Yes, try the build process you described.  Sorry I didn't communicate that 
very well before.

Note that when we build STAF for all the Unix operating systems for which 
we provide binaries, we run the entire STAF build process natively on each 
operating system, so I don't know for sure that this will work, but you 
definitely need to build the codepage files (e.g. run ucm2bin on each .ucm 
file) on the target operating system.

You might find it helpful to look at the STAX job we use to build STAF for 
all operating systems (in src/staf/autobuild/BuildSTAF.xml) and its 
"BuildWinAMD64-CodepageFiles" function which is responsible for building 
the codepage files for the WinAMD64 platform (as we currently don't build 
the rest of STAF for the WinAMD64 platform on the Windows 32-bit build 
machine).

--------------------------------------------------------------
Sharon Lucas
IBM Austin,   luc...@us.ibm.com
(512) 286-7313 or Tieline 363-7313




From:   Erik Mägi <er1k....@gmail.com>
To:     Sharon Lucas/Austin/IBM@IBMUS, 
Date:   06/25/2012 03:59 PM
Subject:        Re: [staf-users] Codepage problems with cross-compiled 
STAF



To make sure I understood correctly, I can build ucm2bin together with the 
rest of STAF source on the host machine for the target but I need to run 
the native ucm2bin on the target to convert the codepage files, and then 
move the converted codepage files back to host for packaging, correct?


On 25 June 2012 18:37, Sharon Lucas <luc...@us.ibm.com> wrote:
The STAF codepage code has to be run on the target platform (i.e. be 
natively built).  For example, for STAF Windows 64-bit we can compile the 
non-STAF codepage code on Windows 32-bit with compile flags for 64-bit, 
but we have to compile the STAF codepage code on Windows 64-bit and then 
our build process transfers the ucm2bin executable to the 32-bit Windows 
build machine for packaging.  We build STAF on each of the target Unix 
platforms. 

--------------------------------------------------------------
Sharon Lucas
IBM Austin,   luc...@us.ibm.com
(512) 286-7313 or Tieline 363-7313




From:        Erik Mägi <er1k....@gmail.com> 
To:        staf-users@lists.sourceforge.net, 
Date:        06/23/2012 05:42 AM 
Subject:        [staf-users] Codepage problems with cross-compiled STAF 




Hi, 

I have cross-compiled (only the basic stuff -- PROJECTS="staf connprov*") 
 STAF 3.4.3 for 2 embedded linux-based platforms that run off of 
Freescale MPC5121 and Intel Atom respectively. When building, I 
encountered problems when the ucm2bin application was executed as it was 
built for the target platform and couldn't run on the building host. I 
disabled the codepage conversion and ucm2bin building by renaming the 
makefile in /codepage and STAF built successfully, I just made sure that 
the alias.txt file was present in the packaged version and all was well -- 
STAF was working on both paltforms. 

We use the QUEUE service a lot in our lab for communication between 
different test nodes and lately we have developed the need to send 
messages that contain non-ASCII characters. It works okay if we send the 
messages between linux pc-s but if I try to send something like this: STAF 
target QUEUE QUEUE MESSAGE 
"<cmd_msg><cmd>TestMessage</cmd><arglist><arg>Fiskemenu: Rødspætter og 
å</arg></arglist></cmd_msg>" NAME test_handle, I get the following on the 
other end:  
STAF local QUEUE LIST HANDLE 29 
Response 
-------- 
P Date-Time Machine  Name     H#  User     Type   Message                 
      
- --------- -------- -------- --- -------- ------ 
----------------------------- 
5 20101011- ssl://19 pyHandle 196 none://a <None> <?xml version=1.0 
?><cmd_msg> 
  08:34:58  2.168.51              nonymous       
 <cmd>TestMessage</cmd><arglis 
terminate called after throwing an instance of 'STAFException' 
            .95@6550                              Aborted 

It seems to be somehow related to the missing codepages. Maybe STAF is 
trying to convert the exotic characters but the codepage it tries to 
access isn't there so it throws an exception, or something like that. The 
question is, can I add support for different character sets after I have 
built STAF by manually running ucm2bin and moving the resulting files in 
the correct directory or does this have to happen at build time? Also, are 
the .bin files created by ucm2bin platform-specific, ie. can I use ucm2bin 
built for my linux pc to convert files for the Freescale device or does 
each platform require its own natively built ucm2bin? 

Thanks, 
Erik
------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
staf-users mailing list
staf-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/staf-users


------------------------------------------------------------------------------
Live Security Virtual Conference
Exclusive live event will cover all the ways today's security and 
threat landscape has changed and how IT managers can respond. Discussions 
will include endpoint security, mobile security and the latest in malware 
threats. http://www.accelacomm.com/jaw/sfrnl04242012/114/50122263/
_______________________________________________
staf-users mailing list
staf-users@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/staf-users

Reply via email to