setedit-users — General list about setedit

Re: Installation problem

[Salvador E. T. <sal...@in...>]

On 11/09/05 13:21, 52...@ni... wrote:

>while installation of setedit in linux redhat 9.0 i m stucking at the
>following message
>install the TURBO version 1st
>  
>

How did you install it? If you used rpm command: did you install 
librhtv-2.0.3-3rh9.i386.rpm first?

>notedly i have installed the following under root directory
>
>setedit-0.5.4-1rh9.i386.rpm
>librhtv-2.0.3-3rh9.i386.rpm
>setedit-0.5.2.RC1.tar.tar
>
>pls give me the soloution
>  
>

Khuda Hafiz, SET

-- 
Salvador Eduardo Tropea (SET). (Electronics Engineer)
Visit my home page: http://welcome.to/SetSoft or
http://www.geocities.com/SiliconValley/Vista/6552/
Alternative e-mail: se...@co... se...@ie... 
Address: Curapaligue 2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA Phone: +(5411) 4759 0013

Installation problem

[<52...@ni...>]

while installation of setedit in linux redhat 9.0 i m stucking at the
following message
install the TURBO version 1st

notedly i have installed the following under root directory

setedit-0.5.4-1rh9.i386.rpm
librhtv-2.0.3-3rh9.i386.rpm
setedit-0.5.2.RC1.tar.tar

pls give me the soloution
-- 
Muhammad Wajid
National University of Science and Technology (NUST)
Pakistan
051-5507197

Multithreaded and forking version of former email-fmt.py

[Grzegorz A. H. <gr...@ti...>]

Hello there! Long time no see.

After spending a lot of time doing nothing at all, my parents have
prodded me to find a job again. Earlier this week I had an interview
where they wanted to ask me about being able to write multithreaded
programs or communicate with other processes.

So I thought I would rewrite the email-fmt.py script I contributed
two years ago.  I still don't know the results of the interview,
but it was quite fun writting this version and it is much better:
it no longer uses temporary files, and since it is in C, it doesn't
waste time loading the huge python interpreter.

I was wondering if those of you using the format script could give
this a try again. In fact, Tim, are you still around here? Tim
wanted the email-fmt program to reformat the lines of text to the
size of the currently opened window. SET, I think this could be
passed as a parameter to the program, but how would you get the
current window size?

And, the previous python version did fit ok as help
documentation. But a C project is much more complex than
copy&paste. Would you like it to as something in a "contrib"
directory?

As a final note, profiling the C version agains the python one shows
on the command line that python is about four times slower. However,
no matter which version I run from inside setedit, it is much slower
than the command line version. Why does this happen?

Here's the C version of email-fmt. Enjoy!

  http://gradha.sdf-eu.org/email-fmt.tar.gz

Re: Tab problem

[Thiago F.G. A. <tf...@te...>]

Right...! I had forgotten about this :)
Thank you.

[],
Thiago

On Thu, 4 Aug 2005, Salvador Eduardo Tropea wrote:

> I think the problem is that you have a custom keys assignment and hence you 
> don't have Tab key assigned to cmcSmartTab command:
>
> Revision 1.654  2005/02/21 18:05:19  set
> * Modified: Now the "Tab" behavior is implemented by an editor command
> (cmcSmartTab). In this way sLisp code can insert "smart tabs". Tabs in
> pseudo macros are translated to this command and the default keyboard
> assigment for Tab is this command. People using a keybind.dat file will
> need to add this assigment in order to keep the same behavior. Added as a
> result of a thread in the mailing list (Arnold and Thiago).

-- 

()  ascii ribbon campaign - against html mail 
/\                        - against microsoft attachments

Re: Tab problem

[Salvador E. T. <sal...@in...>]

Thiago F.G. Albuquerque wrote:

> On Mon, 1 Aug 2005, Salvador Eduardo Tropea wrote:
>
>> Thiago F.G. Albuquerque wrote:
>>
>>>
>>> After I built the new version of SETEDIT in DJGPP, the editor uses 
>>> tabs all the time, even if the option "Use tabs" is not marked in 
>>> "Global Edition".
>>
>>
>>
>> What about the local options?
>
>
> It isn't marked there either.


