ngspice-devel — Ngspice development list (developers are here).

Re: [Ngspice-devel] Fix for Numparam under cygwin

[Hitoshi T. <HDA...@ni...>]

Dear all,

I have tested the function of 'NUMPARAM' and I have got some errors.
Following examples have problems.

(1) The X line is before '.subckt' line.
 In this case, 'X1 1 2 3 4 mos' has been transformed to 'X1'.
Ngspice outputs "error: unknown subckt: x1".

numparam test
.param wx=10u
.param wy=3
**
X1 1 2 3 4 mos
**
vdd 1 3 1
vss 3 0 0
vgg 2 0 1
vbb 4 0 0
**
.subckt mos 1 2 3 4
m1 1 2 3 4 nch w={wx*wy} l=0.5u
+ ad={wx*wy*1u} as={wx*wy*1u} 
+ pd={(wx*wy+1u)*2} ps={(wx*wy+1u)*2} 
+ nrd={0.01u/wx/wy} nrs={0.01u/wx/wy}
.ends mos
.dc vdd 0 3 0.01 vgg 0 3 0.5
.model nch nmos level=14 vtho=0.5 
.end

(2) '.subckt' line has '.params:'. 
 This causes the program freeze.

numparam test
.param wx=10u
vdd 1 3 1
vss 3 0 0
vgg 2 0 1
vbb 4 0 0
**
.subckt mos 1 2 3 4 params:wy=3
m1 1 2 3 4 nch w={wx*wy} l=0.5u
+ ad={wx*wy*1u} as={wx*wy*1u} 
+ pd={(wx*wy+1u)*2} ps={(wx*wy+1u)*2} 
+ nrd={0.01u/wx/wy} nrs={0.01u/wx/wy}
.ends mos
**
X1 1 2 3 4 mos
**
.dc vdd 0 3 0.01 vgg 0 3 0.5
.model nch nmos level=14 vtho=0.5 
.end

The following example is able to analyze normally. 

numparam test
.param wx=10u
.param wy=3
vdd 1 3 1
vss 3 0 0
vgg 2 0 1
vbb 4 0 0
**
.subckt mos 1 2 3 4
m1 1 2 3 4 nch w={wx*wy} l=0.5u
+ ad={wx*wy*1u} as={wx*wy*1u} 
+ pd={(wx*wy+1u)*2} ps={(wx*wy+1u)*2} 
+ nrd={0.01u/wx/wy} nrs={0.01u/wx/wy}
.ends mos
**
X1 1 2 3 4 mos
**
.dc vdd 0 3 0.01 vgg 0 3 0.5
.model nch nmos level=14 vtho=0.5 
.end

Regards,
Hitoshi Tanaka

Re: [Ngspice-devel] Fix for Numparam under cygwin

[Steven B. <ste...@di...>]

Hello,

I have some additional fixes and improvements to the front-end and to 
numparams which I am just testing.  Does  anyone have comments on these 
proposed changes?

The changes are;

1. Addition of end-of-line comments
The numparams module has some code for this but there are some problems 
with it. Instead I've added code to the main front end that gives 
better performance IMHO, and allows eol comments on continued lines 
(ie. lines where the next line starts with a '+'.
I have supported the following formats:
	;	for compatibility with PSpice
	$	for compatibility with HSpice
	//	like C++, and as per numparams
	--	as per numparams
Are any other formats needed? I don't know how standard // and -- are. 
I only put them in because numparams did.

2. An increase in the line buffer size in numparams to 500 (was 250)
This buffer hold the circuit line with any continuation line appended. 
I found that it was all to easy to exceed the original 250 character 
limit with .model lines. Especially if they have lots of white space on 
the continuation lines (to make them look pretty in the circuit file).  
This change increases the buffer to 500. End-of-line comments to don't 
count in this limit as they are stripped.  This hard limit is still 
ugly, but would need a significant overhaul to change that.

Regards,
Steven


On 21 Apr 2005, at 18:17, Hitoshi Tanaka wrote:

> Dear Steven,
>
> I applied your patches for Cygwin and I built it successfully.
> Thank you.
> Subsequently, I am going to do the test of the circuit including param 
> and params.
>
> Regards,
> Hitoshi Tanaka
>
> ----- Original Message ----- From: "steven.borley" 
> <ste...@vi...>
> To: "ngspice-devel" <ngs...@li...>
> Sent: Thursday, April 21, 2005 6:41 AM
> Subject: [Ngspice-devel] Fix for Numparam under cygwin
>
>
>> To compile with numparam enabled under Cygwin the attached patch is
>> required.
>>
>> This is due to feof(p) being a maco that expects p to be a pointer to 
>> a
>> FILE structure, but numparam is passing in a void *.  There is a good
>> chance that this fix could be required in other environments too.
>>
>> Change also tested on MacOSX and is good (although it didn't need it).
>>
>> Regards,
>> Steven
>>
>
>
>
>
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real 
> users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> Ngspice-devel mailing list
> Ngs...@li...
> https://lists.sourceforge.net/lists/listinfo/ngspice-devel
>

Re: [Ngspice-devel] Fix for Numparam under cygwin

[Hitoshi T. <HDA...@ni...>]

Dear Steven,

I applied your patches for Cygwin and I built it successfully.
Thank you.
Subsequently, I am going to do the test of the circuit including param and 
params.

Regards,
Hitoshi Tanaka

----- Original Message ----- 
From: "steven.borley" <ste...@vi...>
To: "ngspice-devel" <ngs...@li...>
Sent: Thursday, April 21, 2005 6:41 AM
Subject: [Ngspice-devel] Fix for Numparam under cygwin


> To compile with numparam enabled under Cygwin the attached patch is
> required.
>
> This is due to feof(p) being a maco that expects p to be a pointer to a
> FILE structure, but numparam is passing in a void *.  There is a good
> chance that this fix could be required in other environments too.
>
> Change also tested on MacOSX and is good (although it didn't need it).
>
> Regards,
> Steven
>
> 

[Ngspice-devel] Fix for Numparam under cygwin

[steven.borley <ste...@vi...>]

To compile with numparam enabled under Cygwin the attached patch is 
required.

This is due to feof(p) being a maco that expects p to be a pointer to a 
FILE structure, but numparam is passing in a void *.  There is a good 
chance that this fix could be required in other environments too.

Change also tested on MacOSX and is good (although it didn't need it).

Regards,
Steven

Re: [Ngspice-devel] Cygwin and MinGW status

[Steven B. <ste...@di...>]

Hello,

With help from Hitoshi Tanaka I have been able to build and run NGspice 
on a WindowsXP laptop using Cygwin.

  It required one fix. Patch is attached. Rather than modify a system 
file I have added a small work-round in src/frontend/resource.c to 
avoid a clash with definitions of BOOLEAN. I also included a fix to 
remove warnings of unused variables.

The patch should be safe on all other operating systems and I have 
already checked that it does not affect MacOSX compiles.

The build after a clean download from CVS was just this...

./autogen.sh
./configure --with-readline=yes
make
make install

There is a build issue with using the --enable-xspice flag (which I'll 
have a look at shortly).

A screen dump of this running on a WindowsXP using Cygwin's X-Server 
can be found here:
http://www.diode.demon.co.uk/ngspice/index.html

That page also has screen dump of it running on my Mac under MacOSX, 
for comparison.

