setedit-users — General list about setedit

Annoying beep

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

In the old days, the message window beeped only when there were no more 
errors. Now, it beeps everytime the compilations ends. Is there a way to 
disable this?

Thiago

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

more lisp questions: bell

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

is there an slisp command to produce a beep from a macro??

I am looking for some 'alert' simpler than the message box.

Arnold

Re: sLisp: undefined variable 'edfWinPrj'

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

Thiago F.G. Albuquerque wrote:

> This sLisp code, which was working until the last version
>
> (SelectWindowNumber edfWinPrj)
>
> now gives the following error:
>
> Error in sLisp interpreter:
> Type of error: syntax
> undefined variable
> Code: ...(SelectWindowNumber edfWinPrj<*>)...
> End of SLP
>
> I noticed that now there are new commands to select the "special 
> windows" (debugger, message, project, etc.) and I modified my 
> menubind.smn to use them. But still, why did the sLisp code break?

Oops! The following patch fixes it:

Index: mainsrc/edconst.cc
===================================================================
RCS file: /cvsroot/setedit/setedit/mainsrc/edconst.cc,v
retrieving revision 1.51
diff -u -r1.51 edconst.cc
--- mainsrc/edconst.cc  6 Sep 2004 16:25:32 -0000       1.51
+++ mainsrc/edconst.cc  23 Sep 2004 19:42:14 -0000
@@ -428,12 +428,12 @@
 {"WinASCII"       ,-dktASCII          }, // SelectWindowNumber constants
 {"WinCalendar"    ,-dktCalendar       },
 {"WinClipboard"   ,-dktClipboard      },
+{"WinDebugger"    ,-dktDbgMsg         },
 {"WinHelp"        ,-dktHelp           },
 {"WinMP3"         ,-dktMP3            },
 {"WinMan"         ,-dktMan            },
 {"WinMessage"     ,-dktMessage        },
 {"WinPrj"         ,-dktPrj            },
-{"WinDebugger"    ,-dktDbgMsg         },
 {"WinWatches"     ,-dktDbgWt          }
 };
 

-------

Wrong cut & paste ;-)

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: lisp questions

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

Arnold Wiegert wrote:

> Thank you, Salvador.
>
> Salvador Eduardo Tropea wrote:
>
>> Arnold Wiegert wrote:
>
>>
>>> My question is: which editor variables - if any - are available in=20
>>> lisp, particularly, is information about the type of file, i.e. C or =

>>> assembler, or lisp which is being edited available.
>>
>>
> I suppose these variables are accessible via getenv() from sLisp; are=20
> there any exceptions?


Read more, you have a function to get this value.

>>
>> All the sLisp documentation can be accessed from the editor=B4s help. =

>> The chapter is called "sLisp macros".
>> MSVC can=B4t generate the docs but you can get them from the binary=20
>> distribution.
>> You can know which syntax highlight is currently used.
>>
>
> Ok, I've found it and looked some more through it; it will take a fair =

> bit more thought, before I delve into all of that :-)
>
>>> Can I inspect the current editor buffer from lisp? and a whole lot=20
>>> more.
>>
>>
> I guess I'm still too much influenced by my old editor, MultiEdit.
> It seemed most of the work was done through an extensive body of=20
> macros; the main program only seemed to provide support for the macro=20
> language.


That=B4s a nice approach, Emacs was designed like this.
The drawback is that it usually consumes much more resources and takes=20
much more time to develope the core.


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

sLisp: undefined variable 'edfWinPrj'

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

This sLisp code, which was working until the last version

(SelectWindowNumber edfWinPrj)

now gives the following error:

Error in sLisp interpreter:
Type of error: syntax
undefined variable
Code: ...(SelectWindowNumber edfWinPrj<*>)...
End of SLP

I noticed that now there are new commands to select the "special windows" 
(debugger, message, project, etc.) and I modified my menubind.smn to use 
them. But still, why did the sLisp code break?

[],
Thiago

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

Re: lisp questions

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

Thank you, Salvador.