I think the problem is that you have a custom keys assignment and hence 
you don't have Tab key assigned to cmcSmartTab command:

Revision 1.654  2005/02/21 18:05:19  set
* Modified: Now the "Tab" behavior is implemented by an editor command
(cmcSmartTab). In this way sLisp code can insert "smart tabs". Tabs in
pseudo macros are translated to this command and the default keyboard
assigment for Tab is this command. People using a keybind.dat file will
need to add this assigment in order to keep the same behavior. Added as a
result of a thread in the mailing list (Arnold and Thiago).

SET

-- 
Salvador Eduardo Tropea (SET). (Electronics Engineer)
Visit my home page: http://welcome.to/SetSoft or
http://www.geocities.com/SiliconValley/Vista/6552/
Alternative e-mail: se...@co... se...@ie... 
Address: Curapaligue 2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA Phone: +(5411) 4759 0013

Re: cmcToggleCharCase patch

[Salvador E. T. <sal...@in...>]

Thiago F.G. Albuquerque wrote:

> Hi,
>
> This patch fixes one anoyance in cmcToggleCharCase. Let me explain.
>
> cmcToggleCharCase behaves like this: if the char under the cursor is a 
> letter, it toggles the case and moves to the right; if it is not, it 
> just doesn't do anything, the cursor stays in the same position. Now 
> suppose you're toggling the case of a identifier with a underscore in 
> the midle, e.g.
>
> set_files
>
> You position the cursor in the first letter, and start "toggling", one 
> letter at a time. When the cursor reaches the '_', it stops. Then you 
> have to move to the right yourself to continue the process. That's the 
> annoyance. When the char under cursor is not a letter, 
> cmcToggleCharCase should move to the right too. Tha't what the patch 
> fixes.

Ok, applied.

SET

-- 
Salvador Eduardo Tropea (SET). (Electronics Engineer)
Visit my home page: http://welcome.to/SetSoft or
http://www.geocities.com/SiliconValley/Vista/6552/
Alternative e-mail: se...@co... se...@ie... 
Address: Curapaligue 2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA Phone: +(5411) 4759 0013

cmcToggleCharCase patch

[Thiago F.G. A. <tf...@te...>]

Hi,

This patch fixes one anoyance in cmcToggleCharCase. Let me explain.

cmcToggleCharCase behaves like this: if the char under the cursor is a 
letter, it toggles the case and moves to the right; if it is not, it just 
doesn't do anything, the cursor stays in the same position. Now suppose 
you're toggling the case of a identifier with a underscore in the midle, e.g.

set_files

You position the cursor in the first letter, and start "toggling", one 
letter at a time. When the cursor reaches the '_', it stops. Then you have 
to move to the right yourself to continue the process. That's the 
annoyance. When the char under cursor is not a letter, cmcToggleCharCase 
should move to the right too. Tha't what the patch fixes.

[],
Thiago

Re: Tab problem

[Thiago F.G. A. <tf...@te...>]

On Mon, 1 Aug 2005, Salvador Eduardo Tropea wrote:

> Thiago F.G. Albuquerque wrote:
>
>> 
>> After I built the new version of SETEDIT in DJGPP, the editor uses tabs all 
>> the time, even if the option "Use tabs" is not marked in "Global Edition".
>
>
> What about the local options?

It isn't marked there either.

[],
Thiago

-- 
()  ascii ribbon campaign - against html mail 
/\                        - against microsoft attachments

Re: DJGPP - Configure errors

[Salvador E. T. <sal...@in...>]

Thiago F.G. Albuquerque wrote:

> It was the last snapshot I downloaded. I don't remember the exact 
> date... It must have been one or two days before 11 Jul, which is the 
> date of the first "DJGPP - Configure Errors" message.

Ok.

> Is there any revision/version information in the snapshot? Where can I 
> find it? If you want, I can look it up for you.

