__/ [Peter D.] on Tuesday 01 November 2005 05:45 \__
> Roy Schestowitz wrote in alt.os.linux.mandrake:
>
>> __/ [TerryD] on Sunday 30 October 2005 19:48 \__
>>
>>> Hi All,
>>>
>>> I can go to a certain website then highlight and copy some text along
>>> with
>>> quotemarks. I can paste that clipboard contents into a Knode
>>> composition
>>> window and it looks 100% -- just like the original article. When I
>>> post my
>>> copied material, the quote marks turn into question marks. Question
>>> marks in the copied material turn out normal.
>>>
>>> Any idea what I need to do to fix this?
>>
>> I experience the same problem on occasions. I use KNode and I often
>> quote snippets from articles that are properly viewable in the Web
>> browser. The 'fancy' quotes appear fine (ones where the right-hand and
>> left-hand quotation marks are distinguishable) until posted and then
>> viewed.
>>
>> Whether only KNode is unable to render/interpret these, I do not know
>> for sure, but I suspect not. I suppose that the best thing to do is to
>> remember to manually change these quotes before posting. These are
>> fine for HTML encodings and I agree that KNode should not render
>> things in the 'compose' window if these would not be gracefully
>> accepted by the news servers. It is somewhat misleading as WYSINWYG.
>>
>> Roy
>
> That looks like the information about the character set of the cut
> and pasted text is known to the display manager, but it is not
> getting to KNode and the final file. If the information is available
> then a sensible translation should take place automatically.
>
> That would have to be worth a wish list entry, maybe it is a bug.
> Consider clicking on Help -> report a bug.
I went to Microsoft.com, but there was no Help and no entry called "Report a
bug" (see below)...
__/ [Robert M. Riches Jr.] on Monday 31 October 2005 17:47 \__
> Yes, email the webmaster of the offending site and tell them
> to fix their text. :-(
>
> In my experience, this kind of thing is caused by web sites
> using strange non-ASCII characters, probably as a result of
> infestation of their development software by malware from
> Microsoft Corp.
>
> I think it was in comp.os.linux.misc that someone recently
> mangled (maybe deliberately) a suggested command line to try
> to make command-line interfaces look bad. Several
> characters, mostly spaces and quotation marks, had bit 7 set
> in the mangled version. The original suggested command line
> worked fine, but the mangled one created some very strange
> error messages.
>
> To test that theory, my suggestion would be to pull up a
> terminal window, do "cat > somefile1" and paste in the text
> in question, followed by a control-D (for EOF). Then, run
> 'od -xc' on the file, maybe piping it through more or less.
>
> If the problem is as I suspect, it would be reasonably easy
> to repair text after the fact. If you had a list of the
> mangled characters, it should be possible to concoct a tr
> command to do the job. There may be pre-packaged solutions,
> but I'm not aware of any. Google might help find one.
So, once again, proprietary cripples communication. Trying to 'fix' this
in KNode would not make a fix, but rather a workaround or a hack for non--
standardised bugs that lie elsewhere.
It is somewhat like browsers that must tolerate and cope with MSOffice-
generated HTML. Firefox appears to do it admirably well, but why should
it? Why should developers devour time embracing Microsoft non-standards-
compliant HTML?
Roy
|
|
