tests fail on MSYS2

[neheb]

No real idea what's happening.

[neheb]

Another problem related to the french locale test:

--- test3outputB	2022-06-14 18:37:06.559941900 -0700
+++ testtry	2022-06-14 18:37:06.594847800 -0700
@@ -87,8 +87,9 @@
 Capture group count = 0
 Starting code units: 0 1 2 3 4 5 6 7 8 9 A B C D E F G H I J K L M N O P 
   Q R S T U V W X Y Z _ a b c d e f g h i j k l m n o p q r s t u v w x y z 
-  ª µ º À Á Â Ã Ä Å Æ Ç È É Ê Ë Ì Í Î Ï Ð Ñ Ò Ó Ô Õ Ö Ø Ù Ú Û Ü Ý Þ ß à á â 
-  ã ä å æ ç è é ê ë ì í î ï ð ñ ò ó ô õ ö ø ù ú û ü ý þ ÿ 
+  ƒ Š Œ Ž š œ ž Ÿ ª ² ³ µ ¹ º À Á Â Ã Ä Å Æ Ç È É Ê Ë Ì Í Î Ï Ð Ñ Ò Ó Ô Õ Ö 
+  Ø Ù Ú Û Ü Ý Þ ß à á â ã ä å æ ç è é ê ë ì í î ï ð ñ ò ó ô õ ö ø ù ú û ü ý 
+  þ ÿ 
 Subject length lower bound = 1
 
 # All remaining tests are in the french locale, so set the default.
@@ -147,7 +148,7 @@
 /[[:alpha:]][[:lower:]][[:upper:]]/IB
 ------------------------------------------------------------------
         Bra
-        [A-Za-z\x83\x8a\x8c\x8e\x9a\x9c\x9e\x9f\xaa\xb5\xba\xc0-\xd6\xd8-\xf6\xf8-\xff]
+        [A-Za-z\x83\x8a\x8c\x8e\x9a\x9c\x9e\x9f\xaa\xb2\xb3\xb5\xb9\xba\xc0-\xd6\xd8-\xf6\xf8-\xff]
         [a-z\x83\x9a\x9c\x9e\xaa\xb5\xba\xdf-\xf6\xf8-\xff]
         [A-Z\x8a\x8c\x8e\x9f\xc0-\xd6\xd8-\xde]
         Ket
@@ -155,9 +156,9 @@
 ------------------------------------------------------------------
 Capture group count = 0
 Starting code units: A B C D E F G H I J K L M N O P Q R S T U V W X Y Z 
-  a b c d e f g h i j k l m n o p q r s t u v w x y z ª µ º À Á Â Ã Ä Å Æ Ç 
-  È É Ê Ë Ì Í Î Ï Ð Ñ Ò Ó Ô Õ Ö Ø Ù Ú Û Ü Ý Þ ß à á â ã ä å æ ç è é ê ë ì í 
-  î ï ð ñ ò ó ô õ ö ø ù ú û ü ý þ ÿ 
+  a b c d e f g h i j k l m n o p q r s t u v w x y z ƒ Š Œ Ž š œ ž Ÿ ª ² ³ 
+  µ ¹ º À Á Â Ã Ä Å Æ Ç È É Ê Ë Ì Í Î Ï Ð Ñ Ò Ó Ô Õ Ö Ø Ù Ú Û Ü Ý Þ ß à á â 
+  ã ä å æ ç è é ê ë ì í î ï ð ñ ò ó ô õ ö ø ù ú û ü ý þ ÿ 
 Subject length lower bound = 3
 
 # End of testinput3 

French locale is available.

[zherczeg]

I suspect the directories are simplified somehow, so the test pass regardless of the current working directory. This simplification is probably not done.

[neheb]

if you're referring to the first example, MSYS2 emulates a POSIX environment. The path looks like

C:\msys2\home\x\

The code filters it to

C:\msys2\

which is wrong.

[NWilson]

@neheb Thank you Rosen for reporting this. I have investigated, and opened a PR with the fix (#660).

The issue is to do with the way Windows handles wildcard expansion. When you run command *.txt in CMD or Batch scripts, the glob is not expanded by the shell as it is on Unix: applications are expected to do this themselves (using the same FindNextFile Win32 function that is used for directory traversal).

pcre2grep does already have code to do this on Windows - however, it uses #ifdef to check first for the sys/stat.h header, and then checks for whether WIN32 is defined. On MSYS, this resulted in the Unix codepath being taken, since Unix headers are available in addition to Win32 headers.

The fix is a one-liner to ensure that the Win32 implementation is used (for the UCRT64 and other MinGW-x64 environments). The "MSYS" environment of MSYS2 (which is cygwin-like, and produces applications that are linked to the cygwin runtime) still uses the Cygwin-emulated Unix environment.

[neheb]

cool.