?.lps library
3) Go to the bug? library ('l' command)
4) Edit the symbol bug? ('>' command)
5) Select the ? (only the '?') in the info label 'spice:R?' with the 
middle mouse key
6) Go to text -> unparameterize
7) segmentation fault

This bug is present in my Slackware 8.0 linux station. I don't know
if the bug is present in other OS or distribution.

(Slackware 8.0 is kernel 2.2.19 glibc 2.2.3 Xcircuit 2.5.4 pre-compiled 
slackware package from Tim's site. )


Another strange behaviour

1) Start Xcircuit from a shell prompt
2) Iconify it
3) you get
X Error of failed request:  BadMatch (invalid parameter attributes)
   Major opcode of failed request:  42 (X_SetInputFocus)
   Serial number of failed request:  9390
   Current serial number in output stream:  9390

and the Xcircuit dies.

Very strange. May be relate to the X (Xfree 4.1.0 and/or  fvwm 2.5.5 
window manager).

Another issue:

As I told in my last e-mail I cannot compile the new 3.0 (8 jul) version
with python 2.0.1 due to lack of PyRun_StringsFlags symbol int the 
python library. I observed that the new Python 2.2.1 has this reference 
in the library. So definitevely I need to upgrade the Python package 
and/or toupgrade to Slackware 8.1. Anyway may be a good idea to put a 
warning in the web site about this issue, 2.0.1 is not enough! :-)
I guess that the Slackware package in the web site was not compiled 
under pure Slackware 8.0 (that has python 2.0.1), it is impossible.

Hope this helps


Regards

Massimo






-- 

                          ''~``
                         ( o o )
+------------------.oooO--(_)--Oooo.------------------+
|                                                     |
|          e-mail:  address ga...@ti...            |
|          PGP key available on keyserver             |
|PGP fingerprint16:                                   |
|  76 80 F2 F9 8D 70 F3 D1  42 2B CD 80 29 49 CB 25   |
|                                                     |
|                    .oooO                            |
|                    (   )   Oooo.                    |
+---------------------\ (----(   )--------------------+
                        \_)    ) /
                               (_/

[Xcircuit-dev] Xcircuit 3.0 Compilation Problems

[Massimo G. <ga...@ti...>]

Dear Tim and xcircuit users,

I am trying to compile the new 3.0 version (8 jul) but I am unable (so 
far) to get an working executable.

I am compiling under Slackware 8.0 with Phyton 2.0.1,
Tcl/Tk 8.3.3 installed and Tclx 8.3.0 installed.


Compiling after a

./configure
make

I get

gcc  -O2 -L/usr/X11R6/lib -L/usr/lib/python2.0/config -o xcircuit 
elements.o events.o filelist.o files.o fontfile.o formats.o functions.o 
help.o keybindings.o libraries.o menucalls.o netlist.o parameter.o 
python.o rcfile.o render.o schema.o selection.o text.o wrapper.o 
tkSimple.o xcircuit.o Xw/libxcXw.a -lXt -lm  -lpython2.0 
  -lpthread -ldl  -lutil          -lm -lXpm -lX11
/usr/lib/python2.0/config/libpython2.0.a(posixmodule.o): In function 
`posix_tmpnam':
posixmodule.o(.text+0x25fe): the use of `tmpnam_r' is dangerous, better 
use `mkstemp'
/usr/lib/python2.0/config/libpython2.0.a(posixmodule.o): In function 
`posix_tempnam':
posixmodule.o(.text+0x255a): the use of `tempnam' is dangerous, better 
use `mkstemp'
python.o: In function `PyRun_ObjString':
python.o(.text+0x33): undefined reference to `PyRun_StringFlags'
collect2: ld returned 1 exit status
make[1]: *** [xcircuit] Error 1
make[1]: Leaving directory `/user/gaspari/compile/xcircuit-3.0'
make: *** [all-recursive] Error 1


Compiling after a

./configure --enable-wrapper
make wrapper

I get

swig -tcl8 -o xcwrap.c xcwrap.i
make: swig: Command not found
make: *** [xcwrap.c] Error 127

I loooked in all the installed packaged, I have not found the
required 'swig'.

I never used Tcl/Tk so I will try to understand what is the required 
swig executable.

