lineak-developers — Anyone can use this list to contact all of the LinEAK developers at once

[Lineak-developers] [PATCH] LinEAKd and XOSD default configuration

[Aurelien J. <aur...@au...>]

Hi,

I have just tried the Lineakd version 0.7, and I found a bug in the
configuration loader: when not specified in the configuration file,
XOSD_timeout, XOSD_hoffset, XOSD_voffset and XOSD_soffset are
initialized to 0 instead of the default value specified in
definitions.h. This is not very important for the last 3 one, but having
XOSD_timeout set to 0 prevent OSD messages to be hide.

The attached patch fix the problem.

Cheers,
Aurelien

[Lineak-developers] lineakkb.dev section for Acer Aspire 1312 laptop

[Guy V. D. B. <su...@li...>]

Hi guys,

First of all: thanks for the great work with lineak!

I have added a section to the /etc/lineakkb.def file for the extra keys I
have on my Acer Aspire laptop. Here it is:

<begin section>
[AALK1312]
  brandname = "Acer"
  modelname = "Laptop Keyboard (Aspire 1312LC)"
  [KEYS]
    Mute                        = 160
    VolumeUp                    = 176
    VolumeDown                  = 174

    Mail                        = 236
    Internet                    = 178
    P1                          = 244
    P2                          = 243
  [END KEYS]
[END AALK1312]
# end Acer Aspire Keyboard 1312LC
</end section>

I thought it might be useful for some other people runnin linux on their
Aspires.

Greets,
Guy

Re: [Lineak-developers] LinEAKD is segfaulting when xosd fails to initialize

[Sheldon Lee-W. <to...@ro...>]

Hi all,

> A Debian user found a problem in LinEAKD when it is used with xosd. If
> xosd fails to initialize, which occurs for example if there is no font
> available for the current configuration (see my previous email), LinEAKD
> segfaults when pressing a key.

Lol. I encountered this just the other day as well.

> This is due to a call to an xosd_ function with the osd pointer equal to
> null. I have seen that the problems have started to be corrected in the
> CVS (via the enabled function of the xosdCtrl class), however the bug is
> still present.

Yup. 

> I have made two patches which fixes the problem with a different
> approach:
>
> - The first one is only fixing some broken parts following the same
>   approache than in the CVS.

Actually, I have already done this, but not committed it.

> - The second patch is radically different. Instead of verifying, before
>   each call to and xosdCtrl function, that xosd is enabled, it makes the
>   object created from xosdCtrl to react the same way if xosd is enabled
>   or not. In fact it is just testing if (osd != NULL) before calling an
>   xosd_ function.

And this is the better way to do it anyway......

> Please feel free to include the patch you want (or even none of them).

Thanks Aurelien, I applied your font patch, and the second xosd segfault 
patch. Much appreciated.

Cheers,
Sheldon.

==================================================================
"... all thoughts of selfish desire, ill-will, hatred and
 violence are the result of a lack of wisdom ... "
 - Buddha

For an awesome fantasy role playing game checkout:
http://lycadican.sourceforge.net

GPG KeyID=04B7F7F8
GPG Fingerprint=4B0F 7202 FAFF D146 5F56  9E83 BE7F D7F7 04B7 F7F8
==================================================================

[Lineak-developers] LinEAKD is segfaulting when xosd fails to initialize

[Aurelien J. <aur...@au...>]

Hello,

A Debian user found a problem in LinEAKD when it is used with xosd. If
xosd fails to initialize, which occurs for example if there is no font
available for the current configuration (see my previous email), LinEAKD
segfaults when pressing a key.

This is due to a call to an xosd_ function with the osd pointer equal to
null. I have seen that the problems have started to be corrected in the
CVS (via the enabled function of the xosdCtrl class), however the bug is
still present.

I have made two patches which fixes the problem with a different
approach:

- The first one is only fixing some broken parts following the same
  approache than in the CVS.

- The second patch is radically different. Instead of verifying, before
  each call to and xosdCtrl function, that xosd is enabled, it makes the
  object created from xosdCtrl to react the same way if xosd is enabled
  or not. In fact it is just testing if (osd != NULL) before calling an
  xosd_ function.

BTW, both patches make LinEAKD more verbose when xosd fails to
initialize.

Please feel free to include the patch you want (or even none of them).