Regards,
Steven

On 20 Apr 2005, at 01:18, Hitoshi Tanaka wrote:

> Dear Steven,
>
> Sorry, I forgot that I comment-outed the line.
> typedef BYTE BOOLEAN, *PBOOLEAN;
> -->/*typedef BYTE BOOLEAN, *PBOOLEAN;*/
>
> Regards,
> Hitoshi Tanaka
>
> ----- Original Message ----- From: "Steven Borley" 
> <ste...@di...>
> To: "ngspice-devel" <ngs...@li...>
> Sent: Wednesday, April 20, 2005 6:19 AM
> Subject: Re: [Ngspice-devel] Cygwin and MinGW status
>

Re: [Ngspice-devel] --enable-numparam compile error

[steven.borley <ste...@vi...>]

On 19 Apr 2005, at 23:17, Greg Lauer wrote:

> Steven,
>
> Yes, looks like two files got updated on the same day.
> I'm still using the 15c source and just updating those few files.
>
> I did get numparam to compile without updating those extra files,

I think you'll find that although it will compile without those other 
files, it will not work properly.

The subckt.c-diff patch file fixes/adds the following;
* Added some debug tracing (normally off)
* Fixes a #ifdef that had a typo and thus stopped numparams working (**)

The inpcom.c-diff patch file fixes/adds the following:
* Renames a local variable to avoid confusion (of the 
developer/maintainer)
* Fixes the renumbering of lines from a .include file (**)

The changes flagged with (**) above are the important ones.

The inpcom.c fix means that include file now work with numparam.

> so I think I'm in good shape. I'll go back and upate those other two as
> well.
>
> If I wanted to update ALL of the 15c files with what's in CVS, what's
> the best way to do this?

Are are some instructions on obtaining all the files from CVS here:
http://sourceforge.net/cvs/?group_id=38962

For update etc I tend to use LinCVS. You could give that a whirl if you 
what a GUI front-end to CVS.

Regards,
Steven

Re: [Ngspice-devel] --enable-numparam compile error

[Greg L. <gl...@ls...>]

Steven,

Yes, looks like two files got updated on the same day.
I'm still using the 15c source and just updating those few files.

I did get numparam to compile without updating those extra files,
so I think I'm in good shape. I'll go back and upate those other two as
well.

If I wanted to update ALL of the 15c files with what's in CVS, what's
the best way to do this?

Thanks,
Greg

steven.borley wrote:

> On 19 Apr 2005, at 22:28, Greg Lauer wrote:
>
>> Hi Steven,
>>
>> From CVS, I see that the latest files are:
>>
>> 5176 Apr 16 15:45 general.h
>> 16873 Apr 16 15:45 mystring.c
>> 28555 Apr 16 15:45 readme.txt
>> 16078 Apr 16 15:45 spicenum.c
>> 37481 Apr 16 15:45 xpressn.c
>>
>> Are there any other files that might have been updated for numparam?
>
>
> From memory you will also need...
>     src/frontend/subckt.c
>     src/frontend/inpcom.c
>
> Paolo is very thorough in keeping the ChangeLog file up-to-date. You 
> might like to check in there too, in case I've forgotten something.
>
>> I'll update these and give it a whirl.
>
>
> Good luck.
>
> Regards,
> Steven
>

Re: [Ngspice-devel] --enable-numparam compile error

[steven.borley <ste...@vi...>]

On 19 Apr 2005, at 22:28, Greg Lauer wrote:

> Hi Steven,
>
> From CVS, I see that the latest files are:
>
> 5176 Apr 16 15:45 general.h
> 16873 Apr 16 15:45 mystring.c
> 28555 Apr 16 15:45 readme.txt
> 16078 Apr 16 15:45 spicenum.c
> 37481 Apr 16 15:45 xpressn.c
>
> Are there any other files that might have been updated for numparam?

 From memory you will also need...
	src/frontend/subckt.c
	src/frontend/inpcom.c

Paolo is very thorough in keeping the ChangeLog file up-to-date. You 
might like to check in there too, in case I've forgotten something.

> I'll update these and give it a whirl.

Good luck.

Regards,
Steven

Re: [Ngspice-devel] --enable-numparam compile error

[Greg L. <gl...@ls...>]

Hi Steven,

 From CVS, I see that the latest files are:

5176 Apr 16 15:45 general.h
16873 Apr 16 15:45 mystring.c
28555 Apr 16 15:45 readme.txt
16078 Apr 16 15:45 spicenum.c
37481 Apr 16 15:45 xpressn.c

Are there any other files that might have been updated for numparam?
I'll update these and give it a whirl.

Thanks!
Greg


steven.borley wrote:

> Greg,
>
> I fixed this a few days ago.  It should be available from CVS now 
> (takes up to 5 hours from from when Paolo commits it for it to be 
> accessible).
>
> Basically I change the names of these two functions to avoid the 
> clash. However, there were other things needed to get numparam to work 
> correctly, not just this, so I suggest you update from CVS when you can.
>
> Regards,
> Steven
>
>
> On 19 Apr 2005, at 00:16, Greg Lauer wrote:
>
>> I'm trying to compile ngspice with --enable-numparam
>> I get the following make error. Looks like it doesn't like
>> a mac osx (powerpc) include file. Anyone have a clue how to get around
>> this?
>>
>> In 'general.h', the lines it complains about look like:
>> Func long round(double d);
>> Func long trunc(double x);
>>
>> In 'math.h', the line look like:
>> extern double round( double );
>> extern double trunc( double );
>>
>> Thanks,
>> Greg
>>
>> Making all in numparam
>> if gcc -DHAVE_CONFIG_H -I. -I. -I../../.. -I../../../src/include  
>> -I/sw/include  -g -O2 -Wall  -MT spicenum.o -MD -MP
>> -MF ".deps/spicenum.Tpo" -c -o spicenum.o spicenum.c; \
>> then mv -f ".deps/spicenum.Tpo" ".deps/spicenum.Po"; else rm -f 
>> ".deps/spicenum.Tpo"; exit 1; fi
>> spicenum.c: In function `nupa_done':
>> spicenum.c:425: warning: implicit declaration of function `exit'
>> if gcc -DHAVE_CONFIG_H -I. -I. -I../../.. -I../../../src/include  
>> -I/sw/include  -g -O2 -Wall  -MT xpressn.o -MD -MP
>> -MF ".deps/xpressn.Tpo" -c -o xpressn.o xpressn.c; \
>> then mv -f ".deps/xpressn.Tpo" ".deps/xpressn.Po"; else rm -f 
>> ".deps/xpressn.Tpo"; exit 1; fi
>> In file included from xpressn.c:10:
>> general.h:198: error: conflicting types for `round'
>> /usr/include/architecture/ppc/math.h:336: error: previous declaration 
>> of `round'
>> general.h:199: error: conflicting types for `trunc'
>> /usr/include/architecture/ppc/math.h:345: error: previous declaration 
>> of `trunc'
>> make[4]: *** [xpressn.o] Error 1
>> make[3]: *** [all-recursive] Error 1
>> make[2]: *** [all-recursive] Error 1
>> make[1]: *** [all-recursive] Error 1
>> make: *** [all] Error 2
>>
>>
>>
>>
>> -------------------------------------------------------
>> This SF.Net email is sponsored by: New Crystal Reports XI.
>> Version 11 adds new functionality designed to reduce time involved in
>> creating, integrating, and deploying reporting solutions. Free 
>> runtime info,
>> new features, or free trial, at: 
>> http://www.businessobjects.com/devxi/728
>> _______________________________________________
>> Ngspice-devel mailing list
>> Ngs...@li...
>> https://lists.sourceforge.net/lists/listinfo/ngspice-devel
>>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: New Crystal Reports XI.
> Version 11 adds new functionality designed to reduce time involved in
> creating, integrating, and deploying reporting solutions. Free runtime 
> info,
> new features, or free trial, at: http://www.businessobjects.com/devxi/728
> _______________________________________________
> Ngspice-devel mailing list
> Ngs...@li...
> https://lists.sourceforge.net/lists/listinfo/ngspice-devel
>

Re: [Ngspice-devel] Cygwin and MinGW status

[Steven B. <ste...@di...>]

Tanaka-san,

Thank you your help.