In the first case I guess that you are using a new version of Python.
Really the 'PyRun_StringFlags' is not defined in my libpython2.0.a

Doing a nm libpython2.0.a I get the 'PyRun_String' reference but not the

PyRun_StringFlags

I have attached the two config.log files.

I will try to understand more details tomorrow.

In addition I am reporting a compilation warning too

gcc -DPACKAGE=\"xcircuit\" -DVERSION=\"3.0\" -DPROTOTYPES=1 
-DHAVE_LIBM=1 -DHAVE_LIBXT=1 -DSTDC_HEADERS=1 -DHAVE_SETENV=1 
-DHAVE_PUTENV=1 -DHAVE_DIRENT_H=1 -DHAVE_U_CHAR=1 -DSCHEMA=1 
-DSCHEM_FLAG=\"-schema\" -DLGF=1 -DINPUT_FOCUS=1 -DGS_EXEC=\"gs\" 
-DHAVE_PYTHON=1 -DHAVE_X11_XPM_H=1 -DHAVE_XPM=1  -I. -I.  -IXw 
-DPROG_VERSION=3.0 -DTEMP_DIR=\"/tmp\" 
-DRESOURCES_DIR=\"/usr/local/share/xcircuit-3.0/app-defaults\" 
-DBUILTINS_DIR=\"/usr/local/share/xcircuit-3.0\" -DBUILTINS_FILE=\"\" 
-DUSER_RC_FILE=\".xcircuitrc\" 
-DPROLOGUE_DIR=\"/usr/local/share/xcircuit-3.0\" 
-DPROLOGUE_FILE=\"xcircps2.pro\" -DSTARTUP_FILE=\"xcstartup.py\" 
-DCONFIG_FILE=\"\" -DLGF_LIB=\"lgf.lps\"  -I/usr/X11R6/include 
-DDOUBLEBUFFER -I/usr/include/python2.0  -g -O2 -c python.c
python.c: In function `PyRun_ObjString':
python.c:149: warning: assignment makes pointer from integer without a cast


Hope this helps.

Regards

Massimo


-- 


                          ''~``
                         ( o o )
+------------------.oooO--(_)--Oooo.------------------+
|                                                     |
|          e-mail:  address ga...@ti...            |
|          PGP key available on keyserver             |
|PGP fingerprint16:                                   |
|  76 80 F2 F9 8D 70 F3 D1  42 2B CD 80 29 49 CB 25   |
|                                                     |
|                    .oooO                            |
|                    (   )   Oooo.                    |
+---------------------\ (----(   )--------------------+
                        \_)    ) /
                               (_/

[Xcircuit-dev] Xcircuit bug report (additional)

[Massimo G. <ga...@ti...>]

Dear all,

In my last bug report I wrote some key combination using bracket. The 
web interface of the sourceforge repository has trouble with this. Some 
characters are disappeared. So I repost in a different way two lines of 
my last e-mail


.
.
.
.
6) Go to the end of the string (after the parameter ?) and out of the 
angle bracket and press Alt+Enter

7) Press Alt+P and then "2" to add the second parameter (the model)
.
.
.


Sorry for the time waste.

Bye

Massimo


-- 

                          ''~``
                         ( o o )
+------------------.oooO--(_)--Oooo.------------------+
|                                                     |
|          e-mail:  address ga...@ti...            |
|          PGP key available on keyserver             |
|PGP fingerprint16:                                   |
|  76 80 F2 F9 8D 70 F3 D1  42 2B CD 80 29 49 CB 25   |
|                                                     |
|                    .oooO                            |
|                    (   )   Oooo.                    |
+---------------------\ (----(   )--------------------+
                        \_)    ) /
                               (_/

[Xcircuit-dev] Patch for parameter selection

[R. T. E. <ti...@st...>]

Dear Massimo,

    There are effectively no differences between xcircuit-2.5.4 and
xcircuit-3.0; the difference between the two versions is some
experimental Tcl/Tk code that is not compiled by default.

    So, anyway, the bug you found is that the parameter numbers shown
for selection when you use "alt-P" are wrong; the routine counts only
the string parameters (not the numerical parameters), but forgot to
add them to the parameter count when mapping the selection number
back to the parameter number.  Consequently, you get a numerical
parameter inserted into the string, causing a core dump.

    Attached is a patch file which cures the problem.  Thanks for the
detailed bug report.

