Re: Zip file default encoding

2014-10-12 Thread Stefan Bodewig
On 2014-10-11, Kristian Rosenvold wrote: Piñata encodes as NFC on my mac, which is correct. I am not sure if the NFD issue is for older macs, but at least it looks good here. Great. It may be older macs or older Java. Stefan

Re: Zip file default encoding

2014-10-11 Thread Hervé BOUTEMY
+1 for UTF-8 by default IMHO, nobody knows: everybody just want that it works (TM) I found the spec: appendix D2 APPENDIX D - Language Encoding (EFS) of http://www.pkware.com/documents/casestudies/APPNOTE.TXT this does not give the date when this bit was reserved, but seems quite old: I don't

Re: Zip file default encoding

2014-10-11 Thread Kristian Rosenvold
Piñata encodes as NFC on my mac, which is correct. I am not sure if the NFD issue is for older macs, but at least it looks good here. Kristian 2014-10-10 18:22 GMT+02:00 Stefan Bodewig bode...@apache.org: On 2014-10-10, Kristian Rosenvold wrote: I will investigate some modern windows. I

Zip file default encoding

2014-10-10 Thread Kristian Rosenvold
Currently plexus-archiver uses platform encoding for zip file names if none is specified. This is compliant with traditional Zip history. But zip understood this is a problem and introduced a LEF flag, which basically means all entries must be UTF-8. I would like to switch defaults in plexus zip

Re: Zip file default encoding

2014-10-10 Thread Anders Hammar
As this is a step towards reproducable builds I see this as a good thing. Depending on platform defaults should be punished...:-) /Anders On Fri, Oct 10, 2014 at 8:26 AM, Kristian Rosenvold kristian.rosenv...@gmail.com wrote: Currently plexus-archiver uses platform encoding for zip file

Re: Zip file default encoding

2014-10-10 Thread Karl Heinz Marbaise
Hi, +1 from too... On 10/10/14 8:45 AM, Anders Hammar wrote: As this is a step towards reproducable builds I see this as a good thing. Depending on platform defaults should be punished...:-) Yes... ;-) /Anders On Fri, Oct 10, 2014 at 8:26 AM, Kristian Rosenvold

Re: Zip file default encoding

2014-10-10 Thread Karl Heinz Marbaise
Hi Kristina, just a question on this topic...does currently jar from plexus-archiver already behave like this and uses UTF-8 encoding always ? Or is this platform dependent? On 10/10/14 8:26 AM, Kristian Rosenvold wrote: Currently plexus-archiver uses platform encoding for zip file names

Re: Zip file default encoding

2014-10-10 Thread Kristian Rosenvold
No. It is platform encoding dependent. Which is why we have a bunch of issues in maven-assembly-plugin about this behaviour. The JAR archiver does this already, since its part of the jar spec. Kristian 2014-10-10 11:58 GMT+02:00 Karl Heinz Marbaise khmarba...@gmx.de: Hi Kristina, just a

Re: Zip file default encoding

2014-10-10 Thread Kristian Rosenvold
I am looking at all the encoding-related issues in assembly now btw. Kristian 2014-10-10 12:16 GMT+02:00 Kristian Rosenvold kristian.rosenv...@gmail.com: No. It is platform encoding dependent. Which is why we have a bunch of issues in maven-assembly-plugin about this behaviour. The JAR

Re: Zip file default encoding

2014-10-10 Thread Karl Heinz Marbaise
Hi Kristian, On 10/10/14 12:16 PM, Kristian Rosenvold wrote: No. It is platform encoding dependent. Which is why we have a bunch of issues in maven-assembly-plugin about this behaviour. Ok...so i think this issue http://jira.codehaus.org/browse/MJAR-135 is related to this behaviour as

Re: Zip file default encoding

2014-10-10 Thread Stefan Bodewig
On 2014-10-10, Kristian Rosenvold wrote: This will break builds for people that actually *want* archives in file.encoding; they will need to explicitly specify encoding to get that encoding. To my understanding this is not really an issue, LEF support is ancient history. At the time where I

Re: Zip file default encoding

2014-10-10 Thread Kristian Rosenvold
I suspect defaulting to UTF-8/LEF will work ok for most people, I will do some tests on windows. There's probably 1 billion chinese that rely on CP936 platform default encoding on windows XP (IE6 all over again) but I suspect they will be better of making an explicit specification of the

Re: Zip file default encoding

2014-10-10 Thread Stefan Bodewig
On 2014-10-10, Kristian Rosenvold wrote: I will investigate some modern windows. I suppose this may cook down to various unicode normalization forms on different OS'es. Does commons compress deal with mac-os unicode normalization ? (e.g does it handle Piñata on mac over to a PC ) ? TBH I'm