> ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
> Attributes = Author, Type, Category, Subject
> Required Attributes = Author, Type, Subject
> Options Type = Deposition{1}, Arc Studies{1}, Conditioning{1}, Vacuum{2},
> Other{3}
> Preset Type = Deposition

The "Preset xxx" option actually never was ment to be used with conditional
attributes. I added that functionality in rev. 1.553. Please give it a try.

> the following should do the job:
> 
> <div style="width: 200px">
> <img src="IMG_3133.jpg" style="width: 100%" />
> </div>
> 
> the configurable parameter should be with obviously :-)

Unfortunately not. Please find attached the screen dump from such a try, together with the HTML code.
As you can see, the little elog icon is stretched to the same width as the upper (large) picture.

> elog sends wrong HTTP headers, charset is empty and not separed from the
> next one "Connection: Keep-Alive ... this bug tells the browser that the
> encoding is "connection:" ... that is wrong.

This has to do with the "%S" vs. "%s" bug described elsewhere in this forum. 
I fixed that. The validator link above now also reports error free.

I added ths missing tags. The validator at least does not complain any more.

> > ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;; 
> > Attributes = Author, Type, Category, Subject 
> > Required Attributes = Author, Type, Subject 
> > Options Type = Deposition{1}, Arc Studies{1}, Conditioning{1}, Vacuum{2}, 
> > Other{3} 
> > Preset Type = Deposition 
>  
> The "Preset xxx" option actually never was ment to be used with conditional 
> attributes. I added that functionality in rev. 1.553. Please give it a try. 
 
it works fine !!! thank you so much (btw, using Preset xxx{1} it doesn't 
works.. but it should be the correct beaviour) 

> > the following should do the job: 
> >  
> > <div style="width: 200px"> 
> > <img src="IMG_3133.jpg" style="width: 100%" /> 
> > </div> 
> >  
> > the configurable parameter should be with obviously :-) 
>  
> Unfortunately not. Please find attached the screen dump from such a try, together with the HTML code. 
> As you can see, the little elog icon is stretched to the same width as the upper (large) picture. 
 
you are right ;-( I used a different code in my test and sent you and old test code, *but* also the last 
one does not works on every browser ... the following code would do the job, but it seems to work only 
in gecko-based browsers such as mozilla, firefox and safari ... 
 
<div style="width: 200px"> 
 <img src="IMG_3133.jpg" style="max-width: 200" /> 
</div> 
 
other browsers do not scale the vertical dimention :-( 

As I wrote you by mail, I implemented your idea in the following way. As you
say, it's only working on Gecko based systems. I did not implement additional
configuration parameters, but made it through the CSS file:

- adding a new CSS style "attachmentframe" around all attachments
- adding class "attachmentframe" to the 'default.css' like:

.attachmentframe {
  border:1px solid grey;
  border-top:1px solid lightgrey;
  border-left:1px solid lightgrey;
  background-color:white;
  padding:5px;
}

.attachmentframe img {
    max-width:100%;
}

The latter restricts the image size to 100%, which is the width of the 
browser window. I like it that, but of course you can write 200px or 
anything else. Please find the modified elogd.c and default.css under CVS.

> > as this url shows http://www.htmlhelp.com/tools/validator/problems.html#amp
> 

> > it should be used an HTML entity instead of the ampersand sign.<p>
> 
> Can you please be a bit more specific? In which URL should the ampersand be
> replaced?

uhmm...

http://midas.psi.ch/elogs/Forum/page2?mode=threaded&expand=0&last=7

the bug is also a "documentation bug", infact similar tricks are also
in FAQs (FAQ 5 for example)

I tried to substitute avery occurrence of hard coded "&" with "&amp;" but it does
not do the job at all ... infact the first time the HTML code is ok, but
following a link let the browser to automatically decode html entities .. and
everything turns wrong again ... maybe the encoding should be done in
printing-to-the-browser-time ..