I am unable to try with Visual C++, but from within cygwin I find I 
cannot compile.

After a clean download of all file from CVS I tried again...

./autogen.sh
./configure --enable-experimental
make

This fail while trying to compile resource.c
/usr/include/w32api/winnt.h:157: error: parse error before ',' token

line 157 of winnt.h is a typedef...
typedef BYTE BOOLEAN, *PBOOLEAN;


Has anyone else tried this?

Regards,
Steven


On 19 Apr 2005, at 20:12, Hitoshi Tanaka wrote:

> Dear Steven,
>
> I am sorry.  The previous comment was concerned with "building on 
> Xwindow".
> I have not been succeeding in the method in "INSTALL".
> But I succeeded in building with Visual C++(without xspice).
>
> Regards,
> Hitoshi Tanaka
>
> ----- Original Message ----- From: "Hitoshi Tanaka" 
> <HDA...@ni...>
> To: <ngs...@li...>
> Sent: Wednesday, April 20, 2005 3:40 AM
> Subject: Re: [Ngspice-devel] Cygwin and MinGW status
>
>
>> Dear Steven,
>>
>> I succeeded in building on Cygwin.
>> The way is very simple.
>> ./configure (--enable-exparimantal)
>> make
>> make install
>>
>> However, it did not succeed with --enable-xspice option.
>> By modifying a part of cfunc.mod files, it succeeded.
>> I don't understand the cause well.
>>
>> Regards,
>> Hitoshi Tanaka
>>
>> ----- Original Message ----- From: "Steven Borley" 
>> <ste...@di...>
>> To: <ngs...@li...>
>> Sent: Wednesday, April 20, 2005 1:55 AM
>> Subject: [Ngspice-devel] Cygwin and MinGW status
>>
>>
>>> Hello,
>>>
>>> Can anyone tell me the status of the MinGW and the Cygwin building.
>>>
>>> I'd like to transfer some of my work to my laptop which is running 
>>> Windows XP. Up to now I've work just on my Mac.
>>>
>>> I tried using the notes in INSTALL for building from within Cygwin 
>>> (both methods), but neither worked. I had errors about missing 
>>> ncurses or termcap libraries.
>>>
>>> I also tried from within MSYS + MinGW....(from memory here, so 
>>> please ignore any typos...)
>>>
>>> ./autogen.sh
>>> ./configure --with-windows --without-x --with-readline=no
>>>
>>>  but this failed looking for the ncurses or terncap libraries again.
>>>
>>> Does anyone build in these environments and know what needs to be 
>>> done?
>>>
>>> Regards,
>>> Steven
>>>
>>>
>>>
>>> -------------------------------------------------------
>>> This SF.Net email is sponsored by: New Crystal Reports XI.
>>> Version 11 adds new functionality designed to reduce time involved in
>>> creating, integrating, and deploying reporting solutions. Free 
>>> runtime info,
>>> new features, or free trial, at: 
>>> http://www.businessobjects.com/devxi/728
>>> _______________________________________________
>>> Ngspice-devel mailing list
>>> Ngs...@li...
>>> https://lists.sourceforge.net/lists/listinfo/ngspice-devel
>>>
>>
>>
>>
>>
>> -------------------------------------------------------
>> This SF.Net email is sponsored by: New Crystal Reports XI.
>> Version 11 adds new functionality designed to reduce time involved in
>> creating, integrating, and deploying reporting solutions. Free 
>> runtime info,
>> new features, or free trial, at: 
>> http://www.businessobjects.com/devxi/728
>> _______________________________________________
>> Ngspice-devel mailing list
>> Ngs...@li...
>> https://lists.sourceforge.net/lists/listinfo/ngspice-devel
>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: New Crystal Reports XI.
> Version 11 adds new functionality designed to reduce time involved in
> creating, integrating, and deploying reporting solutions. Free runtime 
> info,
> new features, or free trial, at: 
> http://www.businessobjects.com/devxi/728
> _______________________________________________
> Ngspice-devel mailing list
> Ngs...@li...
> https://lists.sourceforge.net/lists/listinfo/ngspice-devel
>

Re: [Ngspice-devel] Cygwin and MinGW status

[Hitoshi T. <HDA...@ni...>]

Dear Steven,

I am sorry.  The previous comment was concerned with "building on Xwindow".
I have not been succeeding in the method in "INSTALL".
But I succeeded in building with Visual C++(without xspice).

Regards,
Hitoshi Tanaka

----- Original Message ----- 
From: "Hitoshi Tanaka" <HDA...@ni...>
To: <ngs...@li...>
Sent: Wednesday, April 20, 2005 3:40 AM
Subject: Re: [Ngspice-devel] Cygwin and MinGW status


> Dear Steven,
>
> I succeeded in building on Cygwin.
> The way is very simple.
> ./configure (--enable-exparimantal)
> make
> make install
>
> However, it did not succeed with --enable-xspice option.
> By modifying a part of cfunc.mod files, it succeeded.
> I don't understand the cause well.
>
> Regards,
> Hitoshi Tanaka
>
> ----- Original Message ----- 
> From: "Steven Borley" <ste...@di...>
> To: <ngs...@li...>
> Sent: Wednesday, April 20, 2005 1:55 AM
> Subject: [Ngspice-devel] Cygwin and MinGW status
>
>
>> Hello,
>>
>> Can anyone tell me the status of the MinGW and the Cygwin building.
>>
>> I'd like to transfer some of my work to my laptop which is running 
>> Windows XP. Up to now I've work just on my Mac.
>>
>> I tried using the notes in INSTALL for building from within Cygwin (both 
>> methods), but neither worked. I had errors about missing ncurses or 
>> termcap libraries.
>>
>> I also tried from within MSYS + MinGW....(from memory here, so please 
>> ignore any typos...)
>>
>> ./autogen.sh
>> ./configure --with-windows --without-x --with-readline=no
>>
>>  but this failed looking for the ncurses or terncap libraries again.
>>
>> Does anyone build in these environments and know what needs to be done?
>>
>> Regards,
>> Steven
>>
>>
>>
>> -------------------------------------------------------
>> This SF.Net email is sponsored by: New Crystal Reports XI.
>> Version 11 adds new functionality designed to reduce time involved in
>> creating, integrating, and deploying reporting solutions. Free runtime 
>> info,
>> new features, or free trial, at: http://www.businessobjects.com/devxi/728
>> _______________________________________________
>> Ngspice-devel mailing list
>> Ngs...@li...
>> https://lists.sourceforge.net/lists/listinfo/ngspice-devel
>>
>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: New Crystal Reports XI.
> Version 11 adds new functionality designed to reduce time involved in
> creating, integrating, and deploying reporting solutions. Free runtime 
> info,
> new features, or free trial, at: http://www.businessobjects.com/devxi/728
> _______________________________________________
> Ngspice-devel mailing list
> Ngs...@li...
> https://lists.sourceforge.net/lists/listinfo/ngspice-devel
> 

Re: [Ngspice-devel] Cygwin and MinGW status

[Hitoshi T. <HDA...@ni...>]

Dear Steven,

I succeeded in building on Cygwin.
The way is very simple.
./configure (--enable-exparimantal)
make
make install

However, it did not succeed with --enable-xspice option.
By modifying a part of cfunc.mod files, it succeeded.
I don't understand the cause well.