Only SETEdit have it, in include/vername.h

SET

-- 
Salvador Eduardo Tropea (SET). (Electronics Engineer)
Visit my home page: http://welcome.to/SetSoft or
http://www.geocities.com/SiliconValley/Vista/6552/
Alternative e-mail: se...@co... se...@ie... 
Address: Curapaligue 2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA Phone: +(5411) 4759 0013

Re: DJGPP - Configure errors

[Thiago F.G. A. <tf...@te...>]

On Mon, 1 Aug 2005, Salvador Eduardo Tropea wrote:

> Thiago F.G. Albuquerque wrote:

[snip]

>> And yes, all the other methods in this class are defined the same way: so 
>> why the compiler would give an error only with this one? The only 
>> explanation I can come up with is that the other methods are not used (is 
>> this true?), so the linker doesn't look for them.
>
>
> Nope, you are wrong, that's a bug in the compiler.

Ok, it was just a guess... :)

[snip]

>> Thank you. I had some difficulties applying the patch. It failed in hunk 
>> #2, because in my version of the file, "#include <pathtool.h> // 
>> IsADirectory" is in a different position. So I deleted hunk 2, applied the 
>> patch and then removed this line manually. Everything compiled fine.
>
>
> Where did you get the sources? I'm using CVS code as reference.

It was the last snapshot I downloaded. I don't remember the exact date... 
It must have been one or two days before 11 Jul, which is the date of the 
first "DJGPP - Configure Errors" message.

Is there any revision/version information in the snapshot? Where can I 
find it? If you want, I can look it up for you.

[],
Thiago

-- 
()  ascii ribbon campaign - against html mail 
/\                        - against microsoft attachments

Re: DJGPP - Configure errors

[Oswald B. <os...@kd...>]

On Mon, Aug 01, 2005 at 12:55:01PM -0300, Salvador Eduardo Tropea wrote:
> Oswald Buddenhagen wrote:
> > the gcc folks decided to make this an error in gcc 3.4. i like it
> > that way ...
> > 
> >
> Not at least for my copy:
> 
it might depend on some switch.
dunno why the compiler freaks out at this particular statement, though.
maybe some particularly strange kind of memory corruption. :)

> And I don't like it this way because macros like it:
> 
> #define XXXX(a) YYYY(a);
> 
these macros are an incarnation of pure evil. :)
yes, cpp sucks. particularly with c++, where macros from c headers are a
major namespace polluter.

-- 
Hi! I'm a .signature virus! Copy me into your ~/.signature, please!
--
Chaos, panic, and disorder - my work here is done.

Re: DJGPP - Configure errors

[Salvador E. T. <sal...@in...>]

Oswald Buddenhagen wrote:

>On Mon, Aug 01, 2005 at 10:45:16AM -0300, Salvador Eduardo Tropea wrote:=

> =20
>
>>Empty lines are allowed in C and C++, this semicolon is an empty line, =

>>just like this:
>>
>>void pp()
>>{
>>}
>>
>>// Next is empty
>>;
>>
>>   =20
>>
>that's not an empty line, but an empty statement=20
>
Right.

>(in a context where no statement is required).
>
Not required but allowed, etc.

> the gcc folks decided to make this an error in
>gcc 3.4. i like it that way ...
> =20
>
Not at least for my copy:

kotori:/tmp$ g++-3.4 --version
g++-3.4 (GCC) 3.4.4 20050314 (prerelease) (Debian 3.4.3-13)
Copyright (C) 2004 Free Software Foundation, Inc.
Esto es software libre; vea el c=F3digo para las condiciones de copia.  N=
O hay
garant=EDa; ni siquiera para MERCANTIBILIDAD o IDONEIDAD PARA UN PROP=D3S=
ITO EN
PARTICULAR

I compiled it without warnings: (-Wall used):
#include <stdio.h>

void pp();;;

int main(int argc, char *argv[])
{
 printf("Hello\n");;
 return 0;
}
;

And I don't like it this way because macros like it:

#define XXXX(a) YYYY(a);

Becomes wrong if used like this:

XXXX(a);

And in this case is gcc 2.95.2 and getting confused by one in a dozen of =

lines like it.
BTW: Andris tried the code using 4.0.0 and didn't found it.