Regards,
Aurelien

[Lineak-developers] LinEAKD and xosd font

[Aurelien J. <aur...@au...>]

[CCing lineak-devel as it can interest some people]

Hello,

On some configurations, xosd is not working correctly with LinEAKD. 
This is due to the default font which is only available in ISO-8859-1 
encoding. The problem concern, at least, most of the people of the 
Euroland, who are using an @euro locale that is to say with an 
ISO-8859-15 encoding.

I propose a font, which is maybe not so nice as the default one in
lineakd, but which have the same size as the current one. Moreover, 
it is included by default in XFree, it is available in ISO-8859-1,
-2, -3, -4, -9, -10, -13, -14 and ISO-10646 (Unicode) encodings, and
available in both 75 and 100 dpi. It may be also available in other
encodings by installing some more packages in your distribution. This
font is defined as:

-adobe-helvetica-bold-r-normal-*-*-240-*-*-p-*-*-*

I think it could be also a good idea to change the default font in
LineAKD. With the version currently in the CVS, if a user doesn't like 
it, he can change it in the configuration file. A patch against the 
current CVS version is attached for that. Please feel free to include 
it or not.

Regards,
Aurelien

[Lineak-developers] [BUG] lineakd needs to be restarted after changing the keymap

[Aurelien J. <aur...@au...>]

Hello,

Gabor Fleischer reported me a strange bug about lineakd when using it
with kxkb.

When using mutiple keymaps with kxkb, and switching from a window to
another, kxkb need sometimes to be reloaded or rehash.

I have made some tests, and I have remarked that, when I change the
configuration of the keyboard, lineakd prints out "Reclaiming keyboard
map from MappingNotify event". However when I switch from a keyboard to
another, it doesn't display that.

I don't know if it is a bug in lineakd or kxkb.

Cheers,
Aurelien

Note: I am in vacation for a week and I won't be able to read my emails.

[Lineak-developers] klineakconfig stops whole KDE when exiting

[Aurelien J. <aur...@au...>]

Hello,

This time I have a bug without patch :-(

This strange bug has been reported by Gabor FLEISCHER <de...@fl...>
via the Debian bug tracking system:

| When I run klineakconfig from xterm, and press Quit button on 
| klineakconfig's icon in the panel, it quit's fine. BUT, when I run 
| klineakconfig &, quit KDE, restart it, so that klineakconfig runs 
| automatically, and then I press the Quit button, then it stops the 
| whole KDE!

I have tested that on my computer, it is 100% reproducible. I have
quickly looked at the sources of lineakconfig, I didn't noticed
anything wrong. 

If anybody has an idea...

Cheers,
Aurelien

[Lineak-developers] lineakkb.def in the current CVS is buggy

[Aurelien J. <aur...@au...>]

Hello,

The file lineakkb.def which is in the current CVS of lineakd has a
problem. Two [END] are missing. The following patch correct that:

--- lineakd/lineakkb.def
+++ lineakd/lineakkb.def
@@ -349,6 +349,10 @@
     Mute                  = 160
     Home                  = 178
     Webcamera             = 146
+  [END KEYS]
+[END LTCD]
+# end Logitech Access
+
 [LTCD]
   brandname = "Logitech"
   modelname = "Cordless Desktop"


Regards,
Aurelien

[Lineak-developers] lineak_core_functions::is_running(string process) is buggy

[Aurelien J. <aur...@au...>]

Oops, the patch...

[Lineak-developers] lineak_core_functions::is_running(string process) is buggy

[Aurelien J. <aur...@au...>]

Hello,

It seems that the function lineak_core_functions::is_running(string
process) is buggy, it returns true for every process which contains the
word "lineakd" (for example man lineakd), and not only for the daemon. 

I have rewritten the function (see attached patch against current CVS) 
which get the information from /proc instead of using ps and greping 
the results.

Regards,
Aurelien

[Lineak-developers] Long options doesn't work with the CVS version

[Aurelien J. <aur...@au...>]

Hello,

I have just tried the version of lineakd from CVS, and the long options
don't work. The following patch fixes that.

