Re: 7.4.1648 test failure on OS X

[Bram Moolenaar]

Manuel Ortega wrote:

> I just tried to build 7.4.1648.  "make test" failed thus:
> 
> ===8<==
> Test results:
> 
> >From test_quickfix.vim:
> Found errors in Test_caddbuffer_to_empty():
> function RunTheTest[9]..Test_caddbuffer_to_empty line 4: command did not
> fail: cn
> TEST FAILURE
> make[2]: *** [report] Error 1
> make[1]: *** [test] Error 2
> make: *** [test] Error 2
> 
> ===8<==
> 
> The previous build I made (I think it was 7.4.1646) did not have this
> problem.

I'll remove the check that ":cn" fails.  It depends on 'errorformat'.
It's not relevant for the testing.

-- 
hundred-and-one symptoms of being an internet addict:
119. You are reading a book and look for the scroll bar to get to
 the next page.

 /// Bram Moolenaar -- b...@moolenaar.net -- http://www.Moolenaar.net   \\\
///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\  an exciting new programming language -- http://www.Zimbu.org///
 \\\help me help AIDS victims -- http://ICCF-Holland.org///

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

7.4.1648 test failure on OS X

[Manuel Ortega]

I just tried to build 7.4.1648.  "make test" failed thus:

===8<==
Test results:

>From test_quickfix.vim:
Found errors in Test_caddbuffer_to_empty():
function RunTheTest[9]..Test_caddbuffer_to_empty line 4: command did not
fail: cn
TEST FAILURE
make[2]: *** [report] Error 1
make[1]: *** [test] Error 2
make: *** [test] Error 2

===8<==

The previous build I made (I think it was 7.4.1646) did not have this
problem.

-Manny

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: 7.4.1457 test failure on OS X 10.11.3

[Bram Moolenaar]

Ozaki kiichi wrote:

> I could reproduce this error after many trials.
> But I encountered similar kind of failures a few time in also
> v7.5.1456 test_channel.
> 
> In my case, "write failed" is caused by "Broken pipe". (confirmed by
> adding log message) This means that process had exited from while
> loop* (not by 'return NULL') even though connection had not been
> established.
> 
> * https://github.com/vim/vim/blob/d42119f/src/channel.c#L594-L756
> 
> We may require the following:
> 
> diff --git a/src/channel.c b/src/channel.c
> --- a/src/channel.c
> +++ b/src/channel.c
> @@ -703,7 +703,7 @@ channel_open(
>  * actually connect.
>  * We detect an failure to connect when both read and write fds
>  * are set.  Use getsockopt() to find out what kind of failure. */
> -   if (FD_ISSET(sd, &rfds) && FD_ISSET(sd, &wfds))
> +   if (FD_ISSET(sd, &rfds) || FD_ISSET(sd, &wfds))
> {
> ret = getsockopt(sd,
> SOL_SOCKET, SO_ERROR, &so_error, &so_error_len);
> 
> That is, increased the frequency of checking socket error.

It's unclear to me what the table in the socket(7) man page means.  It
mentions using select() but then the situation that the connection can't
be made is not in the table.

Looks like changing the && into || can't make it worse, so let's see if
that helps.

-- 
If you don't get everything you want, think of
everything you didn't get and don't want.

 /// Bram Moolenaar -- b...@moolenaar.net -- http://www.Moolenaar.net   \\\
///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\  an exciting new programming language -- http://www.Zimbu.org///
 \\\help me help AIDS victims -- http://ICCF-Holland.org///

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: 7.4.1457 test failure on OS X 10.11.3

[Ozaki Kiichi]

I could reproduce this error after many trials.
But I encountered similar kind of failures a few time in also v7.5.1456 
test_channel.

In my case, "write failed" is caused by "Broken pipe". (confirmed by adding log 
message)
This means that process had exited from while loop* (not by 'return NULL') even 
though connection had not been established.

* https://github.com/vim/vim/blob/d42119f/src/channel.c#L594-L756

We may require the following:

diff --git a/src/channel.c b/src/channel.c
--- a/src/channel.c
+++ b/src/channel.c
@@ -703,7 +703,7 @@ channel_open(
 * actually connect.
 * We detect an failure to connect when both read and write fds
 * are set.  Use getsockopt() to find out what kind of failure. */
-   if (FD_ISSET(sd, &rfds) && FD_ISSET(sd, &wfds))
+   if (FD_ISSET(sd, &rfds) || FD_ISSET(sd, &wfds))
{
ret = getsockopt(sd,
SOL_SOCKET, SO_ERROR, &so_error, &so_error_len);

That is, increased the frequency of checking socket error.