SET

--=20
Salvador Eduardo Tropea (SET). (Electronics Engineer)
Visit my home page: http://welcome.to/SetSoft or
http://www.geocities.com/SiliconValley/Vista/6552/
Alternative e-mail: se...@co... se...@ie...=20
Address: Curapaligue 2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA Phone: +(5411) 4759 0013

Re: DJGPP - Configure errors

[Oswald B. <os...@kd...>]

On Mon, Aug 01, 2005 at 05:29:22PM +0200, Oswald Buddenhagen wrote:
> On Mon, Aug 01, 2005 at 10:45:16AM -0300, Salvador Eduardo Tropea wrote:
> > void pp()
> > {
> > }
> > 
> > // Next is empty
> > ;
> > 
> that's not an empty line, but an empty statement (in a context where no
> statement is required).
> 
actually, it's worse. it is an empty statement in a context where no
empty statements are allowed in the first place - there is no such thing
as an empty declaration or definition. so the gcc folks were perfectly
safe in just stopping to accept it.

-- 
Hi! I'm a .signature virus! Copy me into your ~/.signature, please!
--
Chaos, panic, and disorder - my work here is done.

Re: DJGPP - Configure errors

[Oswald B. <os...@kd...>]

On Mon, Aug 01, 2005 at 10:45:16AM -0300, Salvador Eduardo Tropea wrote:
> Empty lines are allowed in C and C++, this semicolon is an empty line, 
> just like this:
> 
> void pp()
> {
> }
> 
> // Next is empty
> ;
> 
that's not an empty line, but an empty statement (in a context where no
statement is required). the gcc folks decided to make this an error in
gcc 3.4. i like it that way ...

-- 
Hi! I'm a .signature virus! Copy me into your ~/.signature, please!
--
Chaos, panic, and disorder - my work here is done.

Re: DJGPP - Configure errors

[Salvador E. T. <sal...@in...>]

Thiago F.G. Albuquerque wrote:

>> > On Tue, 19 Jul 2005, Salvador Eduardo Tropea wrote:
>> > > > ... that in this case include/tv/nointl.h is used where:
>> > > > > static char       *getTextNew(const char *msgid, Boolean 
>> onlyIntl=False)
>> > >                   { return onlyIntl ? NULL : newStr(msgid); };
>> >                                                                ^
>> > That's the problem!--------------------------------------------|
>> > (sorry for those using proportional fonts :)
>> > > I removed this ';' after the closing bracket and it solved the 
>> problem.
>> > I am sending a patch attached.
>>
>> Ugh!! that's quite bizarre, a bug in the compiler? Why? The same 
>> class have the same on other members ...
>
>
> Yes, but this semi-colon shouldn't be there anyway:

Empty lines are allowed in C and C++, this semicolon is an empty line, 
just like this:

void pp()
{
}

// Next is empty
;

> this is a definition, not a declaration. Think about it: if this was a 
> "normal" function -- a function defined ouside the class declaration 
> -- you wouldn't put a semi-colon after the body, would you? And that 
> may explain the unresolved reference error: because of the semi-colon, 
> the compiler saw this as a declaration, and look for the definition 
> elsewhere. As he didn't find it...


> And yes, all the other methods in this class are defined the same way: 
> so why the compiler would give an error only with this one? The only 
> explanation I can come up with is that the other methods are not used 
> (is this true?), so the linker doesn't look for them.


Nope, you are wrong, that's a bug in the compiler.

>
>> This is part of pathtool.cc, it seems to be missing and is only used 
>> by DOS. I'm fixing it, patch attached.
>
>
> Thank you. I had some difficulties applying the patch. It failed in 
> hunk #2, because in my version of the file, "#include <pathtool.h> // 
> IsADirectory" is in a different position. So I deleted hunk 2, applied 
> the patch and then removed this line manually. Everything compiled fine.


Where did you get the sources? I'm using CVS code as reference.

> There is a bug in the editor involving tabs. I'll start another thread 
> to talk about this.
>
Ok.

SET

-- 
Salvador Eduardo Tropea (SET). (Electronics Engineer)
Visit my home page: http://welcome.to/SetSoft or
http://www.geocities.com/SiliconValley/Vista/6552/
Alternative e-mail: se...@co... se...@ie... 
Address: Curapaligue 2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA Phone: +(5411) 4759 0013

Re: Tab problem

[Salvador E. T. <sal...@in...>]

Thiago F.G. Albuquerque wrote:

>
> After I built the new version of SETEDIT in DJGPP, the editor uses 
> tabs all the time, even if the option "Use tabs" is not marked in 
> "Global Edition".


