Mercurial > octave
changeset 31149:7d4cf04665e6
oruntests.m: Convert read file from input encoding (bug #62780).
* scripts/testfun/oruntests.m (has_functions, has_tests): Convert read binary
data from input encoding to UTF-8.
* test/file-encoding: Add tests for this use case.
| author | Markus Mützel <markus.muetzel@gmx.de> |
|---|---|
| date | Wed, 20 Jul 2022 16:37:58 +0200 |
| parents | 94998332d787 |
| children | 60f32b7455e3 |
| files | scripts/testfun/oruntests.m test/file-encoding/CP1251/test_CP1251.m test/file-encoding/file-encoding.tst test/file-encoding/test_CP1252.m |
| diffstat | 4 files changed, 65 insertions(+), 0 deletions(-) [+] |
line wrap: on
line diff
--- a/scripts/testfun/oruntests.m Mon Jul 18 17:30:08 2022 +0200 +++ b/scripts/testfun/oruntests.m Wed Jul 20 16:37:58 2022 +0200 @@ -119,6 +119,11 @@ endif str = fread (fid, "*char")'; fclose (fid); + enc = dir_encoding (fileparts (f)); + if (! strcmp (enc, "utf-8")) + ## convert from input encoding to UTF-8 + str = native2unicode (uint8 (str), enc); + endif retval = ! isempty (regexp (str,'^(?:DEFUN|DEFUN_DLD|DEFUNX)\>', 'lineanchors', 'once')); elseif (n > 2 && strcmpi (f((end-1):end), ".m")) @@ -138,6 +143,11 @@ str = fread (fid, "*char").'; fclose (fid); + enc = dir_encoding (fileparts (f)); + if (! strcmp (enc, "utf-8")) + ## convert from input encoding to UTF-8 + str = native2unicode (uint8 (str), enc); + endif retval = ! isempty (regexp (str, '^%!(assert|error|fail|test|xtest|warning)', 'lineanchors', 'once'));
--- a/test/file-encoding/CP1251/test_CP1251.m Mon Jul 18 17:30:08 2022 +0200 +++ b/test/file-encoding/CP1251/test_CP1251.m Wed Jul 20 16:37:58 2022 +0200 @@ -31,3 +31,7 @@ function test_CP1251 () help ("test_CP1251"); endfunction + +%!assert (double ("ÄÖÜäöü ŠŽšž"), +%! [208 148 208 166 208 172 208 180 209 134 209 140 32 ... +%! 208 137 208 139 209 153 209 155])
--- a/test/file-encoding/file-encoding.tst Mon Jul 18 17:30:08 2022 +0200 +++ b/test/file-encoding/file-encoding.tst Wed Jul 20 16:37:58 2022 +0200 @@ -23,6 +23,10 @@ ## ######################################################################## +## Note: The cache of dir_encoding from .oct-config files in the load path +## persists even after removing the folder from the load path. +## Explictily, delete it when removing the path. + ## test file in current directory %!assert (dir_encoding ("."), "windows-1252") @@ -48,6 +52,7 @@ %! assert (dir_encoding (fullfile (pwd (), "CP1251")), "windows-1251"); %! unwind_protect_cleanup %! path (path_orig); +%! dir_encoding (canonicalize_file_name ("CP1251"), "delete"); %! end_unwind_protect %!test @@ -63,6 +68,7 @@ %! assert (strfind (help_str, ref_str)); %! unwind_protect_cleanup %! path (path_orig); +%! dir_encoding (canonicalize_file_name ("CP1251"), "delete"); %! end_unwind_protect @@ -79,6 +85,7 @@ %! assert (dir_encoding (fullfile (pwd (), "CP1251")), "windows-1251"); %! unwind_protect_cleanup %! path (path_orig); +%! dir_encoding (canonicalize_file_name ("CP1251"), "delete"); %! end_unwind_protect %!test @@ -95,5 +102,45 @@ %! assert (strfind (help_str, ref_str)); %! unwind_protect_cleanup %! path (path_orig); +%! dir_encoding (canonicalize_file_name ("CP1251"), "delete"); +%! end_unwind_protect + + +## oruntests with file in current folder with .oct-config file +%!test <*62780> +%! ## wrap in "evalc" to suppress output to the log +%! evalc ('oruntests (".");'); + +## oruntests with file in different folder (not in load path) with +## "dir_encoding" +%!test <*62780> +%! unwind_protect +%! dir_encoding (canonicalize_file_name ("CP1251"), "windows-1251"); +%! ## use "evalc" to suppress output to the log +%! evalc ('oruntests ("CP1251");'); +%! unwind_protect_cleanup +%! dir_encoding (canonicalize_file_name ("CP1251"), "delete"); %! end_unwind_protect +## oruntests with file in different folder (not in load path) with +## "__mfile_encoding__" +%!test <*62780> +%! old_mfile_encoding = __mfile_encoding__ ("windows-1251"); +%! unwind_protect +%! ## use "evalc" to suppress output to the log +%! evalc ('oruntests ("CP1251");'); +%! unwind_protect_cleanup +%! __mfile_encoding__ (old_mfile_encoding); +%! end_unwind_protect + +## oruntests with file in different folder with .oct-config file (in load path) +%!test <*62780> +%! path_orig = path (); +%! unwind_protect +%! addpath (canonicalize_file_name ("CP1251")); +%! ## use "evalc" to suppress output to the log +%! evalc ('oruntests ("CP1251");'); +%! unwind_protect_cleanup +%! path (path_orig); +%! dir_encoding (canonicalize_file_name ("CP1251"), "delete"); +%! end_unwind_protect
--- a/test/file-encoding/test_CP1252.m Mon Jul 18 17:30:08 2022 +0200 +++ b/test/file-encoding/test_CP1252.m Wed Jul 20 16:37:58 2022 +0200 @@ -31,3 +31,7 @@ function test_CP1252 () help ("test_CP1252"); endfunction + +%!assert (double ("ÄÖÜäöü ŠŽšž"), +%! [195 132 195 150 195 156 195 164 195 182 195 188 32 ... +%! 197 160 197 189 197 161 197 190])