Thank you.
- Ozaki Kiichi

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: 7.4.1457 test failure on OS X 10.11.3

[Bram Moolenaar]

Kazunobu Kuriyama wrote:

> Unfortunately, I got the same experience that Manuel Ortega went through.
> 
> Yesterday, I ran "make test" several times after pulling the patch into my
> repo, and didn't see any failure with it.
> 
> But today, I got it (FWIW, after pulling other new patches into the repo
> and building vim with them).
> 
> The error message I got was the same as Manuel's:
> 
> >From test_channel.vim:
> Found errors in Test_open_delay():
> function Test_open_delay[3]..3_run_server line 46: 'Caught exception:
> Vim(call):E631: ch_evalexpr(): write failed'
> TEST FAILURE
> 
> As Bram suggested, I edited test_channel.vim to have the log file.  But
> then the error was gone as Manuel described.

Darn.  Must be a timing problem.

> The bug is really nasty.  It usually behaves as if it were normal but
> suddenly hits unlucky guys.  If someone tries to inspect it, it goes out of
> sight.
> 
> I feel the patch should be revised or reverted.

The main difference after the patch is that "len" isn't checked to be
zero.  Can you put that back and see if it helps?

-- 
SECOND SOLDIER: It could be carried by an African swallow!
FIRST SOLDIER:  Oh  yes! An African swallow maybe ... but not a European
swallow. that's my point.
 "Monty Python and the Holy Grail" PYTHON (MONTY) PICTURES LTD

 /// Bram Moolenaar -- b...@moolenaar.net -- http://www.Moolenaar.net   \\\
///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\  an exciting new programming language -- http://www.Zimbu.org///
 \\\help me help AIDS victims -- http://ICCF-Holland.org///

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: 7.4.1457 test failure on OS X 10.11.3

[Kazunobu Kuriyama]

Unfortunately, I got the same experience that Manuel Ortega went through.

Yesterday, I ran "make test" several times after pulling the patch into my
repo, and didn't see any failure with it.

But today, I got it (FWIW, after pulling other new patches into the repo
and building vim with them).

The error message I got was the same as Manuel's:

>From test_channel.vim:
Found errors in Test_open_delay():
function Test_open_delay[3]..3_run_server line 46: 'Caught exception:
Vim(call):E631: ch_evalexpr(): write failed'
TEST FAILURE

As Bram suggested, I edited test_channel.vim to have the log file.  But
then the error was gone as Manuel described.

The bug is really nasty.  It usually behaves as if it were normal but
suddenly hits unlucky guys.  If someone tries to inspect it, it goes out of
sight.

I feel the patch should be revised or reverted.

Best regards,
Kazunobu Kuriyama



2016-03-01 5:06 GMT+09:00 Bram Moolenaar :

>
> Manuel Ortega wrote:
>
> > > > I just build 7.4.1457 on OS X 10.11.3, and `make test` failed with
> this
> > > > error:
> > > >
> > > > From test_channel.vim:
> > > > Found errors in Test_open_delay():
> > > > function Test_open_delay[3]..3_run_server line 46: 'Caught
> > > exception:
> > > > Vim(call):E631: ch_evalexpr(): write failed'
> > > > TEST FAILURE
> > > > make[2]: *** [report] Error 1
> > > > make[1]: *** [test] Error 2
> > > > make: *** [test] Error 2
> > > >
> > > > Let me know if I need to provide anything else.
> > >
> > > That is unfortunate.  Did it still pass with 7.4.1456?  Then it must be
> > > the change to channel_open() that causes this.
> > >
> >
> > It did pass with 7.4.1456.
> >
> > It might be useful to have the log.  Edit src/testdir/test_channel.vim
> > > and uncomment the ch_logfile() line (near the end).  Then run the test
> > > and see if you can relate the log with that test failure.
> >
> >
> > I tried doing this, but then I couldn't reproduce the test failure.  This
> > is baffling, since I made sure I could get a test failure more than once
> > (with `make distclean` in between) before I bothered to report it here.
>
> Strange.
>
> > Side note: After doing the suggested uncommenting, I found that running
> > `make distclean` did NOT remove src/testdir/channellog.  That should
> > probably be fixed.  Perhaps there are other test-related things that
> `make
> > distclean` is not deleting, and which cause problems.
>
> "make distclean" removes files that are normally created during
> building.  It doesn't delete core dumps, valgrind logs, etc.  This log
> file will not normally be produced, thus it's good that 'git status'
> shows it, you know it's to be deleted.
>
> > I'll keep trying to reproduce the test failure, and if I can, I'll
> > resurrect this thread.
>
> I had a try myself on a Macbook: tests passed.
>
> --
> A M00se once bit my sister ...
>  "Monty Python and the Holy Grail" PYTHON (MONTY) PICTURES
> LTD
>
>  /// Bram Moolenaar -- b...@moolenaar.net -- http://www.Moolenaar.net
>  \\\
> ///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/
> \\\
> \\\  an exciting new programming language -- http://www.Zimbu.org
> ///
>  \\\help me help AIDS victims -- http://ICCF-Holland.org
> ///
>
> --
> --
> You received this message from the "vim_dev" maillist.
> Do not top-post! Type your reply below the text you are replying to.
> For more information, visit http://www.vim.org/maillist.php
>
> ---
> You received this message because you are subscribed to the Google Groups
> "vim_dev" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to vim_dev+unsubscr...@googlegroups.com.
> For more options, visit https://groups.google.com/d/optout.
>

