make RELEASE=1 ?

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|

make RELEASE=1 ?

denisgolovan
Hi, all!

  Some time I was building LCL using standart command:
  make clean all # for current target cpu,os
  and
  make clean all OS_TARGET=wince CPU_TARGET=arm PP=ppcarm

  ... and everything seem working well.

  The problem is that I got some special bugs - some components such TMemDataset are unable to work with ARM processors without proper data
alignment. Such bugs are hard to find without debug info (-gl switch) and later stack tracing. Another problem is my executable with debug info
about 18Mb size and due its size it just refuses to run on Windows CE :(.
  So I decided to switch -gl only on some potentially problematic units and remove debug info from others. (Strip removes entire debug info)
 
  I tried to build LCL using:

  make clean all RELEASE=1 OS_TARGET=wince CPU_TARGET=arm PP=pparm

  but it just cannot find some base units (i.e. contnrs - fcl-base package). Its because makefile generated by fpcmake contains -n switch (do not
read standart fcp config). Removing -n from makefile (not makefile.fpc of course) does the job and LCL successfully compiles with RELEASE=1.

  Maybe someone give a hint how to use RELEASE=1 conditional without removing -n manually from makefiles?
  Or how to debug on wince :) ?

P.S. Gentoo Linux, Lazarus SVN, FPC SVN

--
Best regards,
Denis Golovan aka MageSlayer
_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: make RELEASE=1 ?

Peter Vreman
> Hi, all!
>
>   Some time I was building LCL using standart command:
>   make clean all # for current target cpu,os
>   and
>   make clean all OS_TARGET=wince CPU_TARGET=arm PP=ppcarm
>
>   ... and everything seem working well.
>
>   The problem is that I got some special bugs - some components such TMemDataset are unable to
> work with ARM processors without proper data
> alignment. Such bugs are hard to find without debug info (-gl switch) and later stack tracing.
> Another problem is my executable with debug info
> about 18Mb size and due its size it just refuses to run on Windows CE :(.
>   So I decided to switch -gl only on some potentially problematic units and remove debug info from
> others. (Strip removes entire debug info)
>
>   I tried to build LCL using:
>
>   make clean all RELEASE=1 OS_TARGET=wince CPU_TARGET=arm PP=pparm
>
>   but it just cannot find some base units (i.e. contnrs - fcl-base package). Its because makefile
> generated by fpcmake contains -n switch (do not
> read standart fcp config). Removing -n from makefile (not makefile.fpc of course) does the job and
> LCL successfully compiles with RELEASE=1.
>
>   Maybe someone give a hint how to use RELEASE=1 conditional without removing -n manually from
> makefiles?
>   Or how to debug on wince :) ?
>
> P.S. Gentoo Linux, Lazarus SVN, FPC SVN

It is known issue and a shortcomming of the lazarus Makefile.fpc's. They rely on the paths
configured in fpc.cfg. Instead of RELEASE=1 you can use 'OPT=-g-' to disable debug info.

Peter


_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal
Reply | Threaded
Open this post in threaded view
|

Re: make RELEASE=1 ?

denisgolovan
On Fri, Feb 29, 2008 at 11:39:19AM +0100, Peter Vreman wrote:

> >   Maybe someone give a hint how to use RELEASE=1 conditional without removing -n manually from
> > makefiles?
> >   Or how to debug on wince :) ?
> >
> > P.S. Gentoo Linux, Lazarus SVN, FPC SVN
>
> It is known issue and a shortcomming of the lazarus Makefile.fpc's. They rely on the paths
> configured in fpc.cfg. Instead of RELEASE=1 you can use 'OPT=-g-' to disable debug info.
>
> Peter

  Thanks, Peter. Seems this is the only way :(

--
Best regards,
Denis Golovan aka MageSlayer
_______________________________________________
fpc-pascal maillist  -  [hidden email]
http://lists.freepascal.org/mailman/listinfo/fpc-pascal