Regards,
Aurelien

Index: cmdprefs.cpp
===================================================================
RCS file: /cvsroot/lineak/lineakd/lineakd/cmdprefs.cpp,v
retrieving revision 1.2
diff -u -1 -b -p -r1.2 cmdprefs.cpp
--- cmdprefs.cpp        8 Jul 2003 02:18:16 -0000       1.2
+++ cmdprefs.cpp        26 Jul 2003 23:31:42 -0000
@@ -25,2 +25,9 @@

+extern "C" {
+#ifdef HAVE_GETOPT_H
+#  define HAVE_GETOPT_LONG
+#  include <getopt.h>
+#endif
+}
+
 CmdPrefs::CmdPrefs(){
 

[Lineak-developers] Compaq Presario 725EA: Can't get any key...

[Coume <cou...@lu...>]

Hello,
Yesterday and the day before I tried to find out what were the key code of =
the EAK of my compaq presario 725EA but I can't...

I tried the 2 methods mentionned in the website, but none works...
-In the xterm win, none of the EAK key worked (no events shown)
-In trying to select other keyboard, and try them... none worked

Is there a 3 way to find out the key?? I would really like to help about th=
is keyboard... The compaq 725EA is already very hard to configure for being=
 under linux, then if I could help for that I would be happy!

If there is another method, just let me know :)

Regards,
Coume

________________________________________________________________________

Tired of Hotmail, Yahoo, Caramail, etc...?
Go for Runbox and Get:
 - 100MB storage space with options to buy unlimited disk space
 - Send and receive large messages, up to 20 MB each.
 - 10 Mo File Storage with options to buy unlimited disk space
 - Access your email on any device via Web, POP, IMAP, SMTP, WAP, SMS, and =
more....
 - Anti-spam features, Anti-Virus, no-ads, and much more!!!
Give it a try: http://1555.runbox.com/
1 month FREE Trial!!!!
________________________________________________________________________

[Lineak-developers] lineak, KDE, KMilo

[George S. <st...@kd...>]

Hi Guys!
  I just heard about your project.  Great work!  We have needed something like 
this for a while.  I'm writing because I have just checked in some somewhat 
related code to KDE CVS for 3.2.0.  It's called "KMilo" in kdeutils.  It 
provides two things in one kded module:

   - a (soon to be skinnable) on-screen display
   - a (eventually plugin based) backend for special hardware

   I wrote this as a generic solution to the problem I have of a Mac PowerBook 
with special keys for brightness and volume that display onscreen in Gnome 
and Mac OS X, but not KDE.  It now displays onscreen volume changes and 
brightness changes.  I would like to extend this to support other types of 
hardware and special keyboards.  I am reminded of my old (acer-built) IBM 
ThinkPad which had all kinds of hardware keys.

   What do you think about this?  How could our projects complement eachother 
best?  Ideally I would like all visual feedback to go through kmilod since it 
is even possible to trigger it via dcop.  It would be nice to have a central 
module to configure these kinds of things as well.  Is it possible to make a 
kmilo module for back-end integration with lineak?  Or does that not make 
sense?

   Thanks

-- 

George Staikos

[Lineak-developers] New keyboard definition - Compaq Presario 90x

[Aurelien J. <aur...@au...>]

Hi all !

A user (mo...@fr...) send me a new keyboard definition for the Compaq
Presario 90x.

Here is the definition for /etc/lineakbb.def:

[CP90x]
  brandname = "Compaq"
  modelname = "Presario 90x"
  [KEYS]
    internet        = 163
    mail            = 158
    search          = 154
    media           = 239
    volumedown      = 174
    volumeup        = 176
    play            = 162
    stop            = 164
    previous        = 144
    next            = 153
  [END KEYS]
[END CP90x]


And attached the picture for lineakconfig.

Regards,
Aurelien

[Lineak-developers] [Lineakconfig] Segfault of lineakconfig [patch]

[Aurelien J. <aur...@au...>]

Hi,

I have found a bug in lineakconfig. When switching between keyboards,
lineakconfig crash or display erroneous name, depending on the version
of gcc (2.95 or 3.2). However using efence it segfault in both case.

The problem is in select_new_kbmodel(GtkWidget *widget, gpointer data).
data is the name of the keyboard, sent by the callback of the widget. In
fact it is a pointer to the name in the dcfp structure.
Doing myEAK.EAKmodel = (char *)data assign this pointer to
myEAK.EAKmodel. Then when running initEAK(), this variable is remplaced
by a copy of the new name, but it is first freed, that is to say data
from the dcfp structure are freed ! Then it causes a segfault when
switching from keyboards...

The following patch corrects the problem.