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: 7.4.1457 test failure on OS X 10.11.3

[Bram Moolenaar]

Manuel Ortega wrote:

> > > I just build 7.4.1457 on OS X 10.11.3, and `make test` failed with this
> > > error:
> > >
> > > From test_channel.vim:
> > > Found errors in Test_open_delay():
> > > function Test_open_delay[3]..3_run_server line 46: 'Caught
> > exception:
> > > Vim(call):E631: ch_evalexpr(): write failed'
> > > TEST FAILURE
> > > make[2]: *** [report] Error 1
> > > make[1]: *** [test] Error 2
> > > make: *** [test] Error 2
> > >
> > > Let me know if I need to provide anything else.
> >
> > That is unfortunate.  Did it still pass with 7.4.1456?  Then it must be
> > the change to channel_open() that causes this.
> >
> 
> It did pass with 7.4.1456.
> 
> It might be useful to have the log.  Edit src/testdir/test_channel.vim
> > and uncomment the ch_logfile() line (near the end).  Then run the test
> > and see if you can relate the log with that test failure.
> 
> 
> I tried doing this, but then I couldn't reproduce the test failure.  This
> is baffling, since I made sure I could get a test failure more than once
> (with `make distclean` in between) before I bothered to report it here.

Strange.

> Side note: After doing the suggested uncommenting, I found that running
> `make distclean` did NOT remove src/testdir/channellog.  That should
> probably be fixed.  Perhaps there are other test-related things that `make
> distclean` is not deleting, and which cause problems.

"make distclean" removes files that are normally created during
building.  It doesn't delete core dumps, valgrind logs, etc.  This log
file will not normally be produced, thus it's good that 'git status'
shows it, you know it's to be deleted.

> I'll keep trying to reproduce the test failure, and if I can, I'll
> resurrect this thread.

I had a try myself on a Macbook: tests passed.

-- 
A M00se once bit my sister ...
 "Monty Python and the Holy Grail" PYTHON (MONTY) PICTURES LTD

 /// Bram Moolenaar -- b...@moolenaar.net -- http://www.Moolenaar.net   \\\
///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\  an exciting new programming language -- http://www.Zimbu.org///
 \\\help me help AIDS victims -- http://ICCF-Holland.org///

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: 7.4.1457 test failure on OS X 10.11.3

[Ozaki Kiichi]

Hmm, on my 10.11.3 environments, I cannot reproduce this failure...

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: 7.4.1457 test failure on OS X 10.11.3

[Manuel Ortega]

On Sun, Feb 28, 2016 at 4:18 PM, Bram Moolenaar  wrote:

>
> Manuel Ortega wrote:
>
> > I just build 7.4.1457 on OS X 10.11.3, and `make test` failed with this
> > error:
> >
> > From test_channel.vim:
> > Found errors in Test_open_delay():
> > function Test_open_delay[3]..3_run_server line 46: 'Caught
> exception:
> > Vim(call):E631: ch_evalexpr(): write failed'
> > TEST FAILURE
> > make[2]: *** [report] Error 1
> > make[1]: *** [test] Error 2
> > make: *** [test] Error 2
> >
> > Let me know if I need to provide anything else.
>
> That is unfortunate.  Did it still pass with 7.4.1456?  Then it must be
> the change to channel_open() that causes this.
>

It did pass with 7.4.1456.

