branch master updated: * tp/tests/encoded/list-of-tests (verbatiminclude_names_latin1_html) (verbatiminclude_names_latin1_html_explicit_encoding): add -c MESSAGE_ENCODING=UTF-8 to have a reproducible output. With ASCII output encoding, there is no guarantee on the representation of non-ASCII characters. Gavin idea.

[Patrice Dumas]

This is an automated email from the git hooks/post-receive script.

pertusus pushed a commit to branch master
in repository texinfo.

The following commit(s) were added to refs/heads/master by this push:
     new 6a2bc80a51 * tp/tests/encoded/list-of-tests 
(verbatiminclude_names_latin1_html) 
(verbatiminclude_names_latin1_html_explicit_encoding): add -c 
MESSAGE_ENCODING=UTF-8 to have a reproducible output.  With ASCII output 
encoding, there is no guarantee on the representation of non-ASCII characters.  
Gavin idea.
6a2bc80a51 is described below

commit 6a2bc80a51216340f43b751d14987cd51c884469
Author: Patrice Dumas <pertusus@free.fr>
AuthorDate: Wed Oct 30 20:32:42 2024 +0100

    * tp/tests/encoded/list-of-tests (verbatiminclude_names_latin1_html)
    (verbatiminclude_names_latin1_html_explicit_encoding): add -c
    MESSAGE_ENCODING=UTF-8 to have a reproducible output.  With ASCII
    output encoding, there is no guarantee on the representation of
    non-ASCII characters.  Gavin idea.
---
 ChangeLog                                          |  8 +++++++
 tp/tests/encoded/list-of-tests                     | 25 ++++++++++++++--------
 .../verbatiminclude_names_latin1.2                 |  8 +++----
 .../verbatiminclude_names_latin1.2                 |  4 ++--
 4 files changed, 30 insertions(+), 15 deletions(-)

diff --git a/ChangeLog b/ChangeLog
index 1a1ca72d24..f5c2c166d1 100644
--- a/ChangeLog
+++ b/ChangeLog
@@ -1,3 +1,11 @@
+2024-10-30  Patrice Dumas  <pertusus@free.fr>
+
+       * tp/tests/encoded/list-of-tests (verbatiminclude_names_latin1_html)
+       (verbatiminclude_names_latin1_html_explicit_encoding): add -c
+       MESSAGE_ENCODING=UTF-8 to have a reproducible output.  With ASCII
+       output encoding, there is no guarantee on the representation of
+       non-ASCII characters.  Gavin idea.
+
 2024-10-30  Patrice Dumas  <pertusus@free.fr>
 
        Add contrib/init_examples directory for texi2any initialization files
diff --git a/tp/tests/encoded/list-of-tests b/tp/tests/encoded/list-of-tests
index f667e6124e..70ff0005ce 100644
--- a/tp/tests/encoded/list-of-tests
+++ b/tp/tests/encoded/list-of-tests
@@ -3,10 +3,21 @@
 # ./texi2any.pl --footnote-style=bâd
 # ./texi2any.pl --paragraph-indent=ïndent
 
+# NOTE in the default case, the locale set for the tests corresponds to
+# an ASCII encoding.  In that setup, non-ASCII encoded characters are usually
+# treated as invalid characters.  Perl Encode::encode use a substitution
+# character which seems to be "?", but there is no guarantee.  With XS/C
+# and GNU iconv, there also seems to be a "?" output but there are also
+# "encoding error at byte XXX" errors output.  For XS/C it could probably
+# be possible to have "?" output reproducibly by replacing encode_with_iconv by
+# gnulib striconveh module mem_cd_iconveh with iconv_ilseq_handler set to
+# iconveh_question_mark.
+#
+# Since there is no guarantee on the non-ASCII character representations
+# in error messages with an ASCII encoding, we set MESSAGE_ENCODING=UTF-8
+# for all the tests.
+
 # check non ascii in verbatiminclude file name
-# TODO with XS there are "encoding error at byte XXX" output.
-# Maybe replace encode_with_iconv by gnulib striconveh module
-# mem_cd_iconveh with iconv_ilseq_handler set to iconveh_question_mark?
 #no_command_line_encoding osé_utf8.texi --html
 
