ant-ivy-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From "Xavier Hanin" <xavier.ha...@gmail.com>
Subject Re: Experience with ivy from SVN (2008-09-16)
Date Wed, 17 Sep 2008 13:20:48 GMT
On Wed, Sep 17, 2008 at 11:57 AM, Seaborne, Andy <andy.seaborne@hp.com>wrote:

>
>
> > -----Original Message-----
> > From: Xavier Hanin [mailto:xavier.hanin@gmail.com]
> > Sent: 17 September 2008 08:21
> > To: ivy-user@ant.apache.org
> > Subject: Re: Experience with ivy from SVN (2008-09-16)
> >
> > On Tue, Sep 16, 2008 at 6:43 PM, Seaborne, Andy
> > <andy.seaborne@hp.com>wrote:
> >
> > > Hi there,
> > >
> > > I just tried building ivy from SVN then using it on a project.  I had
> > two
> > > minor problems which don't seem to be in the CHANGES file (but I'm a
> > > beginner with Ivy so they may be expressed in language I didn't grok)
> > and
> > > one case of something seeming wrong in 2.0.0-beta and SVN.
> > >
> > > Are these intentional changes or am I unlucky?
> >
> > The way Ivy deals with paths has changed in rc1, this is intentional.
> > Paths
> > handling is documented on the "Main Concepts" page available on trunk or
> > in
> > the rc1 attempt uploaded yesterday.
>
> Is this the text you are referring to at the bottom of concepts.html?

yes


>
>
> """
> Paths handling
>
> ...
> These paths can obviously be relative or absolute. We recommend to always
> use absolute paths, so that you don't have to worry about what is the base
> of your relative paths. Ivy provides some variables which can be used as the
> base of your absolute paths.
> """
>
> Minor user feedback - it's confusing because it says "relative or absolute"
>
> I'd suggest "Ivy uses of absolute paths, so that you don't have to worry
> about what is the base of your relative paths."
>
> "Recommend" implies to me you can deviate from the suggestion which is
> reinforced by:
>
> """
> If you really want to use relative paths, the base directory used to
> actually locate the file depends on where the relative path is defined:
> """
>
> Which to my reading says you can (user beware!) use relative paths

Yes, it's true, but if you read on you have:
"In settings files, paths for file inclusion (namely properties file loading
and settings inclusion) are relative to the directory in which the settings
file is located. All other paths must be absolute unless explicitly noted."

Which is explicit. But I agree we should rephase this since it's confusing.
I'm not sure I'm the best person to rephrase it, suggestions are welcome.


>
>
> The filesystem resolver does not mention it.  It talks of "patterns" and
> "[module]/[revision]/[artifact].[ext] works" but does not mention paths
> (absolute or otherwise)

Indeed, we have to update the documentation of filesystem (and possibly
other) resolver. Could you create a JIRA for that?


>
> So user feedback is
> 1/ The text is confusing
> 2/ This is an easy trap for the new users (this is not a cygwin issue -
> that's just more complicated version of the same) but the documentation
> places it a little out of sight.
>
> I grep'ed the documentation for "absolute" and this, settings/trigger.html
> are the only mentions are concepts.html and settings/triggers.html.
>
> Cygwin users may possibly aware that any mention of: "/path/to/..." really
> means "c:/path/to/..."

We can add a note for cygwin users in the paths handling section. We will
also add a link to this section in all resolvers implying paths.



>
>
> >
> > In rc1, most paths in settings files must be absolute. It seems that the
> > File#isAbsolute() method does not return true for a unix like path under
> > cygwin, this is the root of your problem.
>
> Yes - an MS Windows absolute path must include the drive letter
> (File.isAbsolute).
>
> > I'm not sure if this is
> > something
> > that we can address or not, please file a bug in JIRA.
>
> Done - IVY-908
>
> > Another problem I
> > see
> > is that the error reporting is not nice, if the problem is "only" a
> > problem
> > of absolute path, it should be mentionned, because the current error
> > message
> > is very cryptic for a user. This can be reported in a separate JIRA.
>
> I certain agree with the "cryptic"!
>
> IVY-909
>
> >
> > That being said, as there is a workaround and only occurs in some
> > environment (cygwin), I don't think this is a blocking issue for
> > 2.0.0.Since we have just started our 2.0.0 release candidates cycles,
> > I think this
> > will rather be fixed in a 2.0.x version, unless other blocking bugs are
> > raised and this one is fixed along with the blocking ones in a next RC.
> >
> > For your problem #2, this is indeed a change in rc, the solution is to
> > prefix your pattern with ${ivy.settings.dir} or with ${user.home}.
>
> Another solution is to use Ant's <property .... location=/>
>
> >
> > For #3, this may be a blocking issue, but I'm really surprised this has
> > not
> > been reported before. Could you setup a small test case reproducing this
> > problem?
>
> Will try to.
>
> I hope you find the feedback from a nearly first time user on the
> documentation useful,

Absolutely!

Xavier


>
>
> Keep on with the great work!
>
>        Andy
>
> >
> > Xavier
>



-- 
Xavier Hanin - Independent Java Consultant
BordeauxJUG co leader - http://www.bordeauxjug.org/
Blogger - http://xhab.blogspot.com/
Apache Ivy Creator - http://ant.apache.org/ivy/

Mime
  • Unnamed multipart/alternative (inline, None, 0 bytes)
View raw message