It might be useful to have the log.  Edit src/testdir/test_channel.vim
> and uncomment the ch_logfile() line (near the end).  Then run the test
> and see if you can relate the log with that test failure.


I tried doing this, but then I couldn't reproduce the test failure.  This
is baffling, since I made sure I could get a test failure more than once
(with `make distclean` in between) before I bothered to report it here.

Side note: After doing the suggested uncommenting, I found that running
`make distclean` did NOT remove src/testdir/channellog.  That should
probably be fixed.  Perhaps there are other test-related things that `make
distclean` is not deleting, and which cause problems.

I'll keep trying to reproduce the test failure, and if I can, I'll
resurrect this thread.

-Manny

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: 7.4.1457 test failure on OS X 10.11.3

[Bram Moolenaar]

Manuel Ortega wrote:

> I just build 7.4.1457 on OS X 10.11.3, and `make test` failed with this
> error:
> 
> From test_channel.vim:
> Found errors in Test_open_delay():
> function Test_open_delay[3]..3_run_server line 46: 'Caught exception:
> Vim(call):E631: ch_evalexpr(): write failed'
> TEST FAILURE
> make[2]: *** [report] Error 1
> make[1]: *** [test] Error 2
> make: *** [test] Error 2
> 
> Let me know if I need to provide anything else.

That is unfortunate.  Did it still pass with 7.4.1456?  Then it must be
the change to channel_open() that causes this.

It might be useful to have the log.  Edit src/testdir/test_channel.vim
and uncomment the ch_logfile() line (near the end).  Then run the test
and see if you can relate the log with that test failure.

-- 
A fool must search for a greater fool to find admiration.

 /// Bram Moolenaar -- b...@moolenaar.net -- http://www.Moolenaar.net   \\\
///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\  an exciting new programming language -- http://www.Zimbu.org///
 \\\help me help AIDS victims -- http://ICCF-Holland.org///

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

7.4.1457 test failure on OS X 10.11.3

[Manuel Ortega]

I just build 7.4.1457 on OS X 10.11.3, and `make test` failed with this
error:

>From test_channel.vim:
Found errors in Test_open_delay():
function Test_open_delay[3]..3_run_server line 46: 'Caught exception:
Vim(call):E631: ch_evalexpr(): write failed'
TEST FAILURE
make[2]: *** [report] Error 1
make[1]: *** [test] Error 2
make: *** [test] Error 2

Let me know if I need to provide anything else.

-Manny

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: Test failure on OS X

[Manuel Ortega]

The test failures have disappeared after I build 7.4.1290.

-Manny

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: Test failure on OS X

[Bram Moolenaar]

Kazunobu Kuriyama wrote:

> As shown above, on OS X, we got the error messages "Connection refused,"
> which correspond to ECONNREFUSED.
> 
> But I found the actual errno is EINPROGRESS; the errno value is overwritten
> at line 457.

Let's remove that.

> Actually, if line 457 is commented out, I got other error messages
> "Operation now in progress," which correspond to EINPROGRESS.
> 
> This indicates that, although we got the error messages from the block
> 456--462, the real cause of the error came from connect(), as select()
> won't set errno to EINPROGRESS.

So this means select() didn't wait as expected.  I suspect this is
because it also has the read status checked.  In the docs and examples I
find one should only check the write status.

> Reading the code at line 420 and below, we see that  the code doesn't
> consider it an error when the errno is set to EINPROGRESS even if connect()
> returns -1.
> 
> If EINPROGRESS is given a go like this, then wouldn't we have to wait for
> the socket to be connected like this? (for simplicity, timeout to avoid
> infinite loop is omitted):

No, connect() doesn't take a timeout, polling is not a good way to do
this.

I also notice the code uses select() even when connect() returned zero,
that is not right.

Also, when timing out there should be no error message, the caller
should handle this.

[...]
> But I saw this sort of error appeared in previous patches and soon
> disappeared at some point recently.  Isn't it possible to fix that in a
> similar way, though I'm not pretty sure how the previous issue was fixed.
> 
> I know little about the recent development of channel.  So if I miss the
> point and am wrong, sorry about disturbing you.

I'll change a few things, please try again after that patch.
It may require some debugging and trying to make this code portable.
It's the part that is new, it didn't exist in the netbeans code.

