Dear Scott and Kornel,

On 2016-09-21, Scott Kostyshak wrote:
> On Tue, Sep 20, 2016 at 05:51:36PM +0200, Kornel Benko wrote:
>> Am Dienstag, 20. September 2016 um 10:56:29, schrieb Scott Kostyshak 
>> <>
>> > On Tue, Sep 20, 2016 at 06:28:48AM +0000, Guenter Milde wrote:
>> > > On 2016-09-20, Scott Kostyshak wrote:

>> > > The documentation is for documentation, test-use is secondary and
>> > > "exotic tests" don't merit changes that make the documentation
>> > > more difficult to read.
>> > 
>> > I agree this is the right policy. If I convert those comments to a LyX
>> > note or comment inset, this policy is still satisfied because the inset
>> > is just as readable, right? So either:
>> > 
>> > (1) revert the commit I pushed and invert the test

This is the simplest way.
Especially, because we already have a similar case:

diff --git a/development/autotests/invertedTests 
index 8dab991..8ee6a85 100644
--- a/development/autotests/invertedTests
+++ b/development/autotests/invertedTests
@@ -100,6 +100,7 @@ Sublabel: ert
 # Non-ASCII in ERT, fails with inputenc==ASCII (e.g. XeTeX with tex-fonts)
 # inputencoding="utf8-plain" with Xe/LuaTeX: characters with
 # Unicode point > 256 lead to errors with 8-bit fonts

>> Please no. If we never interpret ERT, then this will stay inverted
>> forever. Inverted tests are waiting for correction, at least this was
>> what I have/had in mind.

> I tend to agree. Something feels wrong with inverting. Unless we label
> the issue as a LyX enhancement (and create a trac ticket) because there
> is no way in LyX to produce LyX-readable content in this case that can
> be exported to several different formats.


a) this is something waiting for correction (with low priority, because
   XeTeX + TeX-fonts is "exotic").

   Possible changes include:
   * do not check ERT for unsupported characters (after all, the user is
     responsible for ERT, this "helpfull" feature stands in the way quite
     regularely). However, there is no agreement, other developers prefer
     checking ERT.
   * fix "inputenc" LaTeX package to allow XeTeX and set it to binary mode
   * write a "xeinputenc" package to allow XeTeX and set it to binary mode
   * fix to make
     non-TeX fonts default for compiling with XeTeX.
     (This would not solve the test but make the problem even less urgent.)
   * Change the ERT-comment to a LyX note eventually

b) "inverting this test" (i.e. adding a pattern for this test to
   invertedTests)¹ will

   * add the "WILL_FAIL" test feature
   * add the label "suspended"
   * not add the label "export"

   because the test matches the "catchall pattern" for _texF in

¹ For me, this shows that the test system is too complicated: how
  should I call the action so that everyone understands?

>> > or
>> > 
>> > (2) change the inset for all beamer manuals.

>> +1

> OK I will go for this. I'll wait another day or so to see if Günter
> disagrees.

IMV there are more important tasks but I won't stop you if you like to go
this way.

Actually, only the French manual has this comment in ERT, all other
instances (en, de, ja) just have the "\setbeamercovered{transparent}"
LaTeX macro.

If you fix the beamer manual *and* Additional.lyx, consider a
minimal example for /autotests/export/latex/ and a pattern to
invertedTests like

 # Non-ASCII in ERT, fails with inputenc==ASCII (e.g. XeTeX with tex-fonts)


Reply via email to