						Regards,
						Tim

+--------------------------------+-------------------------------------+
| Dr. R. Timothy Edwards (Tim)   | email: ti...@st...    |
| Mailstop 4-234 (SRI)           | web:   http://bach.ece.jhu.edu/~tim |
| JHU Applied Physics Laboratory | phone: (240) 228-4613               |
| 11100 Johns Hopkins Road       | fax:   (240) 228-7636               |
| Laurel, MD  20723-6099         | cell:  (240) 461-3895               |
+--------------------------------+-------------------------------------+

[Xcircuit-dev] Xcircuit bug report

[Massimo G. <ga...@ti...>]

Dear Tim and Xcircuit users,

May be I found a bug in Xcircuit (version 2.5.4 Slackware 8.0 libc 2.2.1 
  kernel 2.2).


1) "Load new library" the attached "pippo.lps"

2) go to the library page "pippo"

3) Look at the NPN

4) Put the cursor on it and press ">" to edit the symbol

5) put the cursor on the text label "Q?" and press "e" to edit

6) Go to the end of the string (after the parameter ?) and out of the <> 
and press <Alt-Enter>

7) Press <Alt-P> and then "2" to add the second parameter (the model)

8) Segmentation Fault

Is the 3.0 affected?

I haven't applied, so far, the patch for the "more than 9 parameters 
problem" reported few days ago.

For the future I will compile Xcircuit with the -g option.
So my reports will be clearer attaching stuff from the gnu debugger.

Hope this helps

Regards

Massimo


-- 

                          ''~``
                         ( o o )
+------------------.oooO--(_)--Oooo.------------------+
|                                                     |
|          e-mail:  address ga...@ti...            |
|          PGP key available on keyserver             |
|PGP fingerprint16:                                   |
|  76 80 F2 F9 8D 70 F3 D1  42 2B CD 80 29 49 CB 25   |
|                                                     |
|                    .oooO                            |
|                    (   )   Oooo.                    |
+---------------------\ (----(   )--------------------+
                        \_)    ) /
                               (_/

[Xcircuit-dev] RE: xcircuit bug? (!)

[R. T. E. <ti...@st...>]

Dear Massimo,

   Oops, typo!  I guess I never made a part with more than 9 parameters
before, because the file reader will only look at the last digit.  The
fix is to change one word in files.c (in version 3.0, it's line 2772; it
may be a few line numbers different in version 2.5.4):

Change:

         else if (!strcmp(keyword, "beginparm")) { /* parameterized object */
            short tmpnum, i;
            for (--keyptr; *keyptr == ' '; keyptr--);
            for (; isdigit(*keyptr) && (keyptr >= lineptr); keyptr--); 
                                                  ^^^^^^^

to:

         else if (!strcmp(keyword, "beginparm")) { /* parameterized object */
            short tmpnum, i;
            for (--keyptr; *keyptr == ' '; keyptr--);
            for (; isdigit(*keyptr) && (keyptr >= buffer); keyptr--); 
                                                  ^^^^^^

Thanks for the detailed bug report.  It made it very easy to track down
the source of the problem.
						Regards,
						Tim

[Xcircuit-dev] Xcircuit bug?

[Massimo G. <ga...@ti...>]

Dear Xcircuit users


I have discovered a bug (may be!) in xcircuit 2.5.4 downloaded from
Tim's website (slackware package). Using Slackware 8.0
(kernel 2.2.18, libc 2.2.1, package locale not installed) I get trouble
doing:

A) I create a custom library (see attachment libok.lps)

B) this library seems good.

C) I run xcircuit, load the libok.lps

D) Editing the virtual simbol I parameterize the position of the text label

E) Seems good

F) I save the library as libcrash.lps

G) I close xcircuit and then run xcircuit again

F) I load libcrash.lps

then xcircuit get a crash saying


readlabel() error:  Parameter 2 exceeds declared number of parameters (1)!
readlabel() error:  Parameter 3 exceeds declared number of parameters (1)!
readlabel() error:  Parameter 4 exceeds declared number of parameters (1)!
readlabel() error:  Parameter 7 exceeds declared number of parameters (1)!
readlabel() error:  Parameter 5 exceeds declared number of parameters (1)!
readlabel() error:  Parameter 6 exceeds declared number of parameters (1)!
readlabel() error:  Parameter 6 exceeds declared number of parameters (1)!
Segmentation fault


But the line
(0) (0) (k) (1) (0.0) (?) (1.0)  0 16 64 -64 11 beginparm
is declaring eleven parameter and I have eleven parameter.

