Re: [platform-testers] texinfo pretest

[Gavin Smith]

Many thanks for the thorough report.

> below are few results from various platforms.
>
> regards,
>  - assaf
>

> No failures with the following platforms, though many tests were skipped
> (~200).
> Perhaps an optional yet important package was missing.
>     centos 6.5, 7.0
>     debian 7.6
>     gNewSense 3.1
>     Ubuntu 14.04, 15.04
>     Trisquel 6.0.1, 7
>     Hurd 0.6/Debian/i386

It could be a missing Perl module.

> On Fedora 20, 21 - 1 test failed: t/index-apropos.sh
>
>      FAIL: t/index-apropos.sh
>      ========================

I'm not sure what's gone wrong here.

I'm going to change the pseudotty program to log what bytes it
actually manages to receive in case any are going missing.

>           ./pseudotty: getting pty master fd
>      ./pseudotty: unlocking slave device
>      ./pseudotty: getting file name of slave device...
>      ./pseudotty: /dev/pts/1
>      ./pseudotty: entering main loop
>      Fetching PID of ginfo process under test...
>      info: "./t/Infokey-config", line 2: unknown action `xxx-not-recognized'
>      info: "./t/Infokey-config", line 3: cannot bind key sequence to
> menu-digit
>      ...7780
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      ./pseudotty: trying to read
>      Program timed out after 3 seconds
>      ./t/Init-inter.inc: line 117:  7780 Terminated              $GINFO "$@"
>      ./t/Init-test.inc: line 65:  7773 Terminated              sleep 10 3>
> $PTY_TYPE
>      ./t/Init-test.inc: line 65:  7774 Terminated              ./pseudotty >
> $PIPEIN 3< $PTY_TYPE
>      FAIL t/index-apropos.sh (exit status: 1)
>
>
>

> On OpenSuSE 13.2 - 2 tests failed:  t/index-apropos.sh ,
> t/c-u-m-x-scroll-forward.sh
>
>     FAIL: t/c-u-m-x-scroll-forward.sh
>     =================================

I fear that this may because C-u is being interpreted as the terminal
kill byte. I'll have a think about this.

>         FAIL: t/index-apropos.sh
>     ========================

Possibly the same problem as before, whatever it was.

> On FreeBSD 9.3 - build failed due to "Error expanding embedded variable." -
> perhaps 'make' related.
>
>     $ make
>     ...
>     Making all in install-info
>     Making all in .
>     gcc -DHAVE_CONFIG_H -I. -I..  -I..
> -I../gnulib/lib                 -I../gnulib/lib
> -DLOCALEDIR=\"/usr /local/share/locale\"  -D_THREAD_SAFE
> -I/usr/local/include  -g -O2 -MT install-info.o -MD -MP -MF
> .deps/install-info.Tpo -c -o install-info.o install-info.c
>     mv -f .deps/install-info.Tpo .deps/install-info.Po
>     gcc  -g -O2   -o ginstall-info install-info.o ../gnulib/lib/libgnu.a
> /usr/local/lib/libintl.so -L/usr/local/lib /usr/local/lib/libiconv.so
> -Wl,-rpath -Wl,/usr/local/lib
>     Making all in tests
>     Making all in po
>     Error expanding embedded variable.
>     *** [all-recursive] Error code 1

I'm trying to search the Web for information but everyone has said so
far "use GNU make instead". :-(

It could be the use of "$$" in Makefile rules. See
https://bugs.launchpad.net/intltool/+bug/1299862.

> On NetBSD 6.1.4, OpenBSD 5.6
> build failed on due to missing header:
>
>     $ make check
>     ...
>     make  pseudotty
Fixed.

> After the 'make check' build completes, there are tests errors:
> On NetBSD 6.1.4, it seems 'dir-dangling-entry.sh' is stuck.
> The last output to the terminal is:
>    PASS: t/where-dir-file.sh
> And pressing CTRL-C shows:
>    make: *** t/dir-dangling-entry.sh.log removed

I hope it's not the case, but there's a chance there are lingering
"pseudotty" processes on your system. I suggest you try doing "ps -C
pseudotty" to make sure this is not the case. And if there are, I
apologize. I tried to kill these processes on exit but evidently some
tests aren't timing out properly.

>
> On OpenBSD 5.6, two tests fail, then 'dir-dangling-entry.sh' is stuck:

It could be because the program being tested exited too quickly. I'll
have a think about this.

>     $ cat info/t/file.sh.log
>     ./t/file.sh: ./t/Init-test.inc[61]: trap: bad signal SIGINT
>     info: "./t/Infokey-config", line 2: unknown action `xxx-not-recognized'
>     info: "./t/Infokey-config", line 3: cannot bind key sequence to
> menu-digit
>     File: file-menu,   Node: Top
>     File: file-menu,   Node: Node 1
>     File: file-menu,   Node: Node 2
>     File: file-menu,   Node: Node 3
>     File: file-menu,   Node: Has.dot
>     ./t/file.sh[33]: cleanup: not found
>     FAIL t/file.sh (exit status: 127)
>
>     $ cat info/t/file-nodes.sh.log
>     ./t/file-nodes.sh: ./t/Init-test.inc[61]: trap: bad signal SIGINT
>     info: "./t/Infokey-config", line 2: unknown action `xxx-not-recognized'
>     info: "./t/Infokey-config", line 3: cannot bind key sequence to
> menu-digit
>     Arrived at Node 1.
>     Arrived at Node 2.
>     ./t/file-nodes.sh[28]: cleanup: not found
>     FAIL t/file-nodes.sh (exit status: 127)

I have no clue. I don't know why "cleanup" wouldn't be found... unless
trap handlers are inherited by subshells maybe?

Does running "t/file.sh" from the command-line, from within the "info"
directory, work?