Regards,
Hitoshi Tanaka

----- Original Message ----- 
From: "Steven Borley" <ste...@di...>
To: <ngs...@li...>
Sent: Wednesday, April 20, 2005 1:55 AM
Subject: [Ngspice-devel] Cygwin and MinGW status


> Hello,
>
> Can anyone tell me the status of the MinGW and the Cygwin building.
>
> I'd like to transfer some of my work to my laptop which is running Windows 
> XP. Up to now I've work just on my Mac.
>
> I tried using the notes in INSTALL for building from within Cygwin (both 
> methods), but neither worked. I had errors about missing ncurses or 
> termcap libraries.
>
> I also tried from within MSYS + MinGW....(from memory here, so please 
> ignore any typos...)
>
> ./autogen.sh
> ./configure --with-windows --without-x --with-readline=no
>
>  but this failed looking for the ncurses or terncap libraries again.
>
> Does anyone build in these environments and know what needs to be done?
>
> Regards,
> Steven
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: New Crystal Reports XI.
> Version 11 adds new functionality designed to reduce time involved in
> creating, integrating, and deploying reporting solutions. Free runtime 
> info,
> new features, or free trial, at: http://www.businessobjects.com/devxi/728
> _______________________________________________
> Ngspice-devel mailing list
> Ngs...@li...
> https://lists.sourceforge.net/lists/listinfo/ngspice-devel
> 

[Ngspice-devel] Cygwin and MinGW status

[Steven B. <ste...@di...>]

Hello,

Can anyone tell me the status of the MinGW and the Cygwin building.

I'd like to transfer some of my work to my laptop which is running 
Windows XP. Up to now I've work just on my Mac.

I tried using the notes in INSTALL for building from within Cygwin 
(both methods), but neither worked. I had errors about missing ncurses 
or termcap libraries.

I also tried from within MSYS + MinGW....(from memory here, so please 
ignore any typos...)

./autogen.sh
./configure --with-windows --without-x --with-readline=no

  but this failed looking for the ncurses or terncap libraries again.

Does anyone build in these environments and know what needs to be done?

Regards,
Steven

Re: [Ngspice-devel] --enable-numparam compile error

[steven.borley <ste...@vi...>]

Greg,

I fixed this a few days ago.  It should be available from CVS now 
(takes up to 5 hours from from when Paolo commits it for it to be 
accessible).

Basically I change the names of these two functions to avoid the clash. 
However, there were other things needed to get numparam to work 
correctly, not just this, so I suggest you update from CVS when you 
can.

Regards,
Steven


On 19 Apr 2005, at 00:16, Greg Lauer wrote:

> I'm trying to compile ngspice with --enable-numparam
> I get the following make error. Looks like it doesn't like
> a mac osx (powerpc) include file. Anyone have a clue how to get around
> this?
>
> In 'general.h', the lines it complains about look like:
> Func long round(double d);
> Func long trunc(double x);
>
> In 'math.h', the line look like:
> extern double round( double );
> extern double trunc( double );
>
> Thanks,
> Greg
>
> Making all in numparam
> if gcc -DHAVE_CONFIG_H -I. -I. -I../../.. -I../../../src/include  
> -I/sw/include  -g -O2 -Wall  -MT spicenum.o -MD -MP
> -MF ".deps/spicenum.Tpo" -c -o spicenum.o spicenum.c; \
> then mv -f ".deps/spicenum.Tpo" ".deps/spicenum.Po"; else rm -f 
> ".deps/spicenum.Tpo"; exit 1; fi
> spicenum.c: In function `nupa_done':
> spicenum.c:425: warning: implicit declaration of function `exit'
> if gcc -DHAVE_CONFIG_H -I. -I. -I../../.. -I../../../src/include  
> -I/sw/include  -g -O2 -Wall  -MT xpressn.o -MD -MP
> -MF ".deps/xpressn.Tpo" -c -o xpressn.o xpressn.c; \
> then mv -f ".deps/xpressn.Tpo" ".deps/xpressn.Po"; else rm -f 
> ".deps/xpressn.Tpo"; exit 1; fi
> In file included from xpressn.c:10:
> general.h:198: error: conflicting types for `round'
> /usr/include/architecture/ppc/math.h:336: error: previous declaration 
> of `round'
> general.h:199: error: conflicting types for `trunc'
> /usr/include/architecture/ppc/math.h:345: error: previous declaration 
> of `trunc'
> make[4]: *** [xpressn.o] Error 1
> make[3]: *** [all-recursive] Error 1
> make[2]: *** [all-recursive] Error 1
> make[1]: *** [all-recursive] Error 1
> make: *** [all] Error 2
>
>
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by: New Crystal Reports XI.
> Version 11 adds new functionality designed to reduce time involved in
> creating, integrating, and deploying reporting solutions. Free runtime 
> info,
> new features, or free trial, at: 
> http://www.businessobjects.com/devxi/728
> _______________________________________________
> Ngspice-devel mailing list
> Ngs...@li...
> https://lists.sourceforge.net/lists/listinfo/ngspice-devel
>

[Ngspice-devel] --enable-numparam compile error

[Greg L. <gl...@ls...>]

I'm trying to compile ngspice with --enable-numparam
I get the following make error. Looks like it doesn't like
a mac osx (powerpc) include file. Anyone have a clue how to get around
this?

In 'general.h', the lines it complains about look like:
Func long round(double d);
Func long trunc(double x);

In 'math.h', the line look like:
extern double round( double );
extern double trunc( double );

Thanks,
Greg

Making all in numparam
if gcc -DHAVE_CONFIG_H -I. -I. -I../../.. -I../../../src/include  
-I/sw/include  -g -O2 -Wall  -MT spicenum.o -MD -MP
 -MF ".deps/spicenum.Tpo" -c -o spicenum.o spicenum.c; \
then mv -f ".deps/spicenum.Tpo" ".deps/spicenum.Po"; else rm -f 
".deps/spicenum.Tpo"; exit 1; fi
spicenum.c: In function `nupa_done':
spicenum.c:425: warning: implicit declaration of function `exit'
if gcc -DHAVE_CONFIG_H -I. -I. -I../../.. -I../../../src/include  
-I/sw/include  -g -O2 -Wall  -MT xpressn.o -MD -MP
-MF ".deps/xpressn.Tpo" -c -o xpressn.o xpressn.c; \
then mv -f ".deps/xpressn.Tpo" ".deps/xpressn.Po"; else rm -f 
".deps/xpressn.Tpo"; exit 1; fi
In file included from xpressn.c:10:
general.h:198: error: conflicting types for `round'
/usr/include/architecture/ppc/math.h:336: error: previous declaration of 
`round'
general.h:199: error: conflicting types for `trunc'
/usr/include/architecture/ppc/math.h:345: error: previous declaration of 
`trunc'
make[4]: *** [xpressn.o] Error 1
make[3]: *** [all-recursive] Error 1
make[2]: *** [all-recursive] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2

Re: [Ngspice-devel] Re: [Ngspice-users] numparam and gnetlist

[Paolo N. <p....@ie...>]

On Sun, 2005-04-17 at 21:41 +0100, steven.borley wrote:

> 	+uname = $(shell uname)
> 
> should read...
> 	uname = $(shell uname)

Thanks, fixed!