Salvador Eduardo Tropea wrote:
> Arnold Wiegert wrote:
> 
>> One of the things I miss most is the handling of comment lines as in
>> /*
>>  * ----------- I would like an <Enter> here  to automatically add a 
>> space and an asterisk on the next line, so I can continue with the 
>> comment text.
>>
> 
> This will be complex to achieve, but I think is possible.
My preference would be to do it via macros since then it can be more 
versatile for other languages, although most seem to have only single 
line comments, so that may not be a big incentive.
> 
>> My question is: which editor variables - if any - are available in 
>> lisp, particularly, is information about the type of file, i.e. C or 
>> assembler, or lisp which is being edited available.
> 
I suppose these variables are accessible via getenv() from sLisp; are 
there any exceptions?
> 
> 
> All the sLisp documentation can be accessed from the editor´s help. The 
> chapter is called "sLisp macros".
> MSVC can´t generate the docs but you can get them from the binary 
> distribution.
> You can know which syntax highlight is currently used.
> 
Ok, I've found it and looked some more through it; it will take a fair 
bit more thought, before I delve into all of that :-)

>> Can I inspect the current editor buffer from lisp? and a whole lot more.
> 
I guess I'm still too much influenced by my old editor, MultiEdit.
It seemed most of the work was done through an extensive body of macros; 
the main program only seemed to provide support for the macro language.
The worst problems with it, was that is was similar to some current 
languages, but different enough to trip you up.

> You can also consult the macros.slp file to see some examples.
I had looked through that some time ago, but did not really find what I 
was looking for ;-)

Again, thank you for your help; I'll have to study the code a lot more 
and see what I can learn from that.

Arnold

Re: lisp questions

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

Arnold Wiegert wrote:

> Hi,
> now that we have the UNC stuff for MSVC cleared up, I would like=20
> pursue some of the work I was looking at before that. In particular,=20
> I'd like to find out more  the lisp interpreter and how to use it to=20
> write some macros for C specific work.
>
>
> One of the things I miss most is the handling of comment lines as in
> /*
>  * ----------- I would like an <Enter> here  to automatically add a=20
> space and an asterisk on the next line, so I can continue with the=20
> comment text.
>
> From what I have seen, this work seems to be done in the editor code=20
> itself. What I was hoping/looking for was to be able to do such work=20
> from macros, so that they could do some reasonably intelligent work.


This will be complex to achieve, but I think is possible.

> My question is: which editor variables - if any - are available in=20
> lisp, particularly, is information about the type of file, i.e. C or=20
> assembler, or lisp which is being edited available.


All the sLisp documentation can be accessed from the editor=B4s help. The=
=20
chapter is called "sLisp macros".
MSVC can=B4t generate the docs but you can get them from the binary=20
distribution.
You can know which syntax highlight is currently used.

> Can I inspect the current editor buffer from lisp? and a whole lot more=
=2E


Not really. In this case you can select the text before the cursor, copy =

it to the clipboard and then get the clipboard content. You can also=20
know if the cursor is inside a comment.

You can also consult the macros.slp file to see some examples.

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

lisp questions

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

Hi,
now that we have the UNC stuff for MSVC cleared up, I would like pursue 
some of the work I was looking at before that. In particular, I'd like 
to find out more  the lisp interpreter and how to use it to write some 
macros for C specific work.


One of the things I miss most is the handling of comment lines as in
/*
  * ----------- I would like an <Enter> here  to automatically add a 
space and an asterisk on the next line, so I can continue with the 
comment text.

 From what I have seen, this work seems to be done in the editor code 
itself. What I was hoping/looking for was to be able to do such work 
from macros, so that they could do some reasonably intelligent work.

My question is: which editor variables - if any - are available in lisp, 
particularly, is information about the type of file, i.e. C or 
assembler, or lisp which is being edited available.

Can I inspect the current editor buffer from lisp? and a whole lot more.

Is there a manual (page) for this flavor of lisp that I haven't found yet?

I have looked through the code - particularly while debugging the UNC 
stuff and other things, but have not really got to the point where I 
understand it it well enough on my own. So, any hints or corrections are 
most welcome.

