Proof of principle ;-)

Stefan Ritt wrote:

Andreas Luedeke wrote:

If the content has been added with the "elog" command as HTML then it can contain mismatching HTML tags, can't it?
I don't see how this could be avoided by ELOG, unless you want to do a full HTML syntax check of all new and modified entries.

It works also with entries submitted by the "elog" command. The reason is that for the OUTPUT, all "<" characters are transformed into >, which "inactivates" any HTML code, no matter where it came from.

Now I'm confused: if I create an entry with "elog -n 2 ...", then I put HTML code into elog and it is displayed as HTML. This HTML code does NOT convert a "<" into "<", otherwise you could not display any HTML.
But of course this code can be wrongly formatted, for example it can contain a </table> tag without a <table> tag before it. This will definitely spoil the display in ELOG, and that was what I was refering to.
I agree that html tags in plain text entries will not have this problem.

67848

Wed Apr 1 11:41:31 2015

tim.schelfhout@fixbrussel.be

Andreas Luedeke wrote:

67847

Wed Apr 1 11:01:14 2015

Tim Schelfhout

Thank you all of you for your notes ...
Anyway I have no time at the present to debug this ... at the level of configuration file.
I discovered however that the Incognito Mode of Chrome does NOT seem to have this problem??
Chrome version is 41.0.2272.101

Don't know why ??? ... tried disabling all of the added extensions in chrome but to no avail.
Don't know what else can be different in both modes.

Andreas Luedeke wrote:

Stefan Ritt wrote:

The (correct) display tells me that the colours are user-defined, probably by the configuration option

Style <attribute> <value> = <style>

which selects different styles for different rows. Now I do not know why your browser should change behaviour all over sudden, but I would double check the configuration. Like removing all style additions in the config file, then try again, then add one by one. There could also be a class defined with the "style" option which has not been added to the default style file themes directory of the elog installation.

Stefan

Just my two cent:
the content of a particular entry can change the rendering. If you have an HTML end tag like </table> or </font> in your entry, then the display after that entry may be spoiled.
You could try to select the very same entries in both browsers, to see if it does depend on the specific entry content.
If the problem persists, then I would suggest that you post the following:
- a minimal configuration for a logbook that reproduces the problem; and
- the actual entries, exported in XML or RAW format; and
- screenshots on how it displays with IE (add version number) and with Chrome (add version number).

Cheers
Andreas

67846

Wed Apr 1 10:54:27 2015

Andreas Luedeke

andreas.luedeke@psi.ch

Stefan Ritt wrote:

Andreas Luedeke wrote:

the content of a particular entry can change the rendering. If you have an HTML end tag like </table> or </font> in your entry, then the display after that entry may be spoiled.

Actually not. If you have HTML statements in entries, they will be rendered using escape characters (like </table> -> </table>). This is necessary to avoid cross-side-script vulnerabilities (XSS). If this is not working in some case, let me know and I have to fix it. The only exception is if you explicitly allow this via Allow HTML = 1

67845

Tue Mar 31 11:44:27 2015

Andreas Luedeke wrote:

the content of a particular entry can change the rendering. If you have an HTML end tag like </table> or </font> in your entry, then the display after that entry may be spoiled.

67844

Tue Mar 31 11:36:25 2015

Andreas Luedeke

andreas.luedeke@psi.ch

Stefan Ritt wrote:

67843

Mon Mar 30 17:48:06 2015