Ciao,
Paolo

Re: [Ngspice-devel] ngspice compile problems on osx

[<sd...@cl...>]

> 
> 
> --Apple-Mail-6-164616444
> Content-Transfer-Encoding: 7bit
> Content-Type: text/plain;
> 	charset=US-ASCII;
> 	format=flowed
> 
> Hello,
> 
> After the fixes, now in CVS, I can still see one final problem.
> 
> in xspice/icm/makedefs.in there is this line...
> 
> mkdir_p = @mkdir_p@
> 
> The bit between the @@ should get substituted in the configure process. 
> However, in the generated makedef file the line is still there 
> intact....
> 
> mkdir_p = @mkdir_p@
> 
> It gets used during the install process.  Thus install is broken.
> 
> I have attached my generated makedefs (maybe on other systems this 
> works?).  Any suggestions or guesses as to what might be going wrong 
> here?  And is this also broken on Linux and Cygwin etc.?

This is broken in Linux also.  On my box I have just edited the
relevant lines to read:

# mkdir_p = @mkdir_p@
mkdir_p = mkdir -p

And this works for me (as well as the gEDA Suite install CD).  I have
thought about fixing it, but don't have the necessary skills with
makedep and all that.

Anybody know how this is supposed to work?  Any reason to not just
simply implement the above lines?

Stuart

[Ngspice-devel] Re: [Ngspice-users] numparam and gnetlist

[steven.borley <ste...@vi...>]

Paolo,

A spurious '+' has crept in to xspice/icm/makedefs.in during the latest 
commit.

The line...
	+uname = $(shell uname)

should read...
	uname = $(shell uname)

Hardly need a patch file for this, but I've attached one anyway :-)

Regards,
Steven


On 16 Apr 2005, at 23:48, Paolo Nenzi wrote:

> Ciao Steven,
>
> I have applied your patches and tested the code. I have added an
> examples/numparam directory containing the files yuo supplied with your
> patches.
>
> Thanks for your contribution!
>
> Ciao,
> Paolo
>
>
> PS: I have committed al changes made to the CVS, happy downloading :)
>
>
>
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real 
> users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> Ngspice-users mailing list
> Ngs...@li...
> https://lists.sourceforge.net/lists/listinfo/ngspice-users
>

Re: [Ngspice-devel] ngspice compile problems on osx

[Steven B. <ste...@di...>]

Hello,

After the fixes, now in CVS, I can still see one final problem.

in xspice/icm/makedefs.in there is this line...

mkdir_p = @mkdir_p@

The bit between the @@ should get substituted in the configure process. 
However, in the generated makedef file the line is still there 
intact....

mkdir_p = @mkdir_p@

It gets used during the install process.  Thus install is broken.

I have attached my generated makedefs (maybe on other systems this 
works?).  Any suggestions or guesses as to what might be going wrong 
here?  And is this also broken on Linux and Cygwin etc.?

Regards,
Steven

Re: [Ngspice-devel] Subckt.c and other

[Hitoshi T. <HDA...@ni...>]

Dear Paolo and all,

I have checked "greg.bug". It has following three bugs.
(1)Floating node "7_13398_8180#" exists in SUBCKT adc_nand2_big.
(2)There is not a model file "hu_dig".
(3)"/" is included to the node name. (This influences a ploting.)
If those are modified Ngspice can do an analysis correctly(?).
I will send you the corrected file "greg_bug2.cir".

> Now a question: since we are working on subckt.c, do you have an idea
> how a .global <node1> ...  directive that makes a <node1> global to all
> subcircuits (and thus no translation is necessary )can be implemented ?

The addition of the ".global" option is necessary, to use global node.
It is being defined as follows in AIM-Spice. 
General form:
.global [node name 1] < node name 2> ...
Example:
.global vdd vss

I think it can be implemented, if it avoids transforming nodes which 
are designated with ".global" at the expanding of the subckt.

Regards,
Hitoshi Tanaka

----- Original Message ----- 
From: "Paolo Nenzi" <p....@ie...>
To: <ngs...@li...>
Cc: <ngs...@li...>
Sent: Sunday, April 17, 2005 2:51 AM
Subject: [Ngspice-devel] Subckt.c and other


> Dear all,
> 
> First many thanks to Hitoshi for his patch on subckt.c (the new one has
> been not yet committed to the CVS but attached at the end of this
> message. I would like to answer Stuart's question too regarding
> netlists.
> At the following ftp address you can find a file called qxdir.tar.gz.
> That file contains some buggy netlists (netlists plain spice cannot
> compile): ftp://ic.eecs.berkeley.edu/pub/Spice3/
> 
> I think this can the starting point of ngspice fixing, and can parallel
> the fixing activity each one does using the netlists coming from his
> job/hobby.
> 
> Since we are in the process of debugging subckt.c I have attached a
> problematic netlist (from qxdir.tar.gz) called "greg.bug".
> 
> Now a question: since we are working on subckt.c, do you have an idea
> how a .global <node1> ...  directive that makes a <node1> global to all
> subcircuits (and thus no translation is necessary )can be implemented ?
> 
> The idea is to build a list of node names and compare each node in
> translation whith this list, but how this list should be implemented ? 
> 
> 
> Now on the mac porting issues...
> 
> Ciao,
> Paolo  
> 
>

[Ngspice-devel] Re: [Ngspice-users] numparam and gnetlist

[Paolo N. <p....@ie...>]

Ciao Steven,

I have applied your patches and tested the code. I have added an
examples/numparam directory containing the files yuo supplied with your
patches.

Thanks for your contribution!

Ciao,
Paolo


PS: I have committed al changes made to the CVS, happy downloading :)

Re: [Ngspice-devel] MacOS porting

[Khairulmizam S. <xo...@us...>]

Great!!!