Aurelien


diff -u2 -r lineakconfig-0.3.2.orig/src/support.c
lineakconfig-0.3.2/src/support.c
--- lineakconfig-0.3.2.orig/src/support.c       2002-06-11
15:01:50.000000000 +0200
+++ lineakconfig-0.3.2/src/support.c    2003-01-10 00:28:47.000000000
+0100
@@ -329,10 +329,9 @@
   extern gboolean changing_text_from_signal;
   char *temp;
-  myEAK.EAKmodel = (char *)data;

   if (kbtype!=NULL) free(kbtype);

-  if ((temp = lookup_kbtype(myEAK.EAKbrand, myEAK.EAKmodel))==NULL) {
-       g_print("*** PANIC: keyboard of brand '%s' and model '%s' not
        found!\n",myEAK.EAKbrand, myEAK.EAKmodel);
+  if ((temp = lookup_kbtype(myEAK.EAKbrand, (char *)data))==NULL) {
+       g_print("*** PANIC: keyboard of brand '%s' and model '%s' not
found!\n",myEAK.EAKbrand, (char *)data);
        cleanexit();
   } else {

[Lineak-developers] lineak under console?

[Msquared <li...@ms...>]

Hi there!

I was wondering, is it possible to configure lineak so that it works under
consoles?  For example, with an appropriate console keymap, it's possible
to map the windows key to move between consoles.  Can I program any easy
access keys to do similar things?  For example, I'd like to be able to hit
some easy access keys to go to different consoles, instead of alt-F1,
alt-F2, etc...

             2
Regards, /|/|
        /   |

[Lineak-developers] Here is a patch to escape the " character.

[Colin M. <ccm...@ne...>]

Hello again,

	I have completed the patch to escape out the " character in command
lines.  It is now fully operational with lineakd's reading of the config
file as well as lineakconfig (gtk+)'s writing of the config file.  If a
line like this is entered in the lineakconfig box:

mozilla -remote "openurl(http://www.google.com)" 

then when the config file is saved, it will write the command out as
this:

internet = "mozilla -remote \"openurl(http://www.google.com)\""

Thus the " character is escaped transparently for the end user, and I am
able to remote control mozilla from my hotkeys :)

I apologize for sending you the patch in this format, but I really can't
figure out how to get diff working... I try it but it keeps spitting
back the entire file as different, perhaps something to with the editor
I'm using (vi, pico, and kdevelop all produced this result).

Here is the patch:

To the file dcfp.c (located in lineakd-0.4pre1/src/ and
lineakconfig-0.3.2/src/), I add the following lines to the function
dcfp_get_value_from_line:

...
//Approx.  line #234
firstptr = ptr;
if (*ptr == '"') {
   ptr++;
while (*ptr!='\n' && *ptr!='\0' && *ptr!='"') {
//ADDED--v
    if (*ptr=='\\' && *(ptr+1)=='"' && *(ptr+2)!='\n' && *(ptr+2)!='\0')
    ptr++;
//END ADDED--^
	  value = (char *)realloc(value, (i+2)*sizeof(char));
	  value[i] = *ptr;
	  value[i+1] = '\0';
...

To the file support.h (located in lineakconfig-0.3.2/src/), i add the
following line to the declarations:
...
gboolean save_conffile (void);
int file_exists(char *filename);
void delete_file(char *filename);
gboolean get_lineakd_pid(void);
//ADDED--v
gchar *escape_command(char *cmd);
//END ADDED--^

/* available brands */
typedef struct abrands {
  char *name;
...

To the file support.c (located in lineakconfig-0.3.2/src/), i add the
following lines at the end to create the function escape_command:

...
//approx. line #971
  return pspid;
}


/**********************************************************************************************/
//ADDED--v
gchar *escape_command(char *cmd) {
  gchar *escaped, *ptr = NULL;
  int i = 0;

  escaped = (char *)malloc(sizeof(char));
  *escaped = '\0';
  ptr = cmd;

  while (*ptr != '\0')
  {
  if (*ptr == '"') {
  escaped = (char *)realloc(escaped, (i+2)*sizeof(char));
  escaped[i] = '\\';
  escaped[i+1] = '\0';
  i++;
  }
  escaped = (char *)realloc(escaped, (i+2)*sizeof(char));
  escaped[i] = *ptr;
  escaped[i+1] = '\0';
  i++; ptr++;
  }
  return escaped;
}
//END ADDED --^
...

