Fixed for msys/cygwin environments #117
Conversation
…nce the latter properly escapes command line arguments in msys environment. Changed win32/unix detection to rely on presence of "uname", to cover msys/cygwin environments. sys.platform returns "win32" in msys case, tricking the setup script into performing unnecessary escaping of MODULE_NAME macro.
| @@ -68,7 +68,13 @@ | |||
| It is almost fully compliant with the Python database API version 2.0 also | |||
| exposes the unique features of SQLite.""" | |||
|
|
|||
| if sys.platform != "win32": | |||
| try: | |||
| subprocess.check_output(["uname", "-a"]) | |||
There was a problem hiding this comment.
There is no specific reason to use check_output() vs check_call(). check_call() would actually be more preferable.
The "-a" argument also is excessive and can be removed.
| @@ -29,7 +29,7 @@ | |||
| print("Only Python 2.7 is supported.") | |||
| sys.exit(1) | |||
|
|
|||
| import commands | |||
| import subprocess | |||
There was a problem hiding this comment.
subprocess came in 2.4 (https://docs.python.org/2/library/subprocess.html#module-subprocess), so not sure how backwards compatible this change would be. Not following pysqlite history, so no idea if this is something to worry about.
| @@ -122,8 +128,7 @@ class MyBuildExt(build_ext): | |||
| amalgamation = False | |||
|
|
|||
| def _pkgconfig(self, flag, package): | |||
| status, output = commands.getstatusoutput("pkg-config %s %s" % (flag, package)) | |||
There was a problem hiding this comment.
This was giving me "{ is not a recognized command" or something like that under msys. Looked up getstatusoutput (https://docs.python.org/2/library/commands.html) and they say this:
cmd is actually run as { cmd ; } 2>&1, so that the returned output will contain output or error messages
So that would be where the "{" "not found" comes from.
| @@ -68,7 +68,13 @@ | |||
| It is almost fully compliant with the Python database API version 2.0 also | |||
| exposes the unique features of SQLite.""" | |||
|
|
|||
| if sys.platform != "win32": | |||
There was a problem hiding this comment.
Under msys you end up with "win32" here
Replaced commands.getstatusoutput() with subprocess.check_output() since the latter properly escapes command line arguments in msys environment.
Changed win32/unix detection to rely on presence of "uname", to cover msys/cygwin environments. Not sure if this would be the best way to detect unix shell, but the problem is that sys.platform returns "win32" in msys case, tricking the setup script into performing unnecessary escaping of MODULE_NAME macro.