-- 
-kmbs-



Quoting Paolo Nenzi <p....@ie...>:

> Hi ngspice users on Macs,
> 
> It seems that Mac OSX ngspice users are growing. I cannot directly test
> ngspice on a Machintosh and then I rely upon your reports. I have
> applied Steven patch to autogen.sh and verified that does not break
> anything in Linux-i86 (and it does not break anything).
> 
> I have applied patches contained in the ngspice.patch maintained by
> Khairulmizam Samsudin and thus, I think that it should be changed for
> the next release. One of the patch was not MAC specific, it fixed a
> potential problem in a malloc operation that did not accounted for the
> null terminating character of a string.
> 
> Not yet in CVS, please wait some more :)
> 
> Thanks for your support!
> 
> Ciao,
> Paolo
> 
> 
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> Ngspice-devel mailing list
> Ngs...@li...
> https://lists.sourceforge.net/lists/listinfo/ngspice-devel
> 
> 
> 


-------------------------------------------------
This mail sent through IMP: http://horde.org/imp/

[Ngspice-devel] MacOS porting

[Paolo N. <p....@ie...>]

Hi ngspice users on Macs,

It seems that Mac OSX ngspice users are growing. I cannot directly test
ngspice on a Machintosh and then I rely upon your reports. I have
applied Steven patch to autogen.sh and verified that does not break
anything in Linux-i86 (and it does not break anything).

I have applied patches contained in the ngspice.patch maintained by
Khairulmizam Samsudin and thus, I think that it should be changed for
the next release. One of the patch was not MAC specific, it fixed a
potential problem in a malloc operation that did not accounted for the
null terminating character of a string.

Not yet in CVS, please wait some more :)

Thanks for your support!

Ciao,
Paolo

Re: [Ngspice-devel] ngspice compile problems on osx

[Greg L. <gl...@ls...>]

Success! I removed an offending LDFLAG -Ddlsym=dlsym_prepend_underscore
and got it to compile! I made some additional enhancements to
point to the latest 15c build as well. Below is the new ngspice.info
file (updated version, revision, source-md5, and ldflag lines).
KMBS's .patch file is the same and works as expected.

Thanks everyone!
Greg

################# ngspice.info begin ######################
Package: ngspice
Version: 15c
Revision: 1
Description: Circuit simulation 
License: BSD
Maintainer: Khairulmizam Samsudin <xo...@us...>

# Dependencies
Depends: dlcompat-shlibs, x11, readline-shlibs
BuildDepends: dlcompat-dev, x11-dev, readline

# Unpack Phase
Source: mirror:sourceforge:%n/ng-spice-rework-%v.tar.gz
SourceDirectory: ng-spice-rework-15
Source-MD5: 55157484b65073e981ad7c7baf6d4101

# Patch Phase
Patch: %n.patch

# Compile Phase 
ConfigureParams: --x-includes=/usr/X11R6/include --x-libraries=/usr/X11R6/lib 
--with-readline=yes --mandir='$(prefix)/share/man' 
--infodir='$(prefix)/share/info' --enable-xspice
GCC: 3.3
CompileScript: <<
 #!/bin/sh -ex
 LDFLAGS="-L/sw/lib" CPPFLAGS="-I/sw/include" ./configure %c
 cd src/xspice/icm
 make clean
 cd ../../..
 cd doc
 rm *info*
 cd ..
 make
<<

# Install Phase
DocFiles: ANALYSES BUGS ChangeLog FAQ Stuarts_Poly_Notes AUTHORS COPYING DEVICES 
INSTALL NEWS README
InfoDocs: ngspice.info ngspice.info-1 ngspice.info-10 ngspice.info-2 
ngspice.info-3 ngspice.info-4 ngspice.info-5 ngspice.info-6 ngspice.info-7 
ngspice.info-8 ngspice.info-9

# Additional Data
Homepage: http://sourceforge.net/projects/ngspice
DescDetail: <<
 Ngspice is the name of a project aims to build an open source 
 GPLed mixed-mode/mixed-level circuit simulator. As the first step 
 (which gives clues about the naming), the project tries to improve 
 the capabilities of the original Spice3 circuit simulator originating 
 from the EECS Department of the University of California at Berkeley 
 (UCB). The "ng" prefix has a lot of meanings in this context: "Next 
 Generation, New Good" etc. depending on your preference.
<<
DescUsage: <<
 Refer to inline help and documents in package. A howto can be found at 
 http://www.brorson.com/gEDA/SPICE/simulation.html
<<
DescPort: <<
 -fix cmpp compiler for osx
 -fix loadable module support for xspice
 -Use getopt from libc instead of local compile. In case of 10.2, fink 
libgnugetopt & -shlibs
<<
#################### ngspice.info end #####################

