[izpack-users] shortcutSpec.xml UTF-8 Byte Order Mark breaksIzPackoperation

Bartz, Klaus Klaus.Bartz at coi.de
Tue Oct 17 17:37:56 CEST 2006

Hi Hans-Georg,
the COIOSHelper.dll do nothing about shortcuts. In the common
version only the native part of registry support is in it.
The ShellLink.dll makes the native part for creating shortcuts.

I have not written the shortcut part of IzPack, but if I look
into the sources it seems for me, that the parsing of the xml
file will be performed in the java part of the shortcut stuff 
(see ...panels.ShortcutPanel.readShortcutSpec).
OK, but why only a crash with the shortcutSpec.xml???

Really good question. The shortcutSpec will be parsed also by
nanoxml, what else? In opposite to other specs this will be
directly substituted by the VariableSubstitutor. This uses a
InputStreamReader which uses StreamDecoder$CharsetSD. 
If a BOM is there, this will not consumed else read returns
65259 as first character. I do not know why and I do not know
whether this is a bug in the VM or not. I have the impression
that you like to write bug reports. Therefore retrace the bug
and write a report to sunsoft.



>-----Original Message-----
>From: izpack-users-bounces at lists.berlios.de
>[mailto:izpack-users-bounces at lists.berlios.de]On Behalf Of Hans-Georg
>Sent: Monday, October 16, 2006 3:46 PM
>To: izpack-users at lists.berlios.de
>Subject: Re: [izpack-users] shortcutSpec.xml UTF-8 Byte Order Mark
>On Tue, 10 Oct 2006 14:07:11 +0200, Bartz, Klaus wrote:
>>byte order for UTF-8... it is matter of opinion whether a
>>silent add of a BOM is clever or not. May be you should
>>send a bug report to the developer of the editor.
>thanks again for your reply! The editor is probably
>configurable, but my problem was that I didn't even think of the
>Byte Order Mark at first.
>As far as I can see, nanoxml doesn't have any problems with the
>BOM, because the other XML files never had any such problem.
>Paradoxically, only the DLL that makes the shortcuts for Windows
>(COIOSHelper.dll) seems to have it, but I haven't done any
>thorough tests. Perhaps that DLL doesn't use nanoxml?
>I think, the quickest solution for now would be to write a
>warning into the documentation. If it is true that only the
>shortcut generation is affected, then such a warning should go
>into the shortcut chapter.
>Maybe we have the problem that nobody finds the time to do some
>work on the documentation.
>No mail, please.
>izpack-users mailing list
>izpack-users at lists.berlios.de

More information about the izpack-users mailing list