I have recompiled 2.5.4 on my system but the errors and crash are always 
the same
Is this bug affecting 3.0 too?

Hope this helps.

Regards

Massimo


-- 

                          ''~``
                         ( o o )
+------------------.oooO--(_)--Oooo.------------------+
|                                                     |
|          e-mail:  address ga...@ti...            |
|          PGP key available on keyserver             |
|PGP fingerprint16:                                   |
|  76 80 F2 F9 8D 70 F3 D1  42 2B CD 80 29 49 CB 25   |
|                                                     |
|                    .oooO                            |
|                    (   )   Oooo.                    |
+---------------------\ (----(   )--------------------+
                        \_)    ) /
                               (_/

[Xcircuit-dev] Compiling on Solaris 8 intel Architecture

[Mario E. M. <mar...@te...>]

Hi Everybody.

I'm trying to compile XCircuit for Solaris 8 intel architecture. I configure and 
then I "make" getting the following error:

gcc -o menudep menudep.o
./
mksh: Fatal error: Cannot load command `./': Permisson denied
Current working directory /export/home/mario/software/ingenieria/xcircuit-2.3.3
*** Error code 1
make: Fatal error: Command failed for target `menudep.h'
Current working directory /export/home/mario/software/ingenieria/xcircuit-2.3.3
*** Error code 1
make: Fatal error: Command failed for target `all-recursive'
$ 

Anybody can help me? I'm not a professional programmer, so I can't solve the 
problem myself. 

Thank you very much.

Re: [Xcircuit-dev] XCircuit bug report

[Bryce D. <br...@tl...>]

On Mon, 10 Jun 2002, R. Timothy Edwards wrote:
> Dear Bryce,
>    Thanks for the bug report.  The solution is:
> 
> In functions.c, line 1272:  change---
> [snip]
>                goto nextgen;
> [snip]
> nextgen:
> I will make the change to the distribution.

Thanks for the quick response.  It fixed it!
For those who like patches, the equivalent patch is attached.

-Bryce

[Xcircuit-dev] XCircuit bug report

[R. T. E. <ti...@st...>]

Dear Bryce,
   Thanks for the bug report.  The solution is:

In functions.c, line 1272:  change---

-------------------------------------------------------------------
         if ((*bboxgen)->type == LABEL) {
            labelptr btext = TOLABEL(bboxgen);
            if (btext->pin && !(btext->justify & PINVISIBLE)) {
               continue;
            }
         }
#endif
         calcbboxsingle(bboxgen, thisinst, &llx, &lly, &urx, &ury);
      }
      if (newelement != NULL) break;
   }
-------------------------------------------------------------------

to:

-------------------------------------------------------------------
         if ((*bboxgen)->type == LABEL) {
            labelptr btext = TOLABEL(bboxgen);
            if (btext->pin && !(btext->justify & PINVISIBLE)) {
               goto nextgen;
            }
         }
#endif
         calcbboxsingle(bboxgen, thisinst, &llx, &lly, &urx, &ury);
      }
nextgen:
      if (newelement != NULL) break;
   }
-------------------------------------------------------------------

I will make the change to the distribution.

						Regards,
						Tim

Re: [Xcircuit-dev] bug #565180: in 2.5.4, 2.5.5, copying two pins -> freeze

[ken r. <ke...@24...>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

i've seen this too. 

figured it was pilot error, i.e. i was "doing it wrong".