> 
> KMBS,
> 
> Thanks for the fink info/patch pointer. I tried those and get errors, however.
> I read somewhere that these info/patch files were developed for
> 10.2, and I'm running 10.3.8/xcode 1.1. Could that be and issue?
> 
> >From the log, looks like the problem is:
> ld: Undefined symbols:
> _dlsym_prepend_underscore
> 
> Thanks,
> Greg
> 
> Here's the end of my build with the error:
> 
> gcc -g -O2 -Wall -o ngspice conf.o ngspice.o spice.o spicelib/devices/dev.o  
> -L/sw/lib frontend/libfte.a frontend/wdisp/libwindisp.a 
> frontend/plotting/libplotting.a xspice/cm/libcmxsp.a xspice/mif/libmifxsp.a 
> spicelib/devices/asrc/libasrc.a spicelib/devices/bjt/libbjt.a 
> spicelib/devices/bjt2/libbjt2.a spicelib/devices/bsim1/libbsim1.a 
> spicelib/devices/bsim2/libbsim2.a spicelib/devices/bsim3/libbsim3.a 
> spicelib/devices/bsim3v0/libbsim3v0.a spicelib/devices/bsim3v1/libbsim3v1.a 
> spicelib/devices/bsim3v1s/libbsim3v1s.a 
spicelib/devices/bsim3v1a/libbsim3v1a.a 
> spicelib/devices/bsim4/libbsim4.a spicelib/devices/cap/libcap.a 
> spicelib/devices/bsim3soi/libbsim3soi.a 
> spicelib/devices/bsim3soi_pd/libbsim3soipd.a 
> spicelib/devices/bsim3soi_fd/libbsim3soifd.a 
> spicelib/devices/bsim3soi_dd/libbsim3soidd.a spicelib/devices/cccs/libcccs.a 
> spicelib/devices/ccvs/libccvs.a spicelib/devices/cpl/libcpl.a 
> spicelib/devices/csw/libcsw.a spicelib/devices/dio/libdio.a 
> spicelib/devices/ind/libind.a spicelib/devices/isrc/libisrc.a 
> spicelib/devices/hfet1/libhfet.a spicelib/devices/hfet2/libhfet2.a 
> spicelib/devices/hisim/libhisim.a spicelib/devices/jfet/libjfet.a 
> spicelib/devices/jfet2/libjfet2.a spicelib/devices/ltra/libltra.a 
> spicelib/devices/mes/libmes.a spicelib/devices/mesa/libmesa.a 
> spicelib/devices/mos1/libmos1.a spicelib/devices/mos2/libmos2.a 
> spicelib/devices/mos3/libmos3.a spicelib/devices/mos6/libmos6.a 
> spicelib/devices/mos9/libmos9.a spicelib/devices/res/libres.a 
> spicelib/devices/soi3/libsoi3.a spicelib/devices/sw/libsw.a 
> spicelib/devices/txl/libtxl.a spicelib/devices/tra/libtra.a 
> spicelib/devices/urc/liburc.a spicelib/devices/vbic/libvbic.a 
> spicelib/devices/vccs/libvccs.a spicelib/devices/vcvs/libvcvs.a 
> spicelib/devices/vsrc/libvsrc.a spicelib/analysis/libckt.a 
> spicelib/devices/libdev.a xspice/evt/libevtxsp.a xspice/enh/libenhxsp.a 
> xspice/ipc/libipcxsp.a xspice/idn/libidnxsp.a -ldl frontend/parser/libparser.a 
> frontend/help/libhlp.a spicelib/parser/libinp.a maths/deriv/libderiv.a 
> maths/cmaths/libcmaths.a maths/poly/libpoly.a maths/ni/libni.a 
> maths/sparse/libsparse.a misc/libmisc.a -lm -lncurses -lreadline 
> -L/usr/X11R6/lib -lX11 -lXt -lXext -lXmu -lXaw -lSM -lICE
> ld: warning empty table of contents: frontend/wdisp/libwindisp.a (can't load 
> from it)
> ld: warning table of contents of library: spicelib/devices/txl/libtxl.a not 
> sorted slower link editing will result (use the ranlib(1) -s option)
> ld: warning prebinding disabled because dependent library: 
> /usr/X11R6/lib/libXaw.7.dylib is not 
> prebound/sw/lib/libncurses.dylib(lib_termcap.o) definition of _BC
> /sw/lib/libreadline.dylib(terminal.so) definition of _BC
> ld: warning multiple definitions of symbol _UP
> /sw/lib/libncurses.dylib(lib_termcap.o) definition of _UP
> /sw/lib/libreadline.dylib(terminal.so) definition of _UP
> ld: warning multiple definitions of symbol _vendorShellWidgetClass
> /usr/X11R6/lib/libXt.dylib(Vendor.o) definition of _vendorShellWidgetClass
> /usr/X11R6/lib/libXaw.dylib(Vendor.o) definition of _vendorShellWidgetClass
> ld: warning multiple definitions of symbol _vendorShellClassRec
> /usr/X11R6/lib/libXt.dylib(Vendor.o) definition of _vendorShellClassRec
> /usr/X11R6/lib/libXaw.dylib(Vendor.o) definition of _vendorShellClassRec
> ld: warning multiple definitions of symbol _shutdown
> spice.o definition of _shutdown in section (__TEXT,__text)
> /usr/lib/libdl.dylib(shutdown.So) definition of _shutdown
> ld: Undefined symbols:
> _dlsym_prepend_underscore
> make[3]: *** [ngspice] Error 1
> make[2]: *** [all-recursive] Error 1
> make[1]: *** [all-recursive] Error 1
> make: *** [all] Error 2
> ### execution of /var/tmp/tmp.1.TPFzKh failed, exit code 2
> Removing build lock...
> dpkg -r fink-buildlock-ngspice-15b-3
> (Reading database ... 37159 files and directories currently installed.)
> Removing fink-buildlock-ngspice-15b-3 ...
> Failed: phase compiling: ngspice-15b-3 failed
> 
> ld: warning multiple definitions of symbol _PC
> /sw/lib/libncurses.dylib(lib_tputs.o) definition of _PC
> /sw/lib/libreadline.dylib(terminal.so) definition of _PC
> ld: warning multiple definitions of symbol _BC
> 
> > 
> > Hi,
> > 
> > Here's info & patch file for ngspice 15b
> > http://userweb.elec.gla.ac.uk/k/kmbs/files/sw/fink/
> > 
> > -- 
> > -kmbs-
> > 
> > Quoting Greg Lauer <gl...@ls...>:
> > 
> > > I'm having problems compiling ngspice 15c with xspice
> > > on mac 10.3.8. I've tried the suggestions in the archives,
> > > ie patching xspice/icm/makedefs.in with 'LDFLAGS = -bundle'
> > > 
> > > Any suggestions? Does anyone have a fink .info (and patch if necessary)
> > > to get it to compile under fink? I can get ngspice to compile by itself,
> > > but I really need xspice for the spice2-poly conversion.
> > > 
> > > Thanks
> > > Greg
> > > 
> > > PS. Where it crashes, I get the following during make:
> > > Making all in icm
> > > gcc -I../../../../../src/include -I../../../../.. -g -O2 -Wall  -fPIC 
> > > -DCM_LINUX
> > >  -c ifspec.c
> > > gcc -I../../../../src/include -I../../../.. -g -O2 -Wall  -fPIC 
> > > -DCM_LINUX -c dl
> > > main.c
> > > gcc -I../../../../src/include -I../../../.. -g -O2 -Wall  -fPIC 
> > > -DCM_LINUX -bund
> > > le -flat_namespace -undefined suppress -o spice2poly.cm `awk '{ print $1 
> > > }' obje
> > > cts.inc` dlmain.o
> > > ld: icm_spice2poly/cfunc.o bad magic number (not a Mach-O file)
> > > make[6]: *** [spice2poly.cm] Error 1
> > > make[5]: *** [spice2poly-mods] Error 2
> > > make[4]: *** [all] Error 1
> > > make[3]: *** [all-recursive] Error 1
> > > make[2]: *** [all-recursive] Error 1
> > > make[1]: *** [all-recursive] Error 1
> > > make: *** [all] Error 2
> > > 
> > > 
> > > 
> > > 
> > > -------------------------------------------------------
> > > SF email is sponsored by - The IT Product Guide
> > > Read honest & candid reviews on hundreds of IT Products from real users.
> > > Discover which products truly live up to the hype. Start reading now.
> > > http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> > > _______________________________________________
> > > Ngspice-devel mailing list
> > > Ngs...@li...
> > > https://lists.sourceforge.net/lists/listinfo/ngspice-devel
> > > 
> > > 
> > > 
> > 
> > 
> > -------------------------------------------------
> > This mail sent through IMP: http://horde.org/imp/
> 
> 
> 
> 
> -------------------------------------------------------
> SF email is sponsored by - The IT Product Guide
> Read honest & candid reviews on hundreds of IT Products from real users.
> Discover which products truly live up to the hype. Start reading now.
> http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> _______________________________________________
> Ngspice-devel mailing list
> Ngs...@li...
> https://lists.sourceforge.net/lists/listinfo/ngspice-devel

Re: [Ngspice-devel] ngspice compile problems on osx

[Greg L. <gl...@ls...>]

KMBS,

Thanks for the fink info/patch pointer. I tried those and get errors, however.
I read somewhere that these info/patch files were developed for
10.2, and I'm running 10.3.8/xcode 1.1. Could that be and issue?