What about the local options?


SET

-- 
Salvador Eduardo Tropea (SET). (Electronics Engineer)
Visit my home page: http://welcome.to/SetSoft or
http://www.geocities.com/SiliconValley/Vista/6552/
Alternative e-mail: se...@co... se...@ie... 
Address: Curapaligue 2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA Phone: +(5411) 4759 0013

Re: DJGPP - Configure errors

[Thiago F.G. A. <tf...@te...>]

> > On Tue, 19 Jul 2005, Salvador Eduardo Tropea wrote:
> > 
> > > ... that in this case include/tv/nointl.h is used where:
> > > 
> > > static char       *getTextNew(const char *msgid, Boolean onlyIntl=False)
> > >                   { return onlyIntl ? NULL : newStr(msgid); };
> >                                                                ^
> > That's the problem!--------------------------------------------|
> > (sorry for those using proportional fonts :)
> > 
> > I removed this ';' after the closing bracket and it solved the problem.
> > I am sending a patch attached.
> 
> Ugh!! that's quite bizarre, a bug in the compiler? Why? The same class 
> have the same on other members ...

Yes, but this semi-colon shouldn't be there anyway: this is a definition, 
not a declaration. Think about it: if this was a "normal" function -- a 
function defined ouside the class declaration -- you wouldn't put a 
semi-colon after the body, would you? And that may explain the unresolved 
reference error: because of the semi-colon, the compiler saw this as a 
declaration, and look for the definition elsewhere. As he didn't find 
it...

And yes, all the other methods in this class are defined the same way: so 
why the compiler would give an error only with this one? The only 
explanation I can come up with is that the other methods are not used (is 
this true?), so the linker doesn't look for them.

> This is part of pathtool.cc, it seems to be missing and is only used by 
> DOS. I'm fixing it, patch attached.

Thank you. I had some difficulties applying the patch. It failed in hunk 
#2, because in my version of the file, "#include <pathtool.h> // 
IsADirectory" is in a different position. So I deleted hunk 2, applied the 
patch and then removed this line manually. Everything compiled fine.

There is a bug in the editor involving tabs. I'll start another thread to 
talk about this.

[],
Thiago

Index: infview/infbase.cc
===================================================================
RCS file: /cvsroot/setedit/setedit/infview/infbase.cc,v
retrieving revision 1.19
diff -u -r1.19 infbase.cc
--- infview/infbase.cc	19 Jul 2005 16:17:24 -0000	1.19
+++ infview/infbase.cc	27 Jul 2005 20:39:35 -0000
@@ -1,4 +1,4 @@
-/* Copyright (C) 1996-2003 by Salvador E. Tropea (SET),
+/* Copyright (C) 1996-2005 by Salvador E. Tropea (SET),
     see copyrigh file for details */
  /***************************************************************

@@ -41,7 +41,6 @@
  #include <fcntl.h> // open
  #include <gzfiles.h>
  #include <dyncat.h>
-#include <pathtool.h> // IsADirectory

  #ifdef TVCompf_djgpp
  #include <crt0.h>
@@ -1478,6 +1477,12 @@
  }

  static
+int IsADirectory(const char *name)
+{
+ return !access(name,D_OK);
+}
+
+static
  void GuessInfoPath(DynStrCatStruct *InfoPath)
  {
   char *djgpp=getenv("DJDIR");


-- 
()  ascii ribbon campaign - against html mail 
/\                        - against microsoft attachments

Tab problem

[Thiago F.G. A. <tf...@te...>]

After I built the new version of SETEDIT in DJGPP, the editor uses tabs 
all the time, even if the option "Use tabs" is not marked in "Global 
Edition".