- -ken
- -------
On Thu, Jun 06, 2002 at 03:06:54AM -0400, Bryce Denney wrote:
> I've added another source forge bug
> 
> bug #565180: ] 2.5.4, 2.5.5: copying two pins -> freeze
> http://sourceforge.net/tracker/index.php?func=detail&aid=565180&group_id=4952&atid=104952
> 
> Brief summary:
>                       
> Starting with version 2.5.4, xcircuit freezes when you copy more than one
> pin at a time inside a module.  Versions 2.5.3 and 2.4.0 do not have the
> problem.
> 
> To reproduce this, start up version 2.5.4 or later.  Press 'L' for library
> screen, click on ic_templates, click on any module and place a copy on
> page 1. Then move the mouse over it and press '>' to descend into the
> module.
> 
> Now select two or more pins using the middle button.  (Either by dragging
> a box or by clicking on them.)  Press 'c' to start a copy.  Drag them to
> another location and click the left button to place them.  As soon as you
> click, xcircuit freezes and no longer responds, redraws the screen, etc.
> Escape and right button have no effect. The only way to stop the operation
> is a Control-C to kill the process.
> 
> Copying one pin at a time is ok. Copying two non-pin objects is ok.
> Copying one pin and any number of non-pin objects is ok. But if your
> selection includes 2 or more pins, the copy will crash the program.
> 
> For more details, see the source forge bug report.
> Has anyone else seen this?
> 
> Thanks,
> Bryce
> 
> 
> _______________________________________________________________
> 
> Don't miss the 2002 Sprint PCS Application Developer's Conference
> August 25-28 in Las Vegas -- http://devcon.sprintpcs.com/adp/index.cfm
> 
> _______________________________________________
> Xcircuit-dev mailing list
> Xci...@li...
> https://lists.sourceforge.net/lists/listinfo/xcircuit-dev

- -- 
- ---------------
Freedom of the press belongs to those who can afford to pay the hosting bill.
Now, that means you! http://www.nearlyfreespeech.net
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iD8DBQE9APzRe8HF+6xeOIcRApivAJ9Hp5/7ZHYQP1q2SH5jMyIzJAQ60QCg72wa
0CzYJYi1QvcvMSn4wlA/mms=
=JkJU
-----END PGP SIGNATURE-----

[Xcircuit-dev] bug #565180: in 2.5.4, 2.5.5, copying two pins -> freeze

[Bryce D. <br...@tl...>]

I've added another source forge bug

bug #565180: ] 2.5.4, 2.5.5: copying two pins -> freeze
http://sourceforge.net/tracker/index.php?func=detail&aid=565180&group_id=4952&atid=104952

Brief summary:
                      
Starting with version 2.5.4, xcircuit freezes when you copy more than one
pin at a time inside a module.  Versions 2.5.3 and 2.4.0 do not have the
problem.

To reproduce this, start up version 2.5.4 or later.  Press 'L' for library
screen, click on ic_templates, click on any module and place a copy on
page 1. Then move the mouse over it and press '>' to descend into the
module.

Now select two or more pins using the middle button.  (Either by dragging
a box or by clicking on them.)  Press 'c' to start a copy.  Drag them to
another location and click the left button to place them.  As soon as you
click, xcircuit freezes and no longer responds, redraws the screen, etc.
Escape and right button have no effect. The only way to stop the operation
is a Control-C to kill the process.

Copying one pin at a time is ok. Copying two non-pin objects is ok.
Copying one pin and any number of non-pin objects is ok. But if your
selection includes 2 or more pins, the copy will crash the program.

For more details, see the source forge bug report.
Has anyone else seen this?

Thanks,
Bryce

[Xcircuit-dev] backspace

[R. T. E. <ti...@st...>]

Dear Zhengdao,

   Thanks for the update.  Indeed, all the problems with input focus
seem to be with "fvwm2".  I have not heard of any other window manager
having the same problem.  Maybe I'll look into the fvwm source and see
if I can find the error there.

> I noticed that you have removed the "insert backspace"
> menu item in xcircuit 5.3. The item is sometimes
> useful, e.g., when one wants to have both superscript
> and subscript (which can be done now only through
> kerning).

No, I got rid of it because I replaced it with something that works
much better---try using "Tab Stop" to set a tab point, and
"Tab backward" to jump back to it.  For instance,

X <Tab Stop> <Subscript> z <Normalscript> <Tab Backward> <Superscript> y

gives you the subscript + superscript.  Key sequence is:

X <Alt-Tab> <Keypad-Minus> z <Keypad-Enter> <Shift-Tab> <Keypad-Plus> y

The tabbing functions are more cleanly coded, give cleaner output, are
valid under all conditions (the "backspace" function would screw up under
certain conditions), and also allow the implementation of indented lists
and so forth.  If this breaks some of your drawings when trying to reload
xcircuit ("Warning:  Obsolete backspace command ommitted in text"), let
me know and I'll hurry up and implement the few lines which should be
all that's necessary to translate the former backspace command into a
tab stop + tab backward combination (files.c, comment at line 1540).

