commons-issues mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Sebb (JIRA)" <j...@apache.org>
Subject [jira] Updated: (NET-272) Backport / Downgrade of commons-net-2.0 to Java 1.4
Date Sat, 10 Jul 2010 16:18:55 GMT

     [ https://issues.apache.org/jira/browse/NET-272?page=com.atlassian.jira.plugin.system.issuetabpanels:all-tabpanel
]

Sebb updated NET-272:
---------------------

    Affects Version/s: 1.4
                           (was: 2.0)

Affects 1.4, not 2.0

> Backport / Downgrade of commons-net-2.0 to Java 1.4
> ---------------------------------------------------
>
>                 Key: NET-272
>                 URL: https://issues.apache.org/jira/browse/NET-272
>             Project: Commons Net
>          Issue Type: Improvement
>    Affects Versions: 1.4
>         Environment: Java 1.4 VM (as opposed to usually required Java5 VM)
>            Reporter: Christian Gosch
>            Priority: Minor
>   Original Estimate: 24h
>  Remaining Estimate: 24h
>
> I need FTP over TLS used as Client on an IBM WebSphere 6.0 AppServer which only allows
for the built-in Sun/IBM JRE 1.4.2, and the one and only useful open source Java library for
doing so currently seems to be commons-net-2.0 which requires a Java5 VM. (The alternative
EnterpriseDT edtFTPj/PRO may fit fine but is way too expensive.)
> Thus I tried to "back-port" the commons-net-2.0 sources to Java 1.4 which was feasible
up to some 95% in a few hours :-) 
> Of course the nice things about type safety and so on (generics, extended "for" loop
syntax...) disappeared gracefully, the enum in the TFTPServer had to be replaced by an equivalent
implementation, and all annotations have been abandoned, but finally only 1 hard stopper remained:
I could not find a useful replacement for the use of MatchResult in RegexFTPFileEntryParserImpl
and VMSVersioningFTPEntryParser, which could guarantee the separation of changes in the same
way MatchResult does. (The __listeners list in ListenerList is solved by Collections.synchronizedList(new
ArrayList()), which at least guarantees thread safety.)
> I have that code at hand and can donate it to the community, but at least it should only
get into some branch, and finally: Are you or is someone else interested in this code at all?
> If so, there should be some way to publish it, maybe here or elsewhere. Remaining tasks
are: Finding a useful replacement for Java5 MatchResult in Java 1.4, re-integration with the
maven build environment (i did the modifications somewhat "offline").
> If not, sorry for the disturbance...

-- 
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.


Mime
View raw message