Am 24.01.2016 um 19:08 schrieb Georg Baum <georg.b...@post.rwth-aachen.de>:
>
> Stephan Witt wrote:
>
>> Am 24.01.2016 um 18:17 schrieb Georg Baum
>> <georg.b...@post.rwth-aachen.de>:
>>>
>>> Stephan, does the algo2e.tex test succeed if you apply the attached patch
>>> on a clean tree?
>>
>> No, it doesn’t work. With the patch I get:
>
> It turned out that we used two different names for the latin1 encoding:
> "ISO8859-1" and "ISO-8859-1" (iconv understands both). Does it work with the
> attached patch? It should also print all encoding changes, and for e.g.
> CJKutf8.tex you get several:
>
> Changing is encoding from ISO-8859-1 to UTF-8.
> Changing is encoding from UTF-8 to ISO-8859-15.
> Changing is encoding from ISO-8859-15 to UTF-8.
> Changing is encoding from UTF-8 to ISO-8859-15.
> Changing is encoding from ISO-8859-15 to UTF-8.
> Changing is encoding from UTF-8 to ISO-8859-15.
> Changing is encoding from ISO-8859-15 to UTF-8.
> Changing is encoding from UTF-8 to ISO-8859-15.
> Changing is encoding from ISO-8859-15 to UTF-8.
> Changing is encoding from UTF-8 to ISO-8859-15.
> Changing is encoding from ISO-8859-15 to UTF-8.
> Changing is encoding from UTF-8 to ISO-8859-15.
> Changing is encoding from ISO-8859-15 to UTF-8.
> Changing is encoding from UTF-8 to ISO-8859-15.
> Changing is encoding from ISO-8859-15 to UTF-8.
> Changing is encoding from UTF-8 to ISO-8859-15.
> Changing is encoding from ISO-8859-15 to UTF-8.
> Changing is encoding from UTF-8 to ISO-8859-15.
> Changing is encoding from ISO-8859-15 to UTF-8.
> Changing is encoding from UTF-8 to ISO-8859-15.
> Changing is encoding from ISO-8859-15 to UTF-8.
> Changing is encoding from UTF-8 to ISO-8859-15.
> Changing is encoding from ISO-8859-15 to UTF-8.
> Changing is encoding from UTF-8 to ISO-8859-15.
> Changing os encoding from UTF-8 to ISO-8859-15.
> Changing os encoding from ISO-8859-15 to UTF-8.
>
> The two last ones are strange, the output file should be completely utf8.
>
>> Wouldn’t it be better to start over from the beginning with the new
>> encoding in case it’s different from the current encoding (default or
>> given by command line option)?
>
> This is the idea, yes, but before implementing it I'd like to know whether
> it works in principle. This would require a two step approach for the case
> that no command line option is given: Parse in latin1 encoding until
> \usepackage[xxx]{inputenc} is found, close the file and open it again with
> the correct encoding. Another idea would be to combine the fix for this bug
> with the one for http://www.lyx.org/trac/ticket/8218: Guess the encoding of
> the file, and then open it in the guessed encoding. Both approaches are not
> a 100% solution, but would work for many cases.
Yes, this patch works for algo2e. Now the encoding was not changed anymore
for latin1 inputenc. Now we have:
57% tests passed, 12 tests failed out of 28
Label Time Summary:
cmplyx = 20.40 sec (14 tests)
roundtrip = 11.74 sec (14 tests)
Total Test time (real) = 32.20 sec
The following tests FAILED:
12 - tex2lyx/roundtrip/CJK.tex (Failed)
13 - tex2lyx/cmplyx/CJK.tex (Failed)
14 - tex2lyx/roundtrip/CJKutf8.tex (Failed)
15 - tex2lyx/cmplyx/CJKutf8.tex (Failed)
16 - tex2lyx/roundtrip/test-insets-basic.tex (Failed)
17 - tex2lyx/cmplyx/test-insets-basic.tex (Failed)
18 - tex2lyx/roundtrip/test-insets.tex (Failed)
19 - tex2lyx/cmplyx/test-insets.tex (Failed)
20 - tex2lyx/roundtrip/test-memoir.tex (Failed)
21 - tex2lyx/cmplyx/test-memoir.tex (Failed)
24 - tex2lyx/roundtrip/test-refstyle-theorems.tex (Failed)
25 - tex2lyx/cmplyx/test-refstyle-theorems.tex (Failed)
The complete tex2lyx output is attached.
Stephan
LastTest.log
Description: Binary data
