Port request: "KDEPIM/PI" PIM suite


hmc

Active Member
Joined
Dec 19, 2011
Messages
787
Location
Bavaria, Germany
This was a very comprehensive and efficient PIM suite on the Sharp Zaurus, see also my recent posting here:
http://boards.openpandora.org/index.php/topic/13253-efficient-agenda-and-organizer-for-the-pandora/page-2#entry256247

I could not make it compile on the Pandora yet, however, I am not an expert!

Maybe it's easy.

Anyone wants to try? :)

I would be endlessly grateful, if this PIM suite would be available on the Pandora.

I'd gladly create the PND for it, if just anyone could compile the programs successfully!

Here are the sources:

https://sourceforge.net/projects/kdepimpi/files/KDE_Pim/2.2.10/

Thanks!
 
Last edited by a moderator:
In the absence of anyone else stepping up to the plate, I'll return to this and give it another go. My next project was to package up XEphem - but I imagine this might have more demand so I'll put XEphem on hold for a few days/weeks.

If I get KDE/PIM to work, I'll expect hmc to make an 'appropriate doantion' to an appropriate fund!!!!!  :)

I've downloaded QT 3.2.3 source code, which seems like a good starting point. Ill post back if I get that compiled OK and provide appropriate progress reports thereafter. Not sure how long it will take - its very weather dependent (if the weather is good I have less time to spend on OP stuff as here in Scotland, you learn to take advantage of the all too infrequent good weather).
 
Hi Mark,

great, thank you for your efforts!

What fund would you consider appropriate? 

Btw: Your suggestion with the donation makes me feel that you assume that I have some commercial interest in KDEPIM. But this is not the case. I'll gladly donate something, just because I will be able to use this software for myself! (Just wanted to make this clear)

In case you need assistance from the author of KDEPIM/PI, Lutz Rogowski, please let me know. I know him, have called him a few times already to ask for compiling assistance, but since I am not a porting expert, his advice was not too helpful for me. ;-)

Even if it will take longer then - I wish you good weather and much fun outside! :) No hurry with KDEPIM/PI. I've waited for so long, now I can wait a bit longer, too.
 
Hi Mark,

great, thank you for your efforts!

What fund would you consider appropriate? 

Btw: Your suggestion with the donation makes me feel that you assume that I have some commercial interest in KDEPIM. But this is not the case. I'll gladly donate something, just because I will be able to use this software for myself! (Just wanted to make this clear)

In case you need assistance from the author of KDEPIM/PI, Lutz Rogowski, please let me know. I know him, have called him a few times already to ask for compiling assistance, but since I am not a porting expert, his advice was not too helpful for me. ;-)

Even if it will take longer then - I wish you good weather and much fun outside! :) No hurry with KDEPIM/PI. I've waited for so long, now I can wait a bit longer, too.
I have no idea what your intentions re KDE/PIM are - and my intent wasn't for a minute intended to suggest they were in any way commercial  :p  .

Appropriate means just that - if it works out, and you find it useful - throw what you feel is right into the pre order fund, or if we do not have one of those going - somewhere else that will assist with the finances of the OP in general. I'm not going to suggest that the app will be worth x, y or z; that is a very subjective thing. Whatever the amount, large or small, I'm sure it will be appreciated.
 
okay, great. Just to make this sure.

Since I also drive a small business, also recommending the Pandora to customers and occasionally brokering Pandora orders to Michael (but without a profit for me) I am a bit careful. You know, .I have made my experiences back then with the Zaurus, where I did similar things and people assumed that I tried to get profit from voluntary work open-source developers had done. That's definitely not my style. ;-) But I am sensitive for such assumptions now.

For me, KDEPIM/PI is a valuable addition to the wealth of software available for the Pandora, which will also be an argument whenever I recommend the Pandora to others, but this will maybe (hopefully) only increase sales slightly on Michael's side, not for what I offer.

Maybe, if you are successful in compiling, may yo teach me how to compile? So I will be able to apply some enhancements to the code and recompile myself... 

Do you use a crosscompiling setup, or do you compile on the Pandora? 

I never used crosscompilers, always did porting / compiling work only on the Pandora itself. 
 
okay, great. Just to make this sure.

Since I also drive a small business, also recommending the Pandora to customers and occasionally brokering Pandora orders to Michael (but without a profit for me) I am a bit careful. You know, .I have made my experiences back then with the Zaurus, where I did similar things and people assumed that I tried to get profit from voluntary work open-source developers had done. That's definitely not my style. ;-) But I am sensitive for such assumptions now.

For me, KDEPIM/PI is a valuable addition to the wealth of software available for the Pandora, which will also be an argument whenever I recommend the Pandora to others, but this will maybe (hopefully) only increase sales slightly on Michael's side, not for what I offer.