					Regards,
					Tim

[Xcircuit-dev] XCircuit on solaris crashes when iconified

[R. T. E. <ti...@st...>]

Dear Zhengdao,

When the xcircuit is iconized, it crashes with an
error message like the following:

X Error of failed request:  BadMatch (invalid parameter attributes)
  Major opcode of failed request:  42 (X_SetInputFocus)

I've heard of this problem before on some window managers under Linux,
but not on Solaris.  Technically, it's a window manager error, because
it should not be sending InputFocus events to an unmapped window.
Unfortunately, adding XtIsRealized() to the routine doesn't make any
difference.  You can't get input focus in the main window by clicking
on the title bar?
   It may be necessary to add an event handler to notify the program
when the window has been iconified, instead of relying on XtIsRealized.
I'll see what I can do about it.
						
					Regards,
					Tim

[Xcircuit-dev] X_SetInputFocus

[Zhengdao W. <zhe...@ec...>]

Hi there,

I am using the latest xcircuit version 2.5.5 and
encourter the following problem on sun solaris 2.8
system. 

When the xcircuit is iconized, it crashes with an
error message like the following:

X Error of failed request:  BadMatch (invalid parameter attributes)
  Major opcode of failed request:  42 (X_SetInputFocus)
  Serial number of failed request:  10200
  Current serial number in output stream:  10200

I tried to undefine the variable INPUT_FOCUS, but then
I do not have the focus on the xcircuit window, which 
also means that I cannot input text into it.

Any insights or solutions?

Thanks!
Zhengdao Wang

Re: [Xcircuit-dev] xcircuit to pcb flow

[Eric Y. C. <er...@al...>]

Hi Tim.  Thank you very much for the correction.  I must have missed this
going through the tutorial and just thought that ? was a special character.
I will go back and give it another try.  If this works, it should pretty
much complete the flow.

Thanks,
Eric

[Xcircuit-dev] xcircuit to pcb flow

[R. T. E. <ti...@st...>]

Dear Eric,

   I think the problem is that you missed the instructions down at
the bottom of tutorial step 14:

  "After creating the label, use the second mouse button to drag a select
   box over the question mark. Only the question mark should be
   highlighted. Then select menu item Text->Parameterize. As described
   earlier in the tutorial, the PCB netlister will use this parameterized
   string to determine a part number for the diode, or else the part
   number can be explicitly declared by editing the info label from each
   of the four instances of symbol "diode" that we will generate."

A question mark is just a question mark.  A *parameterized* question
mark is a part number.  It has to be a parameter of the symbol, or
xcircuit can't supply a unique value for each instance of the object.

So, the info label for the crystal should look like

		"pcb:X?"

but when editing, the spelled-out version of the label in the message
box should look like:

		"Font=Helvetica p c b : X Parameter(1)< ? >"

I did this, stuck it into a simple circuit, and (with pin names changed
from x.1, x.2 to simple numbers 1 and 2) I get a PCB output with
"X1-1" and "X1-2".  The difference shows up in the output file with
the original having the definition for object "crystal" like this:

/crystal {
begingate
 .
 .
 .

and the new one having this:

/crystal {
(?) 1 beginparm
 .
 . 
 .

Hopefully this clears up the problems.

					Regards,
					Tim

[Xcircuit-dev] converting generic lib to pcb aware

[Eric Y. C. <er...@al...>]

Hi.  I am almost done with setting up the XCircuit to pcb flow.  The flow
on standard parts works (but there still are some bugs, for example, rat
lines in pcb will not refresh unless the program is restarted).  Custom
parts are now working, using some suggestions from the netlisting
tutorials, but it is still a bit inconvenient managing the libraries.
The remaining problem lies in generic parts that one is attempting to
elevate to annotated status by adding info pins.  It does not work.
Adding a single info pin as recommended in Task 14 of the tutorial
does not work, and adding multiple info pins as found by editing parts
in analoglib2.lps does not work.

The problem is that the "?" is propagated into the pcb netlist, where
pcb is unable to resolve it.  This occurs, for example, when one takes
the crystal symbol ( -|[]|- , sort of), places an instance of it, and
either edits the instance, or runs "m" on the instance to generate an
object in the user library.  If diagnostic printout is performed in
writepcb() (in netlist.c), within parseinfo(), the place where the
info pin is read may be checked.  If the crystal is call X or xt, the
token with a ? appears in the Page*.pcb file, without the ? being
replaced by an incremented device index.  This replacement works for
R and C.  It was found that entries in the postscript XCircuit file
(schematic) contained the spurious "?" only for the new object, not
for R and C, so the question mark was supposed to have been trimmed
out before this time.  That means that writepcb(), which is excuted
much later does not have the bug.  The bug is upstream.  The code is
difficult to navigate.  Any ideas where the bug could be?

