Re: [Libxmlplusplus-general] write() broken, maybe fixed

[Christophe de VIENNE <cdevienne alphacent com>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Le Lundi 16 Décembre 2002 17:38, Murray Cumming a écrit :
> >
> > no. The _content attribute is used only in text nodes, and it has always
> > been like this in libxml++.
>
> "always" is very unclear. I don't think that libxml++ was usable before
> 0.16.0 and I think that the current API represents text as "content"
> rather than as child nodes of a certain type.

at least it's the way I was anderstanding the stuff. That led to all these 
misanderstanding between us since the beginning. I hope it will get better... 
if we continue to interpret the API in a different way we won't go far :-(
In fact I was not andestanding why you wanted to put the content in the parent 
node. I was using (it _was_ usable, even if not perfect) libxml++-0.13 and 
this behavior was logic to me, and there was no reason for me to change it.

>
> What you suggest seems very silly:
> NodeText = node.add_content("some text");
> nodeText->set_content("changed text");

Thanks for the silly ;-)
Indeed this looks not logic. But the documentation of libxml++-0.13 was 
somehow clear on this point, and I think people were using it this way, like 
I did.

extract from old README
    123 bool is_content()
    124     Returns true if the node is a content node and false otherwise.
    125
    126 const string & content()
    127     Returns the content of the node.
    128
    129 const string & content(const string &c)
    130     Sets the content of the node.
[...]
    145 XMLNode *add_content(const string &c = string())
    146     Adds a new child that is set to store content.
    147

When starting working on libxml++ this was very clear to me, and working well. 
So my feeling was, why change it since it's clear and working ?
The problems started because of a misanderstanding between us about the 
content handling.

>
> What I suggest seems clearer:
> node->set_content("some text");
> node->set_content("changed text");

agree it's clearer. But to have this working correctly we have to change 
set_content behavior.

>
> > yes. By the way, I want to warn people waiting for these changes that I
> > have a lack of time this week and next one. So the next release shouldn't
> > happen before beginning of january.
>
> Which effectively means the middle of January at the latest.
>
> I wouldn't worry about this temporary problem, but this makes it not so
> temporary.

I'll do my best to finish it sooner but I cannot make my days longer 
unfortunately.

> > If you don't mind, we'll go back on your last change on node.cc, and
> > correct the dom_build example.
> > Since we didn't do any big changes since the last release, we can do a
> > 0.17.1 version release including the fix.
>
> I would prefer to
> a)
> remove add_content() because it seems unnecessary and leads to
> confusion.
> b)
> Hear about any problems with the current code. Maybe you need to repeat
> something from the previous emails.

I think there is no proper solution but change the API.
I'll try to do something sooner to avoid the confusion being too long...

I keep you informed,

Cheers,

Christophe
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.0.6 (GNU/Linux)
Comment: For info see http://www.gnupg.org

iEYEARECAAYFAj3+FbsACgkQB+sU3TyOQjDr5wCfdatCwWJOjqUis5ZW4GtIod61
H3MAoJo7xQZcj6Dg1jldW9IoEE5Kqka8
=VMlf
-----END PGP SIGNATURE-----