Maybe, if you are successful in compiling, may yo teach me how to compile? So I will be able to apply some enhancements to the code and recompile myself... 

Do you use a crosscompiling setup, or do you compile on the Pandora? 

I never used crosscompilers, always did porting / compiling work only on the Pandora itself. 
I always try and compile natively. In the past, I have run into trouble with cross-compilers. I came from the Z well, I had some years commercial experience on unix boxes some years back, but Linux is the first environment where I've have had a truly free hand - which is why I find the OP such a fantastic tool.

Teach you to compile? LMAO!!! You know how - I know you know (thats a known known). No seriously, the best way is to dive right in. Mistakes are seriously valuable lessons.

KDE/PIM might be a useful teaching tool though.

I'm curious, why do you not feel comfortable having a go at KDE/PIM yourself?

BTW, I'm off work tomorrow, so I'll do a first run at QT 3.2.3 then.
 
I actually tried to compile it (natively, some time ago using the devtools PND and the native-development tar, also using opkg to install all the needed packages myself, and recently using the CodeBloxks PND.

No success.

In general I know how to compile, but as soon as I hit problems, I am often lost. Sometimes I find out how to fix problems, e.g. by installing some missing header files, or by changing small portions of sourcecode, after finding out what the actual problem is. But in general, if it's more than a ./configure && make thing (and then packaging into a PND), it's difficult for me.

KDEPIM/PI in addition is different. It doesn't follow the ./configure && make scheme, instead, as a QT application, it needs qmake, some strange additional variables etc. etc. 

And that's what I meant with teaching: It would be great, if you succeed, if you could send me a list with things you needed to do in order to complie this (ideally jsut a bash script that executes all commands automatically ;-).

It's not that I hand over the compile job to someone else because I'm lazy. It's because I was not able to compile it, and it is also because I'd like to learn. Diving into it is one way, but I went that way already. Now I need help. :)

If I had unlimited time, I would maybe be able, sooner or later, to find out everything that's needed to compile. But since I also have a fulltime job, two children and currently an exam to learn for, I feel dependent on someone else's help.
 
OK, we'll take it in stages. The first thing is to get QT sorted. If that compiles OK, I'll try and put together a cheat sheet to get to that stage. It probably won't make sense at that stage for you to try and follow, as it may be necessary to recompile with a slightly different approach (perhaps because of the final runtime paths etc etc). I've certainly encountered this before with the OP. I'll follow the instructions in the documentation, and try and make suggestions and changes that make sense to the OP. And then we'll move on.

3.2.3, according to the doco' is the QT version of choice, so I'm starting with that. Obviously, we'll end up with the FULL QT environment (including the editor etc), but by the same token, we should also end up with all the required headers and development libraries. However, we should be able to pick out from the the full environment only what is needed for runtime - which doubtless could be scripted.

Be back in touch tomorrow.
 
I too ran this on my Z. Those were the days. Mine is still like new in a box in the garage next to my OmniGo 120, OG100, etc., etc. My only issue is that I want whatever I use to sync with Google. Looks like the information about syncing KDEPIM/PI with Google is pretty dated.
 
I actually tried to compile it (natively, some time ago using the devtools PND and the native-development tar, also using opkg to install all the needed packages myself, and recently using the CodeBloxks PND.

No success.

In general I know how to compile, but as soon as I hit problems, I am often lost. Sometimes I find out how to fix problems, e.g. by installing some missing header files, or by changing small portions of sourcecode, after finding out what the actual problem is. But in general, if it's more than a ./configure && make thing (and then packaging into a PND), it's difficult for me.

KDEPIM/PI in addition is different. It doesn't follow the ./configure && make scheme, instead, as a QT application, it needs qmake, some strange additional variables etc. etc. 

And that's what I meant with teaching: It would be great, if you succeed, if you could send me a list with things you needed to do in order to complie this (ideally jsut a bash script that executes all commands automatically ;-).

It's not that I hand over the compile job to someone else because I'm lazy. It's because I was not able to compile it, and it is also because I'd like to learn. Diving into it is one way, but I went that way already. Now I need help. :)

If I had unlimited time, I would maybe be able, sooner or later, to find out everything that's needed to compile. But since I also have a fulltime job, two children and currently an exam to learn for, I feel dependent on someone else's help.

Just to let you know - I haven't forgotten! Its been glorious weather (for once) in sunny Scotland lately, so I've been spending most of my time working outside. Lots of fences need fixing  :angry:  . Cows can be very destructive and patched fencing only keeps them at bay for so long.

I've had a few of runs @ QT - sadly, I haven't yet got configure to run through cleanly. I have progressed though - from a completely meaningless "Internal error" message to something I've at least got a chance at troubleshooting. There are a myriad of flags that can be set for compilation - most of them poorly documented, so its a bit of a guessing game as to what is relevant. I'll just keep plugging away.
 