[],
Thiago

-- 
()  ascii ribbon campaign - against html mail 
/\                        - against microsoft attachments

Re: DJGPP - Configure errors

[Salvador E. T. <sal...@in...>]

Thiago F.G. Albuquerque wrote:

> On Tue, 19 Jul 2005, Salvador Eduardo Tropea wrote:
>
>> Thiago F.G. Albuquerque wrote:
>>
>>> On Mon, 11 Jul 2005, Salvador Eduardo Tropea wrote:
>>
>
> [snip]
>
>> ... that in this case include/tv/nointl.h is used where:
>>
>> static char       *getTextNew(const char *msgid, Boolean onlyIntl=False)
>>                   { return onlyIntl ? NULL : newStr(msgid); };
>
>                                                                ^
> That's the problem!--------------------------------------------|
> (sorry for those using proportional fonts :)
>
> I removed this ';' after the closing bracket and it solved the problem.
> I am sending a patch attached.


Ugh!! that's quite bizarre, a bug in the compiler? Why? The same class 
have the same on other members ...

> ----------------------------------------------------------------
>
> Then I got the error Oswald Buddenhagen mentioned in his email:
>
> libsettv.a(tdiagaid.o)(.text+0x70c):tdiagaid.cc: undefined reference 
> to `TSNoStaticText::TSNoStaticText(char const *)'
>
> I folowed his instructions to fix it.


This should be solved in the last snapshot.

> Ok, so now the editor is built. (yes!)
>
> ----------------------------------------------------------------
>
> Now I have this problem when building infview:
>
> objinf/infbase.o(.text+0x2d39):infbase.cc: undefined reference to 
> `IsADirectory(char const *)'
>
> I didn't have the time to look into this yet.


This is part of pathtool.cc, it seems to be missing and is only used by 
DOS. I'm fixing it, patch attached.

> ----------------------------------------------------------------
>
>> BTW which gcc and binutils versions are you using?
>
>
> gcc:      2.95.2 19991024 (release)
> binutils: 2.9.5

I used 2.95.4 for a long time I'll recommed you it.

SET

-- 
Salvador Eduardo Tropea (SET). (Electronics Engineer)
Visit my home page: http://welcome.to/SetSoft or
http://www.geocities.com/SiliconValley/Vista/6552/
Alternative e-mail: se...@co... se...@ie... 
Address: Curapaligue 2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA Phone: +(5411) 4759 0013

Re: DJGPP - Configure errors

[Thiago F.G. A. <tf...@te...>]

On Tue, 19 Jul 2005, Salvador Eduardo Tropea wrote:

> Thiago F.G. Albuquerque wrote:
>
>> On Mon, 11 Jul 2005, Salvador Eduardo Tropea wrote:

[snip]

> ... that in this case include/tv/nointl.h is used where:
>
> static char       *getTextNew(const char *msgid, Boolean onlyIntl=False)
>                   { return onlyIntl ? NULL : newStr(msgid); };
                                                                ^
That's the problem!--------------------------------------------|
(sorry for those using proportional fonts :)

I removed this ';' after the closing bracket and it solved the problem.
I am sending a patch attached.

----------------------------------------------------------------

Then I got the error Oswald Buddenhagen mentioned in his email:

libsettv.a(tdiagaid.o)(.text+0x70c):tdiagaid.cc: undefined 
reference to `TSNoStaticText::TSNoStaticText(char const *)'

I folowed his instructions to fix it.

Ok, so now the editor is built. (yes!)

----------------------------------------------------------------

Now I have this problem when building infview:

objinf/infbase.o(.text+0x2d39):infbase.cc: undefined reference to 
`IsADirectory(char const *)'

I didn't have the time to look into this yet.

----------------------------------------------------------------

> BTW which gcc and binutils versions are you using?

gcc:      2.95.2 19991024 (release)
binutils: 2.9.5

----------------------------------------------------------------

[],
Thiago

-- 
()  ascii ribbon campaign - against html mail 
/\                        - against microsoft attachments

Windows XP and MS Visual C++ free toolkit

[Arnold W. <awi...@te...>]

Hello Salvador,

After a long detour, I've got a bit more time to work on SETedit.

Meanwhile, for other reasons I had to upgrade to XP Home edition and 
have now downloaded the MS free C++ toolkit and am ready to give it a 
shot at compiling SETedit. As it turns out the toolkit does have a 
debugger, although I have not had any chance to put it through its paces 
yet.