 # check non ascii command line arguments and css files
@@ -42,13 +53,9 @@ manual_include_accented_file_name_latin1_use_locale_encoding 
manual_include_acce
 # The following tests tests how verbatiminclude file name encoding is
 # resolved in raw text in HTML and in raw text converter and how
 # customization options are used.
-# Note that the tests are not perfect, as UTF-8 corresponds to the
-# locale encoding, so the encoding of the file names could be found
-# either from the command line or the locale, depending on the value
-# of DOC_ENCODING_FOR_INPUT_FILE_NAME as seen from the converters.
-verbatiminclude_names_latin1_html verbatiminclude_names_latin1.texi --html
+verbatiminclude_names_latin1_html verbatiminclude_names_latin1.texi --html -c 
MESSAGE_ENCODING=UTF-8
 
-verbatiminclude_names_latin1_html_explicit_encoding 
verbatiminclude_names_latin1.texi --html -c INPUT_FILE_NAME_ENCODING=UTF-8
+verbatiminclude_names_latin1_html_explicit_encoding 
verbatiminclude_names_latin1.texi --html -c INPUT_FILE_NAME_ENCODING=UTF-8 -c 
MESSAGE_ENCODING=UTF-8
 
 verbatiminclude_names_latin1_rawtext verbatiminclude_names_latin1.texi -c 
TEXINFO_OUTPUT_FORMAT=rawtext
 
diff --git 
a/tp/tests/encoded/res_parser/verbatiminclude_names_latin1_html/verbatiminclude_names_latin1.2
 
b/tp/tests/encoded/res_parser/verbatiminclude_names_latin1_html/verbatiminclude_names_latin1.2
index 46e75d054c..81466777d5 100644
--- 
a/tp/tests/encoded/res_parser/verbatiminclude_names_latin1_html/verbatiminclude_names_latin1.2
+++ 
b/tp/tests/encoded/res_parser/verbatiminclude_names_latin1_html/verbatiminclude_names_latin1.2
@@ -1,4 +1,4 @@
-verbatiminclude_names_latin1.texi:6: @verbatiminclude: could not find 
included_ak?ent?d.texi
-verbatiminclude_names_latin1.texi:8: @verbatiminclude: could not find 
vi_not_exist?ng.t?xi
-verbatiminclude_names_latin1.texi:18: @verbatiminclude: could not find 
included_ak?ent?d.texi
-verbatiminclude_names_latin1.texi:20: @verbatiminclude: could not find 
vi_not_exist?ng.t?xi
+verbatiminclude_names_latin1.texi:6: @verbatiminclude: could not find 
included_akçentêd.texi
+verbatiminclude_names_latin1.texi:8: @verbatiminclude: could not find 
vi_not_existïng.téxi
+verbatiminclude_names_latin1.texi:18: @verbatiminclude: could not find 
included_akçentêd.texi
+verbatiminclude_names_latin1.texi:20: @verbatiminclude: could not find 
vi_not_existïng.téxi
diff --git 
a/tp/tests/encoded/res_parser/verbatiminclude_names_latin1_html_explicit_encoding/verbatiminclude_names_latin1.2
 
b/tp/tests/encoded/res_parser/verbatiminclude_names_latin1_html_explicit_encoding/verbatiminclude_names_latin1.2
index c162bd5eff..9caef1067f 100644
--- 
a/tp/tests/encoded/res_parser/verbatiminclude_names_latin1_html_explicit_encoding/verbatiminclude_names_latin1.2
+++ 
b/tp/tests/encoded/res_parser/verbatiminclude_names_latin1_html_explicit_encoding/verbatiminclude_names_latin1.2
@@ -1,2 +1,2 @@
-verbatiminclude_names_latin1.texi:8: @verbatiminclude: could not find 
vi_not_exist?ng.t?xi
-verbatiminclude_names_latin1.texi:20: @verbatiminclude: could not find 
vi_not_exist?ng.t?xi
+verbatiminclude_names_latin1.texi:8: @verbatiminclude: could not find 
vi_not_existïng.téxi
+verbatiminclude_names_latin1.texi:20: @verbatiminclude: could not find 
vi_not_existïng.téxi