That was what I meant with "difficult to compile". ;-D Way beyond my skills and experiences.

Thanks for your efforts! I really appreciate it. 
 
Qt4 version might be easier to compile. It should compile under codeblocks I think (the orginal, with qt3, doens't).
 
Last edited by a moderator:
Hmmmm....

The reason I was going for the really old version was simply that it was mentioned in the old doco' as the version of choice.

I think I will actually try a different approach. I think what I will try is to go back to where I got to before (ie, just go for it with Codeblocks), and find the deprecated function call. Then, track back to where the function was last available in QT and use that as my starting point. I should really have taken this more analytical approach in the first place - my blunderbuss just keeps on missing ATM!!!!

EDIT:

At least this way (and assuming there are not other deprecated function calls) we will found out the minimum version required without resorting to source code surgery!

'scuse the ignorance, but what is this klever.net repo???
 
Last edited by a moderator:
I don't know what klever.net is or what it is for at all.

I was just pointed to that qt4 port of KDEPIM/PI by the author of the original KDEPIM/PI. He said the port was useless to him, but he pointed me to it, in case it helps us.
 
I don't know what klever.net is or what it is for at all.

I was just pointed to that qt4 port of KDEPIM/PI by the author of the original KDEPIM/PI. He said the port was useless to him, but he pointed me to it, in case it helps us.
Well, help has come from stranger quarters!!!

I do a re-run tomorrow, and see if I can identify the minimum version of QT required without source code modification. It may be that people with knowledge of QT can suggest a source code patch to use a more modern version. We'll just keep at it.

Glad you're not in  a hurry! Right now, I'm almost done with XEphem, and I'm getting dangerously confident!
 
Hi Mark,

I just came across a qt4 port of KDEPIM/PI:

http://scm.klever.net/view/cgit/patchwork/kdepimpi.git

The KDEPIM/PI author, Lutz Rogowski, said that this port is useless to him, because it's not compileable anymore under qt2 and qt3, but that may be irrelevant for us maybe?
Sorry, the penny just dropped. I was more tired last night than I thought. This really could be useful, after all, we have QT4. I'll give this one a go shortly.
 
Moving closer. Got all the way through qmake, and make clean. Failed during make:

In file included from ../microkde/kglobal.h:6:0,

                 from calendar.cpp:25:

../microkde/kiconloader.h: In constructor 'KIconLoader::KIconLoader()':

../microkde/kiconloader.h:22:20: error: call of overloaded 'QString(int)' is ambiguous

../microkde/kiconloader.h:22:20: note: candidates are:

In file included from ../microkde/kdebug.h:6:0,

                 from calendar.cpp:24:

/mnt/utmp/codeblocks/usr/include/QtCore/qstring.h:421:43: note: QString::QString(const QByteArray&)

/mnt/utmp/codeblocks/usr/include/QtCore/qstring.h:419:43: note: QString::QString(const char*)

/mnt/utmp/codeblocks/usr/include/QtCore/qstring.h:725:8: note: QString::QString(const QString&)

/mnt/utmp/codeblocks/usr/include/QtCore/qstring.h:89:5: note: QString::QString(QChar)

/mnt/utmp/codeblocks/usr/include/QtCore/qstring.h:88:14: note: QString::QString(const QChar*)

make[1]: *** [obj/unix/calendar.o] Error 1

make[1]: Leaving directory `/media/code/projects/kdepim/p2/libkcal'

make: *** [sub-libkcal-make_default] Error 2

mark-openpandora:/media/code/projects/kdepim/p2$

 

This is the content of the KIconLoader() class definition in KIconLoader.h:

 


class KIconLoader

{

  public:

    KIconLoader()

      : mIconpath(0) {}

      

    KIconLoader( const QString &iconpath )

      : mIconpath(iconpath) {}

 

//US    QPixmap loadIcon( const QString &name, int ); 

 

    QPixmap loadIcon(const QString& name, KIcon::Group group, int size=0,

    int state=KIcon::DefaultState, QString *path_store=0,

    bool canReturnNull=false) const;

 

//US    

    QString setIconPath( const QString &);

    QString iconPath( /*US const QString &, int */) const;

    QString iconPath( const QString &, int ) const;

    QIcon loadIconSet( const QString &name) const; 

    

//US to make this class usable for different applications, we have to add a iconpathvariable

  private:

    QString mIconpath;

};



Anybody see anything odd here?
 
Last edited by a moderator:
When I tried to compile the qt4 port, I also got until a KIconLoader problem and then gave up. ;-)

Hopefully someone can see something wrong here and help out.
 
Back
Top