I have downloaded the latest snapshots and have also installed the BCC 
W32 version on the XP machine, and it does run with some problem when 
trying to save some configurations. I haven't spent much time sorting it 
out as I'd like to compile it fresh so I can debug it.

While I can't say I have kept up with the latest developments, I have 
noticed that there are some changes in the configuration and the 
documentation with reference to confignt.pl, however looking it over it 
talks about djgpp, so, I thought it best to ask first, what the best way 
to try and compile under both BCC and the toolkit might be.

I have tried under BCC55, but it does not find the compiler include 
directories; I'll start work on sorting that out meanwhile.

Any suggestions on trying the toolkit would be welcome. I do have some 
background with Perl, so I'm hoping we can sort out the configuration 
without too much problem.

TIA,

Arnold

Re: compile error

[Salvador E. T. <sal...@in...>]

Oswald Buddenhagen wrote:

>On Tue, Jul 19, 2005 at 03:41:57PM -0300, Salvador Eduardo Tropea wrote:
>  
>
>>>ah, that's because libeasyd.a _must_ (yes, really, it's a standard)
>>>
>>>      
>>>
>>A standard?
>>
>>    
>>
>[gnu ld man page, option -l<lib>]
>  
>
That isn't part of any standard.

>    The  linker  will  search  an archive only once, at the location
>    where it is specified on  the  command  line.   If  the  archive
>    defines  a  symbol  which  was  undefined  in  some object which
>    appeared before the archive on the command line, the linker will
>    include  the  appropriate file(s) from the archive.  However, an
>    undefined symbol in an object appearing  later  on  the  command
>    line will not cause the linker to search the archive again.
>
>    See  the  -(  option  for  a  way  to force the linker to search
>    archives multiple times.
>
>    You may list the same archive  multiple  times  on  the  command
>    line.
>  
>

In my experience it was true before ld started to use collect2.
The library order that you pointed out was there for more than four years.

>>>>This  type  of  archive  searching is standard for Unix linkers.
>>>>        
>>>>
>    However, if you are using ld on AIX, note that it  is  different
>    from the behaviour of the AIX linker.
>
>well, it's not _a_ standard, but it is standard. :)
>
>  
>
>>>come after libsettv.a (which depends on the former) - even if some
>>>linkers grok it nonetheless.
>>>
>>>      
>>>
>>Just some linkers? all binutils I tried do.
>>
>>    
>>
>well, in version 2.16.1 they don't anymore, by default.
>  
>

No problem, I already solved it.

SET

-- 
Salvador Eduardo Tropea (SET). (Electronics Engineer)
Visit my home page: http://welcome.to/SetSoft or
http://www.geocities.com/SiliconValley/Vista/6552/
Alternative e-mail: se...@co... se...@ie... 
Address: Curapaligue 2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA Phone: +(5411) 4759 0013

Re: compile error

[Oswald B. <os...@kd...>]

On Tue, Jul 19, 2005 at 03:41:57PM -0300, Salvador Eduardo Tropea wrote:
> >ah, that's because libeasyd.a _must_ (yes, really, it's a standard)
> >
> 
> A standard?
> 
[gnu ld man page, option -l<lib>]

    The  linker  will  search  an archive only once, at the location
    where it is specified on  the  command  line.   If  the  archive
    defines  a  symbol  which  was  undefined  in  some object which
    appeared before the archive on the command line, the linker will
    include  the  appropriate file(s) from the archive.  However, an
    undefined symbol in an object appearing  later  on  the  command
    line will not cause the linker to search the archive again.

    See  the  -(  option  for  a  way  to force the linker to search
    archives multiple times.

    You may list the same archive  multiple  times  on  the  command
    line.

>>> This  type  of  archive  searching is standard for Unix linkers.
    However, if you are using ld on AIX, note that it  is  different
    from the behaviour of the AIX linker.

well, it's not _a_ standard, but it is standard. :)

> >come after libsettv.a (which depends on the former) - even if some
> >linkers grok it nonetheless.
> >
> 
> Just some linkers? all binutils I tried do.
>
well, in version 2.16.1 they don't anymore, by default.