TIA,
Arnold

Re: New snapshot available

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

Thiago F.G. Albuquerque wrote:

> On Wed, 15 Sep 2004, Salvador Eduardo Tropea wrote:
>
> [snip]
>
>> Then you won=B4t get debug features.
>> Try removing line 340 of config.pl (unlink $ErrorLog;), run the=20
>> configure script and then take a look at "errormsg.txt" file.
>> Where did you install libmigdb? If you used a path other than the=20
>> default one you=B4ll most probably need to indicate it using the C and=
=20
>> C++ flags.
>
>
> I found the problem: there were two versions of libmigdb installed:=20
> one in
>
> /usr/include
> /usr/lib
>
> (the new one) and the other in
>
> /usr/local/include
> /usr/local/lib
>
> (the old one)
>
> That's because the default install prefix for libmigdb changed in this =

> last release.
>
> The old version probably was before in the include path, because=20
> configure was finding it first. And since it was < 0.8.6...
>
> So I deleted the old version, rebuilt setedit and now it's working :)
>
> Some notes:
>
> 1) In my distribuition (RH 7.3), the prefix for libmigdb *must be*=20
> "/usr/local" (setedit doesn't compile if it's in "/usr").


Then you have to do "make PREFIX=3D/usr/local install" when installing.
But I can=B4t beleive that=B4s really needed. All the libraries you insta=
ll=20
from RPMs will be installed in /usr/lib and its headers in /usr/include. =

So I think your system searches in those directories by default.

> 2) For some reason, these directories in makes/linux
>
> drwxr-xr-x    7 root     root         4.0k Sep 18 01:38 infview-0.2.8/
> drwxr-xr-x    2 root     root         4.0k Sep 18 01:38 result/
> drwxr-xr-x    2 root     root         4.0k Sep 18 01:38 resultInf/
> drwxr-xr-x    8 root     root         4.0k Sep 18 01:38 setedit-0.5.4/
>
> belong to root, so "make clean" fails to remove them.

Yes, they are created during the installation process. Just do the=20
clean-up using the same user that installed things.
Note that usually you won=B4t want to erase these files, they are like a =

cache for the next installation.
This annoyance comes from the fact that the install target is in fact=20
the distribution target plus running the installation script from the=20
distribution.
I don=B4t plan to do it in a different way in the near future.

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: New snapshot available

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

On Wed, 15 Sep 2004, Salvador Eduardo Tropea wrote:

[snip]

> Then you won=B4t get debug features.
> Try removing line 340 of config.pl (unlink $ErrorLog;), run the configure=
=20
> script and then take a look at "errormsg.txt" file.
> Where did you install libmigdb? If you used a path other than the default=
 one=20
> you=B4ll most probably need to indicate it using the C and C++ flags.

I found the problem: there were two versions of libmigdb installed: one in

/usr/include
/usr/lib

(the new one) and the other in

/usr/local/include
/usr/local/lib

(the old one)

That's because the default install prefix for libmigdb changed in this=20
last release.

The old version probably was before in the include path, because configure=
=20
was finding it first. And since it was < 0.8.6...

So I deleted the old version, rebuilt setedit and now it's working :)

Some notes:

1) In my distribuition (RH 7.3), the prefix for libmigdb *must be*=20
"/usr/local" (setedit doesn't compile if it's in "/usr").

2) For some reason, these directories in makes/linux

drwxr-xr-x    7 root     root         4.0k Sep 18 01:38 infview-0.2.8/
drwxr-xr-x    2 root     root         4.0k Sep 18 01:38 result/
drwxr-xr-x    2 root     root         4.0k Sep 18 01:38 resultInf/
drwxr-xr-x    8 root     root         4.0k Sep 18 01:38 setedit-0.5.4/

belong to root, so "make clean" fails to remove them.

[],
Thiago

Re: Sept 15 snapshot & MSVC 6.0

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

Arnold Wiegert wrote:

> Just spent some time yesterday trying to compile the latest snapshot.
>
> After some fiddling, I got as far as:
>
> Linking...
>    Creating library .\Debug/set.lib and object .\Debug/set.exp
> LINK : warning LNK4049: locally defined symbol "_pcre_malloc" imported
> holidays.obj : error LNK2001: unresolved external symbol "char *=20
> __cdecl dlerror(void)" (?dlerror@@YAPADXZ)
> holidays.obj : error LNK2001: unresolved external symbol "void __cdecl =

