Re: [WiX-users] Yet another localization question

2010-08-18 Thread Mike Carlisle
Thank you, that confirms my current understanding :) I hadn't released that cultures were just for fall-backs and couldn't understand why light accepted multiple wxls of different locales when the MSI can only hold 1. On 18 August 2010 14:53, Mike Carlisle wrote: > Thank you, that confirms my c

Re: [WiX-users] Yet another localization question

2010-08-18 Thread Nick Ramirez
You are right. An MSI can only be localized for one language. What that means is actually pretty simple. If you look inside the MSI database (with a tool like Orca.exe), you'll see that many of the tables contain text (the Control table, for instance). Once the MSI is built, these strings are ther

Re: [WiX-users] Yet another localization question

2010-08-18 Thread Mike Carlisle
Thank you - the first part of that makes perfect sense. The second part is still a little ambiguous to me. When you say the cultures flag decides which .wxl files are processed this implies that the MSI can contain more than one locale, ( as the cultures flag takes a list of cultures). I didn't th

Re: [WiX-users] Yet another localization question

2010-08-17 Thread Nick Ramirez
I think your question may stem from thinking that you have to specify localization files of different cultures with each -loc flag. However, you could also specify multiple .wxl files with the same culture. For example, if you had one .wxl file to handle all of the localization strings for your di

Re: [WiX-users] Yet another localization question

2010-08-17 Thread Mike Carlisle
Thanks for your help. I have a working solution using Torch to create localized MST files and then use WiSubStg.vbs to embed them in my MSI. However I'm still confused as to the purpose of passing more than 1 WXL file to the linker, does this achieve anything? Does it ignore all but the first -loc

Re: [WiX-users] Yet another localization question

2010-08-17 Thread Blair
Windows Installer, while supporting using transforms for localization purposes, doesn't dictate how those transforms are to be named, distributed, or invoked in any "canonical" way. Thus, given the variety of groups that have engineered different approaches to deal with this, there isn't away for t