Note that since the custom flow works, it should be possible to build
the crystal object from scratch as was done above with ASCII art.  Then,
the suggestions from task 14 may be applied to add info pins so that
pcb can use the netlisted output.  This does work, but it is preferable
to be able to transfer the symbols without having to redraw all of them.

Any suggestions?

Thanks, Eric

Re: [Xcircuit-dev] xcircuit, PCB, and hierarchy

[Bob P. <bpa...@cs...>]

>My feeling is that the problem lies in the LVS tools.  It should not
>be necessary to ensure a perfect correspondence between part numbers;
>if two parts are 74HCT244s, then it should not make a difference that
>the one on the PCB is labeled #1 and the one on the schematic is
>labeled #2. 

For those here that might not know I work for a electronic contracting shop,
http://www.matric.com/ .  I get to deal with things like "#2 on schematic in upper left corner" is really "#49 in the middle of the PCB",
on a daily basis for a living.

IT DRIVES YOU INSANE!  Especially in the repair department, dealing with that kind of thing if it comes back for
service or upgrades.  That insanity translates in to lost cash$ as the end result.

Don't go there.  #2 on the schematic MUST be #2 on the PCB.


I missed the original message(s) on hierarchy so not sure how we got there but I've seen it handled this way in other programs, or 
I've missed the whole point of whats being discussed here (?):

You draw a sub schematic, or object, and label things in side of it such as "D?".  Then you place several of those objects on a higher
level page.  At some point you hit a button that says "Annotate" and all of the "?" are replaced with a unique sequential number.  There is also a
button that sets every thing back to "?" if you do a major revision.  You will only ever have one D7 as the end result.

In my mind that removes the issue of passing things through the layers, as there are only two levels.  The level that we humans see in the
schematic (object:designator), and the output netlist (designator:net).  A part will be only one of two value "X?" or "Xnumber" at any time. 
 If some one mannually makes two D7's when editing it should be flaged as an error.  What XCircuit does inside of it self to deal with connectivity
should be internal where we users don't care now it gets done.

[Xcircuit-dev] xcircuit, PCB, and hierarchy

[R. T. E. <ti...@st...>]

Dear Eric,

> I do agree that hierarchical
> PCB layout processes are important, from the standpoint of simplicity.
> But, on the other hand, hierarchical layout is often avoided unless it
> is imposed by machine constraints.

   XCircuit has no problem flattening designs;  it's really an issue of
naming conventions.  What xcircuit ends up with (and "magic", and other
hierarchically-based layout tools) is a flat representation with
hierarchical names such as rectifier1/diode3.1 and so forth.  To give
that diode a unique name from the top level, so that the diode is just
"D7", say, is a great idea, but xcircuit can't handle passing parameters
through several levels of hierarchy.  I come back to this issue every
now and then, as it is an extraordinarily useful concept that is
extraordinarily difficult to implement.

   On top of that, the tutorials, and probably the xcircuit interface
overall, have a clear VLSI bias, because that's what I do.  Occasionally
I build PCBs but I have yet to get into the habit of doing LVS on them.
Usually, that's not especially useful, as most of my errors stem from
misreading datasheets, where LVS won't do any good anyway.  It's only
when you get into 4+ layers on a PCB, or when you get into big
production runs, that you get into situations where it becomes
prohibitively expensive to fix errors on the board.

> I redid the tutorial without the hierarchical
> grouping, and it worked just fine.

The reason I haven't worked too much on the hierarchy-to-PCB concept
is that it works just fine if you start with a flat design in
xcircuit.  I should probably separate out the VLSI and PCB parts of
the tutorial so that I can properly stress the use of hierarchy for
VLSI but de-stress it for PCB layout.

> Probably, pcb's are not all that hierarchical because if they were,
> it might be less expensive to do the arrays in VLSI, gals or FPGA's.
> If digital logic is used, it might just be easier to use one of
> Lattices new ISP's.  They even have a downloadable tool chain.