To the same file (support.c in lineakconfig-0.3.2/src/), i further
modify the following line in the save_conffile function:

...
//Approx. Line #910
             "MixerDevice\t\t= %s\n"
               "\n# Specific keys of your keyboard\n", CONF_HEADER,
myEAK.EAKbrand, myEAK.EAKmodel, kbtype, cdromdev!=NULL?cdromdev:"",
mixdev!=NULL?mixdev:"");
  key = myEAK.EAKeylist;
  while (key!=NULL) {
//MODIFY LINE--v
	fprintf(dest,"%s%s=
\"%s\"\n",key->EAkeyname,strlen(key->EAkeyname)>7?"\t":"\t\t",key->EAcommand!=NULL?escape_command(key->EAcommand):"");
//END MODIFY LINE--^
	key = key->next;
  }
...

Thank you for your time, and I hope that this patch may be as useful for
you or other end users as it has proven for myself.  If you do use this,
feel free to modify or fix it as you see fit, just drop my name
somewhere in the source code or documentation or whatever.

Thank you sincerely,
Colin McCambridge

again- I apologize for putting the C code in this email... if you want
me to just send the modified files, if that would be easier, please let
me know.
 

[Lineak-developers] lineakd.conf fix: escape the " character

[Colin M. <ccm...@ne...>]

Hello,

I have a potential cange to submit to your lineak project, specifically
the dcfp module.  I have recently purchased a Microsoft Internet
Keyboard and came across your project on sourceforge for enabling use of
the extended keys.  I wish to thank you wholeheartedly for your efforts
in this area. I really like hotkeys, because I'm too lazy to move my
mouse down to GNOME's menu and navigate the lists to open my browser,
etc ;-)

One problem which I did have in using lineakd lied in configuring
lineakd to control mozilla to navigate to a specific page.  This can be
done with the command:

mozilla -remote "openurl(http://www.google.com)"

or

mozilla -remote "openurl(http://www.google.com,new-window)"

I tried inputing this command into lineakconfig, but the button would
not work.  I examined .lineak/lineakd.conf and determined that the
problem lies in the fact that there was no way to escape the " character
in a command.

I am a fairly rudimental programmer (you may have noticed that I am
emailing this change to you instead of submitting it in whatever
appropriate way things are supposed to be submitted, because I don't
know that appropriate way).  However, I took a look through your source
code and traced to problem to dcfp.c, in the function 

static char *dcfp_get_value_from_line(char *line)

I modified one of the while loops to add an escape sequence for ". 
specifically it will now do this:
After starting a value with ":
    If the parser crosses the sequence \", it will check to see if those
are the last two characters in the value.
        If they are, then nothing different happens.   '\' is added to
the value, and " causes the parsing to terminate.
        If they are not, then this happens: ptr is increased by one so
that instead of '\' being added to the value, '"' is added instead, and
then parsing proceeds normally.

No other sequences are affected, and a '\' does not have to be escaped
itself to be used in the value.

Attached you will find the original dcfp.c file, as well as my modified
one.  I placed a chunk of comments in at the modification I made,
briefly describing it.  You certainly don't have to use my patch (it is,
afterall, your project), and if you do, you are free to modify my code
as you see fit, or reimplement it (also, you are free to modify or
remove the comment junk that I stuck in there- I don't know what the
proper commenting is, if any).  The only thing that I ask is that if you
do use my code or a direct dirivitive, please note my name somewhere in
the source code or documentation.

I believe that this is a useful change, and if you do use it, let me
know, and I can find the relevant location in lineakconfig to add a
quick parse of each command line to escape out the " marks.

Thank you sincerely for your time, and I aplogize for anything I may
have botched in the standard procedure for submitting an update or
modifying code (for example, i have yet to learn how to use diff and
i've heard stuff about cvs... perhaps that is where I was supposed to
submit this to?)

-Colin McCambridge
ccm...@ne...

p.s.- I forgot to mention that I have tested my update with a variety of
value sequences and encountered no bugs with it. e.g., the line
search = "mozilla -remote \"openurl(http://www.google.com,new-window)\"
|| mozilla http://www.google.com"
currently resides in my ~/.lineak/lineakd.conf file and causes mozilla
to either open a new browser with google if it is already running, or
launches it to google if it is not yet running.