On Sat, Sep 30, 2023 at 10:21:13PM -0700, Mike Fulton wrote:
> Hi,
>
> I am trying to understand why testcase ./t/c-u-m-x-scroll-forward.sh fails
> for me on z/OS.
> From the trace, I see that it is reading the characters from the
> pseudotty in and correctly
> processing them, until it gets to the 'print node' sequence which is coded
> as:
>
> \rDq
>
> This sequence isn't printing out the node and so I fail when it tries to:
>
> diff t/c-u-m-x-scroll-forward.sh.out ./t/node-target
>
> because the t/c-u-m-x-scroll-forward.sh.out isn't created.
>
> I have confirmed that I can interactively use M-x print-node to print the
> node out if
> I have done:
> export INFO_PRINT_COMMAND='>/tmp/out'
> first, which seems consistent with the docs.
>
> Hoping someone can provide some guidance here.
>
> thanks, mike
Could you send the file c-u-m-x-scroll-forward.sh.log to confirm that
it is getting that far?
Yes. I have attached it
"print-node" is used in many of the info tests so this one isn't different.
It is odd that no output at all is being output for it, though.
I have a number of tests fail - of the 85 I run, 49 pass and 36 fail.
This is the first one I am investigating.
The only thing that I expect may be different about this test is setting
the screen geometry with the LINES and COLUMNS environment variables.
You could check what happens with the TIOCWINSZ ioctl in pseudotty.c.
Look at what it says in the log file after "./pseudotty: attempting to
set window size".
That appears to work. Here is that part of the output
pseudotty: getting pty master fd
pseudotty: unlocking slave device
pseudotty: getting file name of slave device...
pseudotty: /dev/ttyp0029
pseudotty: opening slave device
pseudotty: attempting to set window size
pseudotty: ...succeeded
pseudotty: opening control channel
Another way of investigating this would be to insert sleep commands, like
diff --git a/info/t/c-u-m-x-scroll-forward.sh b/info/t/c-u-m-x-scroll-forward.sh
index bf3038179a..1b71df8d17 100755
--- a/info/t/c-u-m-x-scroll-forward.sh
+++ b/info/t/c-u-m-x-scroll-forward.sh
@@ -31,6 +31,7 @@ run_ginfo -f intera -n 'Scroll four lines'
# C-u M-x scroll-forward should scroll four lines, not four screens
printf '\025\033xscroll-forward\r' >&7
+sleep 1
printf '\rDq' >&7
timeout_test
sleep 1 did not make a difference unfortunately.
This would help to discover if there were timing-related problems.
c-u-m-x-scroll-forward.sh.log