> dlclose(void *)" (?dlclose@@YAXPAX@Z)
> holidays.obj : error LNK2001: unresolved external symbol _CLY_snprintf
> holidays.obj : error LNK2001: unresolved external symbol "void *=20
> __cdecl dlsym(void *,char const *)" (?dlsym@@YAPAXPAXPBD@Z)
> holidays.obj : error LNK2001: unresolved external symbol "void *=20
> __cdecl dlopen(char *,int)" (?dlopen@@YAPAXPADH@Z)

Hmmm ... Nothing changed in this source to produce it.
If the compiler knows how to optimize things like "if (0) { ... }" these =

reference should disappear.
Make sure the code is using the right configuration header.
If the compiler doesn=B4t know how to optimize "if (0) { ... }" try=20
changing line 55 of setedit/holidays.cc from:

#ifdef SEComp_BCPP

to:

#if defined(SEComp_BCPP) || defined(SEComp_MSC)

> .\Debug/set.exe : fatal error LNK1120: 5 unresolved externals
> Error executing link.exe.
>
> set.exe - 6 error(s), 1 warning(s)
>
> Once I get the linker resolved, I'll have a look at resolving the=20
> snprintf problem.
>
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D=3D=3D
> other fixes:
>
> tvision\classes\tdirlist.cc in
> static
> TStringCollection *ListDirectory( char *path, char *end )
> {
> at line 360:
>
>     return col;
>  =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D

Thanks, I modified too much things there and they are compiler dependant.=


> easydiag\tsviewco.cpp(49): w=3Dmax((size_t)w,strlen(d->getTitle(0))+14)=
;


> set\infview\manview.cpp(148):       int len=3Dmin((int)l->len,w);
> set\infview\manview.cpp(195):    int  xm=3Dmin((int)(l->len),xe);


Please double check those 3, they are fixed in the tarball.

> set\setedit\pathlist.cpp(280):    int l=3Dmin(strlen(str),(size_t)avail=
);

Ok, I=B4m applying this way:
int l=3Dmin((int)strlen(str),avail);

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

Sept 15 snapshot & MSVC 6.0

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

Just spent some time yesterday trying to compile the latest snapshot.

After some fiddling, I got as far as:

Linking...
    Creating library .\Debug/set.lib and object .\Debug/set.exp
LINK : warning LNK4049: locally defined symbol "_pcre_malloc" imported
holidays.obj : error LNK2001: unresolved external symbol "char * __cdecl 
dlerror(void)" (?dlerror@@YAPADXZ)
holidays.obj : error LNK2001: unresolved external symbol "void __cdecl 
dlclose(void *)" (?dlclose@@YAXPAX@Z)
holidays.obj : error LNK2001: unresolved external symbol _CLY_snprintf
holidays.obj : error LNK2001: unresolved external symbol "void * __cdecl 
dlsym(void *,char const *)" (?dlsym@@YAPAXPAXPBD@Z)
holidays.obj : error LNK2001: unresolved external symbol "void * __cdecl 
dlopen(char *,int)" (?dlopen@@YAPAXPADH@Z)
.\Debug/set.exe : fatal error LNK1120: 5 unresolved externals
Error executing link.exe.

set.exe - 6 error(s), 1 warning(s)

Once I get the linker resolved, I'll have a look at resolving the 
snprintf problem.

=================================
other fixes:

tvision\classes\tdirlist.cc in
static
TStringCollection *ListDirectory( char *path, char *end )
{
at line 360:

     return col;
  ===========================
easydiag\tsviewco.cpp(49): w=max((size_t)w,strlen(d->getTitle(0))+14);
set\infview\manview.cpp(148):       int len=min((int)l->len,w);
set\infview\manview.cpp(195):    int  xm=min((int)(l->len),xe);
set\setedit\pathlist.cpp(280):    int l=min(strlen(str),(size_t)avail);

Arnold

Re: New snapshot available

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

Thiago F.G. Albuquerque wrote:

> On Tue, 14 Sep 2004, Salvador Eduardo Tropea wrote:
>
>> Thiago F.G. Albuquerque wrote:
>>
>>> Hi,
>>>
>>> I downloaded, compiled and installed these latest snapshots, but the =

>>> debug stuff is not working. The debug menu was not showing, so I=20
>>> figured: "the variable DEBUG in menubind.smn must be set to false".=20
>>> I erased the enclosing "$if DEBUG/$endif" pair, so now the menu=20
>>> shows, but the commands do nothing.
>>>
>>> Is there a special compilation option to enable this? I lookup up in =