-- 
If Microsoft would build a car...
... the oil, water temperature, and alternator warning lights would
all be replaced by a single "General Protection Fault" warning light.

 /// Bram Moolenaar -- b...@moolenaar.net -- http://www.Moolenaar.net   \\\
///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\  an exciting new programming language -- http://www.Zimbu.org///
 \\\help me help AIDS victims -- http://ICCF-Holland.org///

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: Test failure on OS X

[Kazunobu Kuriyama]

As shown above, on OS X, we got the error messages "Connection refused,"
which correspond to ECONNREFUSED.

But I found the actual errno is EINPROGRESS; the errno value is overwritten
at line 457.

Actually, if line 457 is commented out, I got other error messages
"Operation now in progress," which correspond to EINPROGRESS.

This indicates that, although we got the error messages from the block
456--462, the real cause of the error came from connect(), as select()
won't set errno to EINPROGRESS.

Reading the code at line 420 and below, we see that  the code doesn't
consider it an error when the errno is set to EINPROGRESS even if connect()
returns -1.

If EINPROGRESS is given a go like this, then wouldn't we have to wait for
the socket to be connected like this? (for simplicity, timeout to avoid
infinite loop is omitted):

diff --git a/src/channel.c b/src/channel.c
index d621798..7db37ce 100644
--- a/src/channel.c
+++ b/src/channel.c
@@ -421,7 +421,20 @@ channel_open(char *hostname, int port_in, int
waittime, void (*close_cb)(void))
 SOCK_ERRNO;
 if (ret < 0)
 {
+#ifdef __APPLE__
+ if (errno == EINPROGRESS)
+do
+{
+usleep(1000);
+errno = 0;
+ret = connect(sd, (struct sockaddr *)&server,
sizeof(server));
+}
+while (ret < 0 && errno == EALREADY);
+
+ if (errno != EWOULDBLOCK && errno != EINPROGRESS && errno != EISCONN)
+#else
  if (errno != EWOULDBLOCK && errno != EINPROGRESS)
+#endif
  {
 CHERROR("channel_open: Connect failed with errno %d\n", errno);
 CHERROR("Cannot connect to port\n", "");


Should it be a fix, the story couldn't put an end.  We'd get another error:


>From test_channel.vim:
Found errors in Test_communicate():
function Test_communicate line 51: Expected 'ok' but got ''
cat testdir/messages


>From test_channel.vim:
Executing Test_communicate()
Executing Test_server_crash()
Executing Test_two_channels()
Executed 3 tests
1 FAILED


But I saw this sort of error appeared in previous patches and soon
disappeared at some point recently.  Isn't it possible to fix that in a
similar way, though I'm not pretty sure how the previous issue was fixed.

I know little about the recent development of channel.  So if I miss the
point and am wrong, sorry about disturbing you.

Best regards,
Kazunobu Kuriyama





2016-02-06 23:28 GMT+09:00 Bram Moolenaar :

>
> Manuel Ortega wrote:
>
> > On Fri, Feb 5, 2016 at 4:51 PM, Manuel Ortega 
> wrote:
> >
> > > Vim 7.4.1262 and 7.4.1263 on Mac OS X 10.11.3 both fail the `make test`
> > > phase.  I'm pasting the error message below.
> >
> > The failure is still there on 7.4.1264, but the output is different:
> I'll
> > paste it below:
> >
> > >From test_channel.vim:
> > Executing Test_communicate()
> > Executing Test_server_crash()
> > Executing Test_two_channels()
> > Executed 3 tests
> > 6 FAILED
> >
> > >From test_channel.vim:
> > Found errors in Test_communicate():
> > Caught exception in Test_communicate(): Vim(let):E902: Cannot connect to
> > port: Connection refused @ function
> > Test_communicate[1]..3_start_server, line 34
> > Found errors in Test_server_crash():
> > Caught exception in Test_server_crash(): Vim(let):E902: Cannot connect to
> > port: Connection refused @ function
> > Test_server_crash[1]..3_start_server, line 34
> > Found errors in Test_two_channels():
> > Caught exception in Test_two_channels(): Vim(let):E902: Cannot connect to
> > port: Connection refused @ function
> > Test_two_channels[1]..3_start_server, line 34
> >
> > Test results:
> >
> >
> > >From test_channel.vim:
> > Found errors in Test_communicate():
> > Caught exception in Test_communicate(): Vim(let):E902: Cannot connect to
> > port: Connection refused @ function
> > Test_communicate[1]..3_start_server, line 34
> > Found errors in Test_server_crash():
> > Caught exception in Test_server_crash(): Vim(let):E902: Cannot connect to
> > port: Connection refused @ function
> > Test_server_crash[1]..3_start_server, line 34
> > Found errors in Test_two_channels():
> > Caught exception in Test_two_channels(): Vim(let):E902: Cannot connect to
> > port: Connection refused @ function
> > Test_two_channels[1]..3_start_server, line 34
> > TEST FAILURE
> > make[2]: *** [report] Error 1
> > make[1]: *** [test] Error 2
> > make: *** [test] Error 2
>
> This is most likely due to the timeout on connect().  That's tricky new
> code that might not work the same way on different systems.  I had
> already added ignoring an error that seems Linux specific.
>
> I hope someone can debug this and make a fix.
>
> --
> Although the scythe isn't pre-eminent among the weapons of war, anyone who
> has been on the wrong end of, say, a peasants' revolt will know that in
> skilled hands it is fearsome.
> -- (Terry Pratchett, Mort)
>
>  /// Bram Moolenaar -- b...@moolenaar.net -- http://www.Moolenaar.net
>  \\\
> ///sponsor Vim, vote for features --

