Hervé BOUTEMY wrote:
Le lundi 07 septembre 2009, Lukas Theussl a écrit :
Hi Guys,
Just some remarks:
\u0092 is not a valid xhtml character entity reference [1] (ie it is not
defined by the xhtml 1.0 DTD),
not really.
\u0092 has no named character entity, but it can be represented as a numeric
character reference [2]
so IMO MPIR should not use it in the first
place.
when it was done in MPIR-59, it worked well with HTML output. At that time, we
considered it as a great workaround :)
I think that \u0027 (apos) should be used instead but MPIR-136
states that this leads to test failures?
Yes, simple quote caused a problem with maven-reporting-impl 2.0.4.1.
I just renamed the issue from MNG-2855 to MSHARED-130 to better track the
issue (fixed in 2.0.4.2)
I'd guess that this should be
fixed instead?
Yes, it was fixed in maven-project-info-report-plugin 2.1.2 MPIR-136: there is
not more Unicode \u0092 hack, simple quote doesn't break the french locale
report
Then for people wanting to use french locale of maven-project-info-report-
plugin:
- it simply doesn't work with 2.0.1
- it works for HTML site with 2.1 and 2.1.1, but cause a problem with pdf
(quote rendered as #)
- it works nicely with 2.1.2+
So in other words, r811631 is not needed anymore, just use pir 2.1.2 and
everything is fine, or is there some downside?
In any case, I also don't like it if we arbitrarily re-write some
characters just to work around a bug elsewhere. If a user wants to use this
entity (ie declares it in an external DTD and gets the right font), then he
may very well complain if it gets replaced by a 'too clever' pdf plugin.
+1 totally agree
Finally, in case you are not aware, there is a test document in the pdf
plugin to render all xhtml entities, check the file
target/test-output/pdf/unnamed.pdf. As you see there are a few characters
rendered as # (specifically for me: U+2032, U+2033, U+203E, U+2308, U+2309,
U+230A, U+230B, 7 out of 253). I always attributed that to missing fonts
but maybe it's actually a fop issue..
I must admit I tried to find where the problem lies (to submit a patch to FOP
if necessary) but was not able to really understand. Giving up...
Me too... :( I just found out that U+2032, U+2033 render with the Symbol font but
there are a couple of characters I couldn't get to display with any font: zwnj
(0x200C), zwj (0x200D), lrm (0x200E), rlm (0x200F), oline (0x203E), lceil
(0x2038), rceil (0x2039), lfloor (0x203A), rfloor (0x203B). I don't think they are
used very frequently...
-Lukas
Regards,
Hervé
Cheers,
-Lukas
[1]
http://en.wikipedia.org/wiki/List_of_XML_and_HTML_character_entity_referenc
es
[2] http://en.wikipedia.org/wiki/Unicode_and_HTML
Vincent Siveton wrote:
Hi Hervé,
2009/9/5 Hervé BOUTEMY <[email protected]>:
I'm not convinced this is a good idea: MPIR is fixed now, but this hack
will prevent anybody to output \u0092 when it is the real character they
want.
Using \u0092 char will be displayed as # in the pdf so I don't think
user want to use this char.
MPIR 2.1.x uses \u0092 instead of quote due to an old hack MPIR-59 so
IMHO we need to add this workaround to correctly generate PDF in
French.
Cheers,
Vincent
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