>>> the docs, but didn't find anything.
>>
>>
>> Did you install libmigdb?
>
>
> Yes.
>
> I found these lines in configure.cache:
>
> migdbShipped =3D> no
> HAVE_GDB_MI =3D> no


Then you won=B4t get debug features.
Try removing line 340 of config.pl (unlink $ErrorLog;), run the=20
configure script and then take a look at "errormsg.txt" file.
Where did you install libmigdb? If you used a path other than the=20
default one you=B4ll most probably need to indicate it using the C and C+=
+=20
flags.

> Also, some examples from libmigdb are not compiling. I am sending the=20
> output of make attached.

Don=B4t worry about them, I don=B4t check the examples very often.

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: New snapshot available

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

On Tue, 14 Sep 2004, Salvador Eduardo Tropea wrote:

> Thiago F.G. Albuquerque wrote:
>
>> Hi,
>> 
>> I downloaded, compiled and installed these latest snapshots, but the debug 
>> stuff is not working. The debug menu was not showing, so I figured: "the 
>> variable DEBUG in menubind.smn must be set to false". I erased the 
>> enclosing "$if DEBUG/$endif" pair, so now the menu shows, but the commands 
>> do nothing.
>> 
>> Is there a special compilation option to enable this? I lookup up in the 
>> docs, but didn't find anything.
>
> Did you install libmigdb?

Yes.

I found these lines in configure.cache:

migdbShipped => no
HAVE_GDB_MI => no

Also, some examples from libmigdb are not compiling. I am sending the 
output of make attached.

Thiago

Re: New snapshot available

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

Thiago F.G. Albuquerque wrote:

> Hi,
>
> I downloaded, compiled and installed these latest snapshots, but the 
> debug stuff is not working. The debug menu was not showing, so I 
> figured: "the variable DEBUG in menubind.smn must be set to false". I 
> erased the enclosing "$if DEBUG/$endif" pair, so now the menu shows, 
> but the commands do nothing.
>
> Is there a special compilation option to enable this? I lookup up in 
> the docs, but didn't find anything.

Did you install libmigdb?

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: New snapshot available

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

Hi,

I downloaded, compiled and installed these latest snapshots, but the debug=
=20
stuff is not working. The debug menu was not showing, so I figured: "the=20
variable DEBUG in menubind.smn must be set to false". I erased the=20
enclosing "$if DEBUG/$endif" pair, so now the menu shows, but the commands=
=20
do nothing.

Is there a special compilation option to enable this? I lookup up in the=20
docs, but didn't find anything.

TIA,
Thiago

--=20
()  ascii ribbon campaign - against html mail=20
/\                        - against microsoft attachments


On Fri, 10 Sep 2004, Salvador Eduardo Tropea wrote:

> Hi All!
>
> The version of libmigdb needed for current editor is 0.8.6. Now the confi=
gure=20
> script is checking its version
>
> http://prdownloads.sourceforge.net/libmigdb/libmigdb-0.8.6.tar.bz2?downlo=
ad
> http://sourceforge.net/project/showfiles.php?group_id=3D114758&package_id=
=3D124329&release_id=3D267129
>
> The snapshots are:
>
> http://setedit.sourceforge.net/snap.html
> and
> http://tvision.sourceforge.net/snap.html
>
> This new snapshot contains no new features, just fixes and implementation=
 of=20
> options that didn=B4t work.
> Most debug features are documented and have context sensitive help.
> You can take it as a first release candidate.
>
> SET
>
> P.S. This snapshot have SPARC support.
>
>

New snapshot available

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

Hi All!

The version of libmigdb needed for current editor is 0.8.6. Now the=20
configure script is checking its version

http://prdownloads.sourceforge.net/libmigdb/libmigdb-0.8.6.tar.bz2?downlo=
ad
http://sourceforge.net/project/showfiles.php?group_id=3D114758&package_id=
=3D124329&release_id=3D267129

The snapshots are:

http://setedit.sourceforge.net/snap.html
and
http://tvision.sourceforge.net/snap.html

This new snapshot contains no new features, just fixes and=20
implementation of options that didn=B4t work.
Most debug features are documented and have context sensitive help.
You can take it as a first release candidate.