Re: Test failure on OS X

[Bram Moolenaar]

Manuel Ortega wrote:

> On Fri, Feb 5, 2016 at 4:51 PM, Manuel Ortega  wrote:
> 
> > Vim 7.4.1262 and 7.4.1263 on Mac OS X 10.11.3 both fail the `make test`
> > phase.  I'm pasting the error message below.
> 
> The failure is still there on 7.4.1264, but the output is different:  I'll
> paste it below:
> 
> >From test_channel.vim:
> Executing Test_communicate()
> Executing Test_server_crash()
> Executing Test_two_channels()
> Executed 3 tests
> 6 FAILED
> 
> >From test_channel.vim:
> Found errors in Test_communicate():
> Caught exception in Test_communicate(): Vim(let):E902: Cannot connect to
> port: Connection refused @ function
> Test_communicate[1]..3_start_server, line 34
> Found errors in Test_server_crash():
> Caught exception in Test_server_crash(): Vim(let):E902: Cannot connect to
> port: Connection refused @ function
> Test_server_crash[1]..3_start_server, line 34
> Found errors in Test_two_channels():
> Caught exception in Test_two_channels(): Vim(let):E902: Cannot connect to
> port: Connection refused @ function
> Test_two_channels[1]..3_start_server, line 34
> 
> Test results:
> 
> 
> >From test_channel.vim:
> Found errors in Test_communicate():
> Caught exception in Test_communicate(): Vim(let):E902: Cannot connect to
> port: Connection refused @ function
> Test_communicate[1]..3_start_server, line 34
> Found errors in Test_server_crash():
> Caught exception in Test_server_crash(): Vim(let):E902: Cannot connect to
> port: Connection refused @ function
> Test_server_crash[1]..3_start_server, line 34
> Found errors in Test_two_channels():
> Caught exception in Test_two_channels(): Vim(let):E902: Cannot connect to
> port: Connection refused @ function
> Test_two_channels[1]..3_start_server, line 34
> TEST FAILURE
> make[2]: *** [report] Error 1
> make[1]: *** [test] Error 2
> make: *** [test] Error 2

This is most likely due to the timeout on connect().  That's tricky new
code that might not work the same way on different systems.  I had
already added ignoring an error that seems Linux specific.

I hope someone can debug this and make a fix.

-- 
Although the scythe isn't pre-eminent among the weapons of war, anyone who
has been on the wrong end of, say, a peasants' revolt will know that in
skilled hands it is fearsome.
-- (Terry Pratchett, Mort)

 /// Bram Moolenaar -- b...@moolenaar.net -- http://www.Moolenaar.net   \\\
///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\  an exciting new programming language -- http://www.Zimbu.org///
 \\\help me help AIDS victims -- http://ICCF-Holland.org///

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: Test failure on OS X

[Manuel Ortega]

On Fri, Feb 5, 2016 at 4:51 PM, Manuel Ortega  wrote:

> Vim 7.4.1262 and 7.4.1263 on Mac OS X 10.11.3 both fail the `make test`
> phase.  I'm pasting the error message below.
>

The failure is still there on 7.4.1264, but the output is different:  I'll
paste it below:

>From test_channel.vim:
Executing Test_communicate()
Executing Test_server_crash()
Executing Test_two_channels()
Executed 3 tests
6 FAILED

>From test_channel.vim:
Found errors in Test_communicate():
Caught exception in Test_communicate(): Vim(let):E902: Cannot connect to
port: Connection refused @ function
Test_communicate[1]..3_start_server, line 34
Found errors in Test_server_crash():
Caught exception in Test_server_crash(): Vim(let):E902: Cannot connect to
port: Connection refused @ function
Test_server_crash[1]..3_start_server, line 34
Found errors in Test_two_channels():
Caught exception in Test_two_channels(): Vim(let):E902: Cannot connect to
port: Connection refused @ function
Test_two_channels[1]..3_start_server, line 34

Test results:


>From test_channel.vim:
Found errors in Test_communicate():
Caught exception in Test_communicate(): Vim(let):E902: Cannot connect to
port: Connection refused @ function
Test_communicate[1]..3_start_server, line 34
Found errors in Test_server_crash():
Caught exception in Test_server_crash(): Vim(let):E902: Cannot connect to
port: Connection refused @ function
Test_server_crash[1]..3_start_server, line 34
Found errors in Test_two_channels():
Caught exception in Test_two_channels(): Vim(let):E902: Cannot connect to
port: Connection refused @ function
Test_two_channels[1]..3_start_server, line 34
TEST FAILURE
make[2]: *** [report] Error 1
make[1]: *** [test] Error 2
make: *** [test] Error 2

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Test failure on OS X

[Manuel Ortega]

Vim 7.4.1262 and 7.4.1263 on Mac OS X 10.11.3 both fail the `make test`
phase.  I'm pasting the error message below.

-Manny

>From test_channel.vim:
Executing Test_communicate()
Executing Test_server_crash()
Executing Test_two_channels()
Executed 3 tests
1 FAILED

>From test_channel.vim:
Found errors in Test_communicate():
function Test_communicate line 12: Expected 'ok' but got ''

Test results:


>From test_channel.vim:
Found errors in Test_communicate():
function Test_communicate line 12: Expected 'ok' but got ''
TEST FAILURE
make[2]: *** [report] Error 1
make[1]: *** [test] Error 2
make: *** [test] Error 2

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: Test failure on OS X

[Bram Moolenaar]

Jun Takimoto wrote:

> On 2015/01/29, at 6:39, Bram Moolenaar  wrote:
> > It's better to not skip the test.  Can you try if it works by using
> > "unnamed" instead of "unnamedplus" and then using the * register instead
> > of +?  The * register should work everywhere.
> 
> Yes.
> The test succeeds (and is not skipped) on my Mac with the patch below.
> I'm not sure whether this is also OK on Linux/Windows etc.

I'll try it out (unless someone else has already).

> Another possibility is not to rely on the &clipboard option but to
> explicitly specify the register to use; i.e., replace
>   5y
> by
>   5y +
> (in this case save/set/restore &clipboard is not necessary.)
> I feel this is simpler, unless (one of) the intention of the test is
> to confirm that the &clipboard option is working as expected.

The goal of this test was not to test the clipboard itself but how it's
handled with a nested windo + tabdo.
 
> BTW,
> on Mac (and may be on other OS?) the clipboard can contain not only a
> simple text but also various types of data, such as graphics or movies.
> If I copy a graphics to the clipboard and run 'make check' then the
> graphics is lost. Does a similar problem happen on Linux/Windows?
> On Linux, is there any difference between "* and "+ ?

Yeah, if there is something else than text it can't be restored.

> I personally feel this is not a serious problem, because if a
> user copy a graphics to the clipboard, she/he will usually (or
> almost always) paste it immediately before running 'make check'.

Well, it can be annoying.  But there are other places where the
clipboard is used, e.g. when making a Visual selection.


-- 
hundred-and-one symptoms of being an internet addict:
140. You'd rather catch a score on the web than watch the game as
 it is being played on tv.

 /// Bram Moolenaar -- b...@moolenaar.net -- http://www.Moolenaar.net   \\\
///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\  an exciting new programming language -- http://www.Zimbu.org///
 \\\help me help AIDS victims -- http://ICCF-Holland.org///

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: Test failure on OS X

[Jun T.]

On 2015/01/29, at 6:39, Bram Moolenaar  wrote:
> It's better to not skip the test.  Can you try if it works by using
> "unnamed" instead of "unnamedplus" and then using the * register instead
> of +?  The * register should work everywhere.

