[Player-dev] AMD64: porting status of Helix player
Daniel Yek dyek at real.comThis is looking good, Max! You are doing well.
More inline...
At 02:11 PM 1/7/2007, Massimo Perga wrote:
>Hello All,
> I was finally able to get a build without any failure.
>
>These are the details about my build:
>
>[0] Set BIF branch (hxplay_gtk_current)
>[1] Set Target(s) (player_all)
>[2] Set Profile
>(/home/massimo/amd64_helix/build/umakepf/helix-client-all-defines-free)
>
>Before all, I ported my fixes for the Helix Client on AMD64-native (please
>check the Helix Client ML) on common/container/hxstrfmt.cpp and .
>
>What I had to do specifically for the player were:
>
>1. to update the common/import/gecko-sdk/nspr/include/prcpucfg.h to a
>newer version, because the file version currently on the repository
>doesn't take care of the AMD64
>
>2. changing the script player/installer/archive/make_tempdir.mak to point to
> -> ../../../player/app/plugin/dbg/nphelix.so (earlier it was pointing
> to ../../../player/app/plugin/nphelix.so)
> -> ../../../player/app/gtk/dbg/hxplay.bin (earlier it was pointing to
> ../../../player/app/gtk/hxplay.bin)
> As it's a generated file, which is the original script who generates it ?
The following Umakefil generates it:
player/installer/archive/make_tempdir
> Is it something related to my configuration ?
You can troubleshoot the problem by backtracing in the following steps:
(1) player/installer/archive/make_tempdir:
inst.CopyModuleDlls() and inst.CopyModuleExes() aren't working right
(2) installer/common/scripts/inst.py, unixinstaller.py, and installer.py:
defines "class inst", CopyModuleDlls(), and CopyModuleExes().
It appears to me that CopyModuleBuildFile() and self.module_output_dir
are not working/defined correctly.
Insert outmsg.send() and then execute umake in:
player/installer/archive
to print out relevant values.
I attached the patch below to do exactly this for your convenience.
(3) $BUILD_ROOT/umakecf/unix.cf:
This file sets the value of project.output_dir to dbg.
I believe this is set properly. So, you will have to find out why it
is not being used.
>So I finally got :
>
>- HelixPlayer-1.1.0.2697-20070107.amd64.rpm
>- HelixPlayer-1.1.0.2697-20070107.src.rpm
>- hxplay-1.1.0.2694-linux-2.6-glibc23-amd64.bin
>- hxplay-1.1.0.2697-linux-2.6-glibc23-amd64.tar.bz2
>
>As I'm working on Debian, I tried to run
>the hxplay-1.1.0.2694-linux-2.6-glibc23-amd64.bin file and the
>installation process seemed to work properly (including the GTK GUI): I've
>one issue because it seems not finding the player. Am I missing something
>or it should be distributed in such a ".bin" file ?
You need to be more specific than this.
Did you execute the realplay script, or you execute realplay.bin?
What exactly meant by "it seems not finding the player"? Did you run the
standalone Helix Player application? Or you tried the embedded player on
the web page?
Did you install using root privilege (login as root during install)? Or
user privilege?
>I've attached the build log of my successfull build; please let me know
>your feedbacks about what I had to fix, so that I could know what to write
>in the CR I'll open for AMD64 porting.
You can just send a CR to the mailing list. If it is checked in soon
enough, you don't have to open a problem report in the bug database.
>Thanks,
>Regards,
> Max
Index: installer.py
===================================================================
RCS file: /cvsroot/installer/common/scripts/installer.py,v
retrieving revision 1.5.6.1
diff -u -w -r1.5.6.1 installer.py
--- installer.py 5 Jan 2007 22:51:45 -0000 1.5.6.1
+++ installer.py 8 Jan 2007 06:17:00 -0000
@@ -61,6 +61,8 @@
self.archive_info_header = "archive_info.h"
self.inst_output_dir = project.output_dir
self.module_output_dir = project.output_dir
+ outmsg.send("self.module_output_dir = %s" % (self.module_output_dir))
+ outmsg.send("self.inst_output_dir = %s" % (self.inst_output_dir))
self.include_archive_prog = 0
self.stage1_name = "stage1"
self.installer_name = None
--
Daniel Yek