[prev in list] [next in list] [prev in thread] [next in thread] 

List:       kde-core-devel
Subject:    Re: Success: kde 3.1 cvs running on sgi IRIX!
From:       Dominique Devriese <fritmebufstek () pandora ! be>
Date:       2002-08-11 9:32:44
[Download RAW message or body]

Andreas Sturm <a_sturm@naimasoft.com> writes:

> Hi!
> 
> FYI: i have compiled a kde 3.1 cvs of around 8th august using MIPSpro 
> compilers on sgi IRIX. It looks quite nicely; good work! I have attached a 
> file with my changes to arts, libs, artwork and partly network (not yet 
> finished). I would like you to comment this changes but please be gentle with 
> me as i am not a c++ programmer yet (i am trying to get better with that; 
> prommise!). Shurely you have better solutions than me.
> Again: realy great work; i do like the new desktop already!
> 
> TMEzz37 (Andy)

Here are some comments :

1 how about a patch ? ( check out "man diff" ;)

<snip>

> kdeui/ktoolbarbutton.h:			cc-1117 CC: ERROR File = ./ktoolbarbutton.h, Line = 141
>   					An expression appears after a "return" in a "void" function.
> 					removed the return statement

this is a bug in your compiler, you should prolly send a bug report to
SGI ( or whoever wrote it... ), but in this case removing the return
indeed does no harm...

<snip>
> interfaces/kmediaplayer/player.cpp:	added using namespace KParts
> right above the include moc-file

You should be careful with "using namespace"s, it could be harmful for
--enable-final stuff...

<snip>
> kate/plugins/isearch/ISearchPlugin.moc: 72, 77, 82, 88, added ISearch to
> 					 Plugin::qt_cast( clname );
> 					 Plugin::qt_invoke(_id,_o);
> 					 Plugin::qt_emit(_id,_o);
> 					 Plugin::qt_property( _id, _f,
> _v);

You didn't alter an moc file, right ?  Since those are auto-generated,
that would be _very useless_.  You should find another solution.  If
necessary, you should prolly send a bug report/patch to Qt about the
moc program...

<snip>
> kstyles/keramik/genembed.cpp: changed cmath to math.h and cstring to
> string.h 

hm... seems like a bug in your STL...
<snip>

> kcontrol/kicker/main.cpp: 	cc-1080 CC: ERROR File = main.cpp, Line = 47
>   				A storage class cannot be specified at
> this point. 
> 
>   				static int
> KickerConfig::kickerconfig_screen_number = 0; 
> 				
> 				Don't know what to do; removed static

that's idd the correct fix, it seems to be in cvs already...

> kscreensaver/kdesavers/gravity.cpp: had to add a few float casts to use pow
> cc-1282 CC: ERROR File = gravity.cpp, Line = 347
>   More than one instance of overloaded function "pow" matches the argument list.
> 
>             Function symbol function "pow(double, double)" is ambiguous by
>                       inheritance.
>             Function symbol function "pow(float, float)" is ambiguous by
>                       inheritance.
>             Function symbol function "pow(float, int)" is ambiguous by
>                       inheritance.
>             The argument types are:  (float, double).
>
> particle[loop].z=(particle[loop].zo*sin(particle[loop].index))*
> pow(particle[loop].index/particle[loop].indexo,8.0);

Does this mean pow on IRIX is declared pow( float, double ) ??
weird...

> kget/kmainwidget.cpp: cc-1020 CC: ERROR File = kmainwidget.cpp, Line = 2150
>  			 The identifier "AF_IPX" is undefined.
> 
>      			 ipx_sock = socket(AF_IPX, SOCK_DGRAM, 0);
> 			 
> 			 don't know what that is; commented it out
> (any ideas?)

hm.. shouldn't that be PF_IPX  ?

> kpf/src/ActiveMonitorItem.cpp:65, fixed return in void function

s/fixed/worked around/ ;) since it _is_ valid c++ ...

> krfb/srvloc/kinetaddr.cpp: added #include <strings.h> for bzero

hm.. shouldn't that be string.h ?

> krfb/libvncserver: cc-1020 cc: ERROR File = main.c, Line = 404
>   			The identifier "source" is undefined.
> 
>      			source: "\000\102\044\030\044\102\000",
>      			^
> 
> 		   cc-1067 cc: ERROR File = main.c, Line = 404
>   			A right brace is expected.
> 
>      			source: "\000\102\044\030\044\102\000",
>            			^
> 			stopped working here (it's gotten a way to
> late now.....) 

i think you can just remove all the labels like "source:" there
(provided they are in the right order...) ( although i think it's way
more readable like it is now, and it seems valid to me too... )

Anyway, very nice work, if you provide a patch, i'm sure it will
easily get committed..
Might be a cool thing for the famous "planned features" page: IRIX
port
cheers
domi

-- 
Isn't this my STOP?!
[prev in list] [next in list] [prev in thread] [next in thread]