From the log, looks like the problem is:
ld: Undefined symbols:
_dlsym_prepend_underscore

Thanks,
Greg

Here's the end of my build with the error:

gcc -g -O2 -Wall -o ngspice conf.o ngspice.o spice.o spicelib/devices/dev.o  
-L/sw/lib frontend/libfte.a frontend/wdisp/libwindisp.a 
frontend/plotting/libplotting.a xspice/cm/libcmxsp.a xspice/mif/libmifxsp.a 
spicelib/devices/asrc/libasrc.a spicelib/devices/bjt/libbjt.a 
spicelib/devices/bjt2/libbjt2.a spicelib/devices/bsim1/libbsim1.a 
spicelib/devices/bsim2/libbsim2.a spicelib/devices/bsim3/libbsim3.a 
spicelib/devices/bsim3v0/libbsim3v0.a spicelib/devices/bsim3v1/libbsim3v1.a 
spicelib/devices/bsim3v1s/libbsim3v1s.a spicelib/devices/bsim3v1a/libbsim3v1a.a 
spicelib/devices/bsim4/libbsim4.a spicelib/devices/cap/libcap.a 
spicelib/devices/bsim3soi/libbsim3soi.a 
spicelib/devices/bsim3soi_pd/libbsim3soipd.a 
spicelib/devices/bsim3soi_fd/libbsim3soifd.a 
spicelib/devices/bsim3soi_dd/libbsim3soidd.a spicelib/devices/cccs/libcccs.a 
spicelib/devices/ccvs/libccvs.a spicelib/devices/cpl/libcpl.a 
spicelib/devices/csw/libcsw.a spicelib/devices/dio/libdio.a 
spicelib/devices/ind/libind.a spicelib/devices/isrc/libisrc.a 
spicelib/devices/hfet1/libhfet.a spicelib/devices/hfet2/libhfet2.a 
spicelib/devices/hisim/libhisim.a spicelib/devices/jfet/libjfet.a 
spicelib/devices/jfet2/libjfet2.a spicelib/devices/ltra/libltra.a 
spicelib/devices/mes/libmes.a spicelib/devices/mesa/libmesa.a 
spicelib/devices/mos1/libmos1.a spicelib/devices/mos2/libmos2.a 
spicelib/devices/mos3/libmos3.a spicelib/devices/mos6/libmos6.a 
spicelib/devices/mos9/libmos9.a spicelib/devices/res/libres.a 
spicelib/devices/soi3/libsoi3.a spicelib/devices/sw/libsw.a 
spicelib/devices/txl/libtxl.a spicelib/devices/tra/libtra.a 
spicelib/devices/urc/liburc.a spicelib/devices/vbic/libvbic.a 
spicelib/devices/vccs/libvccs.a spicelib/devices/vcvs/libvcvs.a 
spicelib/devices/vsrc/libvsrc.a spicelib/analysis/libckt.a 
spicelib/devices/libdev.a xspice/evt/libevtxsp.a xspice/enh/libenhxsp.a 
xspice/ipc/libipcxsp.a xspice/idn/libidnxsp.a -ldl frontend/parser/libparser.a 
frontend/help/libhlp.a spicelib/parser/libinp.a maths/deriv/libderiv.a 
maths/cmaths/libcmaths.a maths/poly/libpoly.a maths/ni/libni.a 
maths/sparse/libsparse.a misc/libmisc.a -lm -lncurses -lreadline 
-L/usr/X11R6/lib -lX11 -lXt -lXext -lXmu -lXaw -lSM -lICE
ld: warning empty table of contents: frontend/wdisp/libwindisp.a (can't load 
from it)
ld: warning table of contents of library: spicelib/devices/txl/libtxl.a not 
sorted slower link editing will result (use the ranlib(1) -s option)
ld: warning prebinding disabled because dependent library: 
/usr/X11R6/lib/libXaw.7.dylib is not 
prebound/sw/lib/libncurses.dylib(lib_termcap.o) definition of _BC
/sw/lib/libreadline.dylib(terminal.so) definition of _BC
ld: warning multiple definitions of symbol _UP
/sw/lib/libncurses.dylib(lib_termcap.o) definition of _UP
/sw/lib/libreadline.dylib(terminal.so) definition of _UP
ld: warning multiple definitions of symbol _vendorShellWidgetClass
/usr/X11R6/lib/libXt.dylib(Vendor.o) definition of _vendorShellWidgetClass
/usr/X11R6/lib/libXaw.dylib(Vendor.o) definition of _vendorShellWidgetClass
ld: warning multiple definitions of symbol _vendorShellClassRec
/usr/X11R6/lib/libXt.dylib(Vendor.o) definition of _vendorShellClassRec
/usr/X11R6/lib/libXaw.dylib(Vendor.o) definition of _vendorShellClassRec
ld: warning multiple definitions of symbol _shutdown
spice.o definition of _shutdown in section (__TEXT,__text)
/usr/lib/libdl.dylib(shutdown.So) definition of _shutdown
ld: Undefined symbols:
_dlsym_prepend_underscore
make[3]: *** [ngspice] Error 1
make[2]: *** [all-recursive] Error 1
make[1]: *** [all-recursive] Error 1
make: *** [all] Error 2
### execution of /var/tmp/tmp.1.TPFzKh failed, exit code 2
Removing build lock...
dpkg -r fink-buildlock-ngspice-15b-3
(Reading database ... 37159 files and directories currently installed.)
Removing fink-buildlock-ngspice-15b-3 ...
Failed: phase compiling: ngspice-15b-3 failed

ld: warning multiple definitions of symbol _PC
/sw/lib/libncurses.dylib(lib_tputs.o) definition of _PC
/sw/lib/libreadline.dylib(terminal.so) definition of _PC
ld: warning multiple definitions of symbol _BC

> 
> Hi,
> 
> Here's info & patch file for ngspice 15b
> http://userweb.elec.gla.ac.uk/k/kmbs/files/sw/fink/
> 
> -- 
> -kmbs-
> 
> Quoting Greg Lauer <gl...@ls...>:
> 
> > I'm having problems compiling ngspice 15c with xspice
> > on mac 10.3.8. I've tried the suggestions in the archives,
> > ie patching xspice/icm/makedefs.in with 'LDFLAGS = -bundle'
> > 
> > Any suggestions? Does anyone have a fink .info (and patch if necessary)
> > to get it to compile under fink? I can get ngspice to compile by itself,
> > but I really need xspice for the spice2-poly conversion.
> > 
> > Thanks
> > Greg
> > 
> > PS. Where it crashes, I get the following during make:
> > Making all in icm
> > gcc -I../../../../../src/include -I../../../../.. -g -O2 -Wall  -fPIC 
> > -DCM_LINUX
> >  -c ifspec.c
> > gcc -I../../../../src/include -I../../../.. -g -O2 -Wall  -fPIC 
> > -DCM_LINUX -c dl
> > main.c
> > gcc -I../../../../src/include -I../../../.. -g -O2 -Wall  -fPIC 
> > -DCM_LINUX -bund
> > le -flat_namespace -undefined suppress -o spice2poly.cm `awk '{ print $1 
> > }' obje
> > cts.inc` dlmain.o
> > ld: icm_spice2poly/cfunc.o bad magic number (not a Mach-O file)
> > make[6]: *** [spice2poly.cm] Error 1
> > make[5]: *** [spice2poly-mods] Error 2
> > make[4]: *** [all] Error 1
> > make[3]: *** [all-recursive] Error 1
> > make[2]: *** [all-recursive] Error 1
> > make[1]: *** [all-recursive] Error 1
> > make: *** [all] Error 2
> > 
> > 
> > 
> > 
> > -------------------------------------------------------
> > SF email is sponsored by - The IT Product Guide
> > Read honest & candid reviews on hundreds of IT Products from real users.
> > Discover which products truly live up to the hype. Start reading now.
> > http://ads.osdn.com/?ad_id=6595&alloc_id=14396&op=click
> > _______________________________________________
> > Ngspice-devel mailing list
> > Ngs...@li...
> > https://lists.sourceforge.net/lists/listinfo/ngspice-devel
> > 
> > 
> > 
> 
> 
> -------------------------------------------------
> This mail sent through IMP: http://horde.org/imp/