Re: New: net/wildfly

On 2017/08/29 13:39, Ian Darwin wrote:
> Ping?
>
> On Mon, Aug 21, 2017 at 03:58:37PM -0400, Ian Darwin wrote:
> > On Tue, Aug 15, 2017 at 09:49:24PM -0400, Ian Darwin wrote:
> > > Here is a port of Wildfly 10, the Java-based app server formerly known as
> > > JBoss AS.
> > >
> > > It packages and works.
> > >
> > > Like Tomcat it goes through revisions a lot, so there's a v10 subdirectory
> > > (v11 is already in beta).
> > >
> > > It needs write access to its configuration files in several directories,
> > > which is why it's all
> > > under /var, with @mode +w for the places that need write access.
> >
> > Improved packaging, now removes cleanly (with -c)
> >
> > OK?
>

Comments from my earlier mail (offlist) still apply..

----- Forwarded message from Stuart Henderson <stu@spacehopper.org> -----

From: Stuart Henderson <stu@spacehopper.org>
Date: Wed, 16 Aug 2017 09:17:26 +0100
Subject: Re: test of attach

[..]

No attachment in this one. The one you sent to ports@ was an uncompressed tar
but named .tgz.

Sorry I didn't comment on your earlier mail, but 130-odd MB in /var seems
like maybe a bit much,

For the find(1) invocation, -delete would be safer than piping to xargs,
especially given that upstream isn't averse to including files with spaces
in the names. (Or "find .. -print0 | xargs -r0", but now we finally have
-delete we might as well use it :). I suppose the powershell (.ps1) files
aren't much use here either.


----- End forwarded message -----

Also, the USER/GROUP in Makefile are unused, and I wonder about
ownership/permissions, do *all* the files need to be owned by _wildfly?
If it's only the u+w ones that need to be writable I'd be a lot happier
if the rest were root-owned.. Additionally if it's expecting to write
to the u+w files, they probably need to be renamed/@sample'd, otherwise
if they're changed, upgrades won't work properly..