SET

P.S. This snapshot have SPARC support.

--=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... Address: Curapaligue=20
2124, Caseros, 3 de Febrero
Buenos Aires, (1678), ARGENTINA Phone: +(5411) 4759 0013

Re: more MSVC6 changes

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

Arnold Wiegert wrote:

> In the rush, I forgot one other change I had to make in tvision code
>
> MSVC 6.0 defines _MSC_VER as 1200
>
[snip]

Ok, but: Why MSVC 6.0 can=B4t compile 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: Aug 24 snapshot under BCC 5.5

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

Arnold Wiegert wrote:

> Just compiled  the latest snapshot under BCC 5.5
>
> just a small thing I noticed - besides the new 'Glaciar P. Moreno' ;-)
>
> the top line of the about box when the program first signs on is too=20
> long to fit all of the copyright date. The last digit is dropped.
>
> The regular 'About' box is ok.

That=B4s strange, this box should adjust its size. I=B4ll try to check it=
=20
next time I run the BC++ compiled binary.

Thanks, 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

Aug 24 snapshot under BCC 5.5

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

Just compiled  the latest snapshot under BCC 5.5

just a small thing I noticed - besides the new 'Glaciar P. Moreno' ;-)

the top line of the about box when the program first signs on is too 
long to fit all of the copyright date. The last digit is dropped.

The regular 'About' box is ok.

Arnold

Re: WheelMouse not supported under Windows NT, WinMe/98/95

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

Thiago F.G. Albuquerque wrote:
> On Thu, 26 Aug 2004, Arnold Wiegert wrote:
> 
> [snip]
> 
>> So even though, MFC supports the mouse wheel, the Win32 API does NOT!!!!
> 
> 
> Why don't you have a look at the sources of MFC?
> 
> My guess is that mouse wheel actions are translated into ordinary 
> scrollbar messages.
> 
> This way, the support for the wheel does not require any changes to the 
> Win32 API and even old aplications respond to the wheel unmodifed.
> 
> The question is: does SETEDIT know about scrollbar messages? If not, 
> then this might be the source of your problem.
> 
> Thiago
> 
I had thought along those lines, or something similar to the existing 
X11 support.
Tvision does know about scroll bar messages, but unless we can get those 
wheel messages we're stuck.

As far as MFC goes ...
a) the application I have to test is not an MFC app.
b) I don't have the MFC sources installed at present and I don't know - 
haven't checked yet - whether they are part of my CDs
c) even if they are, it means I have to rearrange my systems again to 
install them - I've had to jump through a lot of hoops to get this far :-(
d) even if I do, I have no guarantee that I'll be able to get the 
results I need and want :-(

At present, the gains do not seem worth the pain.

The app I have - a sample from MSVC 6 - is a plain old windows app which 
handles WM_MOUSEWHEEL messages; they do arrive as part of the regular 
WM_xxx messages. This app does nothing but display the fact that the 
message arrived.
 From my tests, these same messages do not arrive via the existing call 
to ReadConsoleInput() of tvision, even though the other mouse events 
arrive via that call.

While looking through the code just now, I realized that this interface 
does support
fHasWheel = GetSystemMetrics( SM_MOUSEWHEELPRESENT );

Under MSVC for the console app, I had to do a lot of digging and 
cheating just to get it to compile that one line.

So for now, it looks like I'll have to suspend work on this and get on 
with some other things.

Arnold

Re: WheelMouse not supported under Windows NT, WinMe/98/95

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

On Thu, 26 Aug 2004, Arnold Wiegert wrote:

[snip]
> So even though, MFC supports the mouse wheel, the Win32 API does NOT!!!!

Why don't you have a look at the sources of MFC?

My guess is that mouse wheel actions are translated into ordinary 
scrollbar messages.

This way, the support for the wheel does not require any changes to the 
Win32 API and even old aplications respond to the wheel unmodifed.

The question is: does SETEDIT know about scrollbar messages? If not, then 
this might be the source of your problem.