Funny you should mention that.  My "counterexample", the 34-channel
biosonar frontend board, has more Lattice ISP chips on it than
probably any other board ever built (104, to be exact. . . these are
the ispPAC programmable analog chips, which is why there are so many
of them).  I made a subcell of 3 ispPAC chips, capacitors, and wiring,
and copied it 34 times.  In PCB, I had to be absolutely sure that
subcell was correct before copying it, because further changes would
have to be done to each part individually instead of once, to the
subcell.

Lattice's downloadable tool chain has been an annoyance from the start.
For the analog chips, I wrote my own software so I could program all
102 chips at once, feeding them high-level parameters such as filter
frequency and Q.  I used one of their digital chips, the 6192FF, as an
interface chip, but I can only compile the schematic on their Windows-
only tool.  I do the schematics in Windows, too.  It nearly drove me
to write an EDIF output style for xcircuit, but I just didn't have
the time.

Generally speaking, though, most people use FPGAs to implement great
gobs of logic like state machines and glue logic, which isn't
hierarchical at all.  The benefit of the FPGA is that 1) you can fix
it if it's wrong, and 2) you know all the timing delays for simulation
without having to lay out the board.

> Somehow, it seems that there should be a way to insure that the
> parts correspond between the two tools.

My feeling is that the problem lies in the LVS tools.  It should not
be necessary to ensure a perfect correspondence between part numbers;
if two parts are 74HCT244s, then it should not make a difference that
the one on the PCB is labeled #1 and the one on the schematic is
labeled #2.  Or for that matter, that the one on the schematic has
a hierarchical name and the one on the PCB has a flat name.  From
this standpoint, I'd like to get into "pcb" sometime and fix the
way it handles netlists.

I've also been thinking about writing an inherently hierarchical
LVS tool.  For highly arrayed designs, it would be orders of magnitude
faster, take less memory, and produce more meaningful results.  Just
another thing I'll get around to if I have the time.

						Regards,
						Tim

[Xcircuit-dev] netlist to pcb layout flow

[Eric Y. C. <er...@al...>]

Hi.  I saw an e-mail in the archive dated around 10/2001 discussing the
netlist to pcb layout flow.  It discussed pin numbering/naming, and some
hacks that are required.  I am trying to work through the tutorial for
xcircuit with the bridge rectifier.  Recall that the bridge rectifier
exercise defines a new user library element for the four diode
structure.  In the netlist, it looks like the top level instance name
followed by a slash, then the diode name, then a dash and a pin number.
For pcb, it would be most appropriate to construct an element which
represents a bridge rectifier that one could buy at an electronics
store (i.e. a square with 4 leads and a hole in the middle, maybe).
If one does this, however, it would seem that its single instance
should have a unique name (assigned in pcb) rather than the multitude
of names in the input netlist (from xcircuit).  Of course, one could
always hand edit the .pcb input netlist, but is there a more elegant
way to do this?  Or is there something that I'm not seeing?

I was wondering if there is a tutorial or "HOW-TO" for xcircuit-pcb
integration.  I searched google and did not find much of anything,
but there was that suggestive and helpful e-mail in the archive.
Anything else?

Thanks,
Eric

[Xcircuit-dev] Re: xcircuit colormap err

[Eric Y. C. <er...@al...>]

Hi Tim.  Do you need any help or investigation on this matter?  Of
course, my system does not have the LookupColor bug, but I can
easily retrigger the limited color table size problem.  How does
removing the line help the case in which the pixel value that 
comes back from LookupColor is bogus?  I first thought that the
color table generation routine had problems, but it is in fact what
saved the day when the colors were limited.

Eric

>Dear Eric,
>
>> The second thing that worked is to just revert the Cvt routine.
>
>Now I'll have to find a way to resolve the differences between the
>old and the new versions of Cvt...
>
>> Obviously, the line was taken out for a reason, but I cannot figure out
>> why.  If there is a non-zero return code, .pixel should still be set, or
>> else who knows what its value will be.
>
>There is at least one system out there where LookupColor returns a
>completely bogus value in .pixel.  It looks to me like a server error.
>Now I'm stuck with the problem that getting around the sever error
>breaks the code on systems with properly-working X servers.  I may
>be reduced to adding it as an ugly compile-time switch.
>
>Thanks for checking out the bug fix suggestions and reporting back.
>
>					Regards,
>					Tim