Hye-Shik Chang wrote:

On 10/8/05, Nick Coghlan <ncoghlan@iinet.net.au> wrote:

...

Anyone else seeing any problems with test_cmd_line? I've got a few failures in test_cmd_line on Kubuntu 5.10 with GCC 4.0 relating to a missing "\n" line ending.

Same problem here. (FreeBSD 6.0 with GCC 3.4.4) In my short inspection, popen2.popen4.read() returned just an empty string.

Good to know it isn't just a system quirk, as that's the same behaviour I'm getting. I noticed that the ones which appear to be failing (-E, -O, -S, -Q) are the ones which expect an interactive session to open. The tests which pass (-V, -h, directory as argument or stdin) are the ones which don't actually start the interpreter. If I explicitly write Ctrl-D to the subprocess's stdin for the tests which open the interpreter, then the tests pass. So it looks like some sort of buffering problem with standard out not getting flushed before the test tries to read the data. Cheers, Nick. -- Nick Coghlan | ncoghlan@gmail.com | Brisbane, Australia --------------------------------------------------------------- http://boredomandlaziness.blogspot.com

On 10/8/05, Neal Norwitz <nnorwitz@gmail.com> wrote:

On 10/8/05, Nick Coghlan <ncoghlan@gmail.com> wrote:

...

Hye-Shik Chang wrote:

...

On 10/8/05, Nick Coghlan <ncoghlan@iinet.net.au> wrote:

...

Anyone else seeing any problems with test_cmd_line? I've got a few failures in test_cmd_line on Kubuntu 5.10 with GCC 4.0 relating to a missing "\n" line ending.

If I explicitly write Ctrl-D to the subprocess's stdin for the tests which open the interpreter, then the tests pass. So it looks like some sort of buffering problem with standard out not getting flushed before the test tries to read the data.

Sorry, that's a new test I added recently. It works for me on gentoo. The test is very simple and shouldn't be hard to fix. Can you fix it? I assume Guido (or someone) added you as a developer. If not, if you can give me enough info, I can try to fix it.

I guess Neil's test was expecting at least one line of output from python at all times, but on most systems it is completely silent when the input is empty. I fixed the test (also in 2.4) to allow empty input as well as input ending in \n. -- --Guido van Rossum (home page: http://www.python.org/~guido/)