Yes.
The test succeeds (and is not skipped) on my Mac with the patch below.
I'm not sure whether this is also OK on Linux/Windows etc.

Another possibility is not to rely on the &clipboard option but to
explicitly specify the register to use; i.e., replace
5y
by
5y +
(in this case save/set/restore &clipboard is not necessary.)
I feel this is simpler, unless (one of) the intention of the test is
to confirm that the &clipboard option is working as expected.

BTW,
on Mac (and may be on other OS?) the clipboard can contain not only a
simple text but also various types of data, such as graphics or movies.
If I copy a graphics to the clipboard and run 'make check' then the
graphics is lost. Does a similar problem happen on Linux/Windows?
On Linux, is there any difference between "* and "+ ?

I personally feel this is not a serious problem, because if a
user copy a graphics to the clipboard, she/he will usually (or
almost always) paste it immediately before running 'make check'.



diff -r 84171683fd66 src/testdir/test_eval.in
--- a/src/testdir/test_eval.in  Tue Jan 27 22:52:15 2015 +0100
+++ b/src/testdir/test_eval.in  Thu Jan 29 09:57:38 2015 +0900
@@ -131,19 +131,19 @@
 if has('clipboard')
 " Save and restore system clipboard.
 " If no connection to X-Server is possible, test should succeed.
-let _clipreg = ['+', getreg('+'), getregtype('+')]
+let _clipreg = ['*', getreg('*'), getregtype('*')]
 let _clipopt = &cb
-let &cb='unnamedplus'
+let &cb='unnamed'
 5y
-AR +
+AR *
 tabdo :windo :echo "hi"
 6y
-AR +
+AR *
 let &cb=_clipopt
 call call('setreg', _clipreg)
 else
-  call AppendRegParts('+', 'V', "clipboard contents\n", "['clipboard 
contents']", "clipboard contents\n", "['clipboard contents']")
-  call AppendRegParts('+', 'V', "something else\n", "['something else']", 
"something else\n", "['something else']")
+  call AppendRegParts('*', 'V', "clipboard contents\n", "['clipboard 
contents']", "clipboard contents\n", "['clipboard contents']")
+  call AppendRegParts('*', 'V', "something else\n", "['something else']", 
"something else\n", "['something else']")
 endif
 $put ='{{{1 Errors'
 call ErrExe('call setreg()')
diff -r 84171683fd66 src/testdir/test_eval.ok
--- a/src/testdir/test_eval.ok  Tue Jan 27 22:52:15 2015 +0100
+++ b/src/testdir/test_eval.ok  Thu Jan 29 09:57:38 2015 +0900
@@ -317,8 +317,8 @@
 {{{2 setreg('=', ['"abc/ "'])
 =: type v; value: abc/  (['abc/ ']), expr: "abc/ " (['"abc/ "'])
 {{{1 System clipboard
-+: type V; value: clipboard contents  (['clipboard contents']), expr: 
clipboard contents  (['clipboard contents'])
-+: type V; value: something else  (['something else']), expr: something else  
(['something else'])
+*: type V; value: clipboard contents  (['clipboard contents']), expr: 
clipboard contents  (['clipboard contents'])
+*: type V; value: something else  (['something else']), expr: something else  
(['something else'])
 {{{1 Errors
 Executing call setreg()
 Vim(call):E119: Not enough arguments for function: setreg


-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: Test failure on OS X

[Bram Moolenaar]

Jun Takimoto wrote:

> On 2015/01/28, at 11:33, Manuel Ortega  wrote:
> > Test results:
> > test_eval FAILED
> 
> The failure is in the section '{{{1 System clipboard'.
> 
> There is only one clipboard on Mac OS X, and it seems
> both "* and "+ refer to it. For example both
>   "*yyand "+yy
> yank the current line to the clipboard. But
>   let &clipboard='unnamedplus'
> does not work on Mac; even with this option set,
>   :5y
> yanks the 5th line to the default register "".
> 
> If I replace
>   if has('clipboard') (test_eval.in, line 131)
> by
>   if has('unnamedplus')
> then the test succeeds (or is skipped) on my Mac.
> But I don't know whether this causes the test to be
> skipped even on Linux etc.

It's better to not skip the test.  Can you try if it works by using
"unnamed" instead of "unnamedplus" and then using the * register instead
of +?  The * register should work everywhere.

-- 
Amazing but true: If