-- 
Hi! I'm a .signature virus! Copy me into your ~/.signature, please!
--
Chaos, panic, and disorder - my work here is done.

Re: compile error

[Salvador E. T. <sal...@in...>]

Oswald Buddenhagen wrote:

>On Tue, Jul 12, 2005 at 04:38:26PM +0200, Oswald Buddenhagen wrote:
>  
>
>>libsettv.a(tdiagaid.o): In function `CreateAddInsDelDialog(int, int, char const*, int, int, int)':
>>../settvuti/tdiagaid.cc:206: undefined reference to `TSNoStaticText::TSNoStaticText(char const*)'
>>collect2: ld returned 1 exit status
>>
>>    
>>
>ah, that's because libeasyd.a _must_ (yes, really, it's a standard)
>

A standard?

>come after libsettv.a (which depends on the former) - even if some linkers
>grok it nonetheless.
>  
>

Just some linkers? all binutils I tried do.
Anyways, I agree with you, they should be in this order.

>it must be in editor.gpr.
>  
>

Yes, but they can't be manually sorted.

>manually i fixed the LIBRARIES line in editor.umk.
>  
>

Ok, I updated other .umk and .imk files with the same problem.
Now the tool to extract the dependencies from the .gpr file (extrimk) 
can be configured with the "priority" for each library.

>>for some weird reason the compiler command lines are not printed
>>
>>    
>>
>huh, that was a > /dev/null in my build script. :}
>i'm just wondering why i put it there ... i suppose i'm spoiled by the
>kde build system that makes no noise unless something goes wrong. :)
>  
>

:-) ok.

SET

-- 
Salvador Eduardo Tropea (SET). (Electronics Engineer)
Visit my home page: http://welcome.to/SetSoft or
http://www.geocities.com/SiliconValley/Vista/6552/
Alternative e-mail: se...@co... se...@ie... 
Address: Curapaligue 2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA Phone: +(5411) 4759 0013

Re: DJGPP - Configure errors

[Salvador E. T. <sal...@in...>]

Thiago F.G. Albuquerque wrote:

> On Mon, 11 Jul 2005, Salvador Eduardo Tropea wrote:
>
>>> Looking for install tool:Unable to find 'install' on this system.
>>> Please install it and be sure it's in your path.
>>> Also use `install' or `ginstall' name for the binary.
>>> Died at conflib.pl line 1464.
>>
> [snip]
>
>> Right, I fixed it now. The problem was "injected" when I copied some 
>> code from TV's conflib.pl. Now is fixed on CVS.
>
>
> Thank you, now 'configure' is ok. But I still have problems:
>
> 1)
>
[snip]

> ../infview/infbase.cc:1500: request for member `str' in `InfoPath', 
> which is of non-aggregate type `DynStrCatStruct *'

[snip]

> These I fixed with the patch attached.


Ok, applied. Not tested because that's djgpp code.

> 2)

[snip]

> obj/bufun.o(.text+0x180d):bufun.cc: undefined reference to 
> `TVIntl::getTextNew(char const *, bool)'

[snip]

> I didn't try to fix this one. Maybe it has to do with the fact that I 
> don't have gettext and so internationalization is disabled?


Most probably. But what I don't understand is ...

> I configure'd TV with the --no-intl option.


... that in this case include/tv/nointl.h is used where:

 static char       *getTextNew(const char *msgid, Boolean onlyIntl=False)
                    { return onlyIntl ? NULL : newStr(msgid); };

is defined. That's an inline member so I don't see how can be undefined.
 From your mail it looks like you are using -O2 so I don't see how an 
inline function could be missing.
Please play a little bit with include/tv/nointl.h to ensure that's the 
one included. You could also edit classes/tvintl.cc to add "non inline" 
copies of these members.

BTW which gcc and binutils versions are you using?

SET

-- 
Salvador Eduardo Tropea (SET). (Electronics Engineer)
Visit my home page: http://welcome.to/SetSoft or
http://www.geocities.com/SiliconValley/Vista/6552/
Alternative e-mail: se...@co... se...@ie... 
Address: Curapaligue 2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA Phone: +(5411) 4759 0013