Use Advanced Search to search the entire archive.
[jsr363-experts] Re: Road to Final Draft
- From: Werner Keil <
>
- To: "
" <
>
- Subject: [jsr363-experts] Re: Road to Final Draft
- Date: Fri, 18 Mar 2016 11:59:33 +0100
OK, if it's simply a small change, that seems best in master (unless you
prefer PR from a local fork) for all bigger suggestions, e.g. merge
Bootstrap with ServiceProvider or similar, please do in a fork or branch
first and propose changes to the master if it works as intended.
Werner
On Fri, Mar 18, 2016 at 11:38 AM, Martin Desruisseaux <
>
wrote:
>
Le 18/03/16 10:45, Werner Keil a écrit :
>
> Bootstrap is heavily based on JSR 354 (if we find mistakes like the
>
> Error type not being caught, I'd say it may also have to be raised
>
> there in JIRA)
>
>
To not caught java.lang.Error is not necessarily a mistake. We are
>
usually not supposed to catch those errors, except sometime
>
OutOfMemoryError.
>
>
So I suggest to remove that try ... catch block completely. It is almost
>
useless in its current form (except maybe for catching
>
SecurityException, or ConcurrentChangeException if two threads use the
>
class is same time), and expanding its scope for catching
>
java.lang.Error is usually not recommended. Furthermore the current
>
error handling is questionable (stop at the first error, ignore all
>
potentially good providers remaining in the iterator).
>
>
Martin
>
>
>