[rfc][icedtea-web] u51 classpath manifest entry implementation

[Jiri Vanek]

On 02/07/2014 07:49 PM, Jacob Wisor wrote:
> On 02/07/2014 04:22 PM, Andrew Azores wrote:
>> On 02/06/2014 11:59 AM, Jiri Vanek wrote:
>>>
>>>>
>>>>> +CBCheckUnsignedFail= Codebase DO NOT matches codebase manifest attribute,
>>>>> but application is
>>>>> unsigned. Continuing
>>>>
>>>> The application's codebase does NOT match the codebase specified in its
>>>> manifest, but the
>>>> application is unsigned. Continuing
>>>>
>>>
>>> followed
>>
>> This one didn't get changed. Otherwise, I think this is okay to push now.
fied (Ihope)
>>
>> Do you have any plans for making this validation information more visible? I
>> don't think most people are even going to notice the log messages.
 >

Pushed

>
> I do not want to sound too grim here but probably most users will not understand what this messages
It not. I value your "human"	point of view. I just mostly dont know what to do with it:(

> is all about. Even many admins do not know what a jar or application manifest is. :-\ Perhaps you
> should consider rewording this message towards what the effects of non-matching codebases are than
> just stating a plain fact.

Hm hm.... My verbose skill is probably to low ;( I heve added url to redeploy warnning Oracle's pages.
>
> Sometimes I can not escape the impression that developers and JCP committees are throwing new
> features and quick dirty fixes at users just to extend the life span of products with broken
> concepts. They just patch technical details instead of really solving problems. But hey, who am I to
> complain. ;) All I am asking of you is to please think instead of to blindly follow a "reference
> implementation".
>
Yes, I see your point and I agree. However - what can we do? To replace current security 
implementation of JDK will take *decades* especially considering the backward comatinbility :(


J.