Thiago

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

Re: WheelMouse not supported under Windows NT, WinMe/98/95

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

Hi Arnold:

Arnold Wiegert wrote:

> I had been trying to sort out implementation of the wheel mouse for=20
> the Win32 version of setedit, but it looks like it is a dead end until =

> I upgrade my OSes :-/
>
> Just - finally  - came across this notice on MSDN at
> http://msdn.microsoft.com/library/default.asp?url=3D/library/en-us/dllp=
roc/base/mouse_event_record_str.asp=20
>
>
> excerpt:
>
> dwEventFlags
>     Type of mouse event. If this value is zero, it indicates a mouse=20
> button being pressed or released. Otherwise, this member is one of the =

> following values.
>     Value     Meaning
>     DOUBLE_CLICK     The second click (button press) of a double-click =

> occurred. The first click is returned as a regular button-press event.
>     MOUSE_MOVED     A change in mouse position occurred.
>     MOUSE_WHEELED     The mouse wheel was rolled.
>
> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++=
++
>         Windows NT and Windows Me/98/95:  This value is not supported.
> +++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++=
++
>
> So even though, MFC supports the mouse wheel, the Win32 API does NOT!!!=
!


My Win32 API reference says something similar. But IE supports the mouse =

wheel in Win98SE. It means that even when the API doesn=B4t support wheel=
=20
events as mouse events it have some support. I think that=B4s something=20
independent with a different mechanism. What I don=B4t know if that=B4s s=
ome=20
"undocumented" stuff (MS likes to do it to have advantage over other=20
manufacturers) or just some extra DLL provided in one of the SDKs.

> It pretty well looked like that when I started intercepting the mouse=20
> events - none of the wheel events triggered the interrupts at the=20
> application level - and this makes it more or less official.


I didn=B4t analyzed how IE gets access to it, but it does, and my mouse=20
isn=B4t an MS mouse ...
In fact I think I never installed any kind of extra mouse driver.

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

WheelMouse not supported under Windows NT, WinMe/98/95

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

I had been trying to sort out implementation of the wheel mouse for the 
Win32 version of setedit, but it looks like it is a dead end until I 
upgrade my OSes :-/

Just - finally  - came across this notice on MSDN at
http://msdn.microsoft.com/library/default.asp?url=/library/en-us/dllproc/base/mouse_event_record_str.asp

excerpt:

dwEventFlags
     Type of mouse event. If this value is zero, it indicates a mouse 
button being pressed or released. Otherwise, this member is one of the 
following values.
     Value 	Meaning
     DOUBLE_CLICK 	The second click (button press) of a double-click 
occurred. The first click is returned as a regular button-press event.
     MOUSE_MOVED 	A change in mouse position occurred.
     MOUSE_WHEELED 	The mouse wheel was rolled.

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
         Windows NT and Windows Me/98/95:  This value is not supported. 

+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++

So even though, MFC supports the mouse wheel, the Win32 API does NOT!!!!

It pretty well looked like that when I started intercepting the mouse 
events - none of the wheel events triggered the interrupts at the 
application level - and this makes it more or less official.

Arnold

Re: compile under MSVC 6.0

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

Arnold Wiegert wrote:

> Finally got a compile under MSVC 6.0 with the following changes:
> I believe I was using the Aug 24 snapshot.
>
> in compatlayer.h
> #ifdef Uses_snprintf
> #pragma message("Uses_snprintf")
>  #ifdef CLY_Have_snprintf
>   #undef  Include_stdio
>   #define Include_stdio 1
>  #if 1                        // AWW
>   #ifdef TVComp_MSC                // AWW
>    #pragma message("TVComp_MSC")        // AWW
>    #if _MSC_VER =3D=3D1200                // AWW
>     #pragma message("_MSC_VER=3D=3D1200")    // AWW
>     #define CLY_snprintf  _snprintf    // AWW
>    #else                        // AWW
>     #define CLY_snprintf  snprintf    // AWW
>    #endif    // 1200                // AWW
>  #else                        // AWW
>     #define CLY_snprintf  snprintf        // AWW
>   #endif    // TVComp_MSC            // AWW
>  #endif                        // AWW
>   #ifdef TVComp_MSC
>    #if _MSC_VER>=3D1200 // from MSVC .NET first release - v13.0. AWW
>       #define CLY_vsnprintf _vsnprintf
>    #else
>       #define CLY_vsnprintf vsnprintf
>    #endif
>   #else
>    #define CLY_vsnprintf vsnprintf
>   #endif
>  #else
>   #undef  Include_stdarg
>   #define Include_stdarg 1
>  #endif
> #endif


More about it at the end of the message.

> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D
> :\set-test\snapjuly11\msvc6\set\easydiag\tsviewco.cpp(49) : error=20
> C2782: 'T __cdecl max(const T &,const T &)' : template parameter 'T'=20
> is ambiguous
>         could be 'unsigned int'
>         or       'int'
> changed to make it compile:
>  w=3Dmax((unsigned int)w,strlen(d->getTitle(0))+14);    // AWW
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D
> d:\set-test\snapjuly11\msvc6\set\infview\manview.cpp(148) : error=20
> C2782: 'T __cdecl min(const T &,const T &)' : template parameter 'T'=20
> is ambiguous
>         could be 'int'
>         or       'unsigned int'
> d:\set-test\snapjuly11\msvc6\set\infview\manview.cpp(195) : error=20
> C2782: 'T __cdecl min(const T &,const T &)' : template parameter 'T'=20
> is ambiguous
>         could be 'int'
>         or       'unsigned int'
> changed to make it compile:
> 148:
>        int len=3Dmin((int)l->len,w);    // AWW
> 195:
>    int  xm=3Dmin((int)l->len,xe);        // AWW
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D
> :\set-test\snapjuly11\msvc6\set\setedit\calendar.cpp(144) : error=20
> C2782: 'T __cdecl max(const T &,const T &)' : template parameter 'T'=20
> is ambiguous
>         could be 'int'
>         or       'unsigned int'
>
> changed to make it compile:
>  int l=3Dmax((int)strlen(monthName),15)+5;    // AWW
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D


Ok, I see the compiler is quite picky about the max/min arguments.

> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D
> Once it compiles without errors, it throws an assert in the call=20
> marked below
>
> ushort messageBox( ushort aOptions, const char *fmt, ... )
> {
>     va_list argptr;
>
>     char *intlFmt=3DTVIntl::getTextNew(fmt);
>
>     va_start( argptr, fmt );                                //asserts=20
> in this call
>     int l=3DCLY_vsnprintf(NULL, (size_t)0, intlFmt, argptr); <=3D=3D=3D=
=3D=3D=3D=3D=3D=3D=3D
>     va_end( argptr );
>     AllocLocalStr(msg,l+1);
>
>     va_start( argptr, fmt );
>     CLY_vsnprintf(msg, (size_t)l+1, intlFmt, argptr);
>     va_end( argptr );
>
>     DeleteArray(intlFmt);
>     return messageBoxRect( makeRect(), msg, aOptions | mfDontTranslate =
);
> }
>
> =3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=3D=
=3D=3D=3D=3D=3D=3D
> I don't quite understand this call at all and am not suprised - with=20
> the way I see it at present - that it does assert.


If it asserts there then we must consider using the snprintf=20
replacement. The purpose of this call is to find the size of the buffer.
It looks like Microsoft implemented it in the wrong way. The Open Group=20
page for it is:

http://www.opengroup.org/onlinepubs/009695399/functions/snprintf.html

Note that=B4s POSIX/IEEE 1003.1 standard and it clearly says:

=2E.. "If /n/ is zero, nothing shall be written and /s/ may be a null=20
pointer." ...

And when the result value is explained says:

"Upon successful completion, the /snprintf/() function shall return the=20
number of bytes that would be written to /s/ had /n/ been sufficiently=20
large excluding the terminating null byte."

As you can see the pointer can be NULL if the size is 0. And the return=20
value is the len of the text (no EOS counted), the full text.
In this way the first call finds the len, then we allocate len+1 in the=20
stack and call the function again, this time we will get all the text=20
and we can=B4t damage the stack. Robert=B4s code (he added this nice vers=
ion=20
of messageBox) had a fixed buffer of 256 bytes, but a file name in Linux =

can be upto 4096 bytes long, so we could easily destroy the stack in=20
message box naming a file.

So I=B4ll assume Microsoft implementation is useless until somebody can=20
tell me when they fixed it.

> As far as I remember, those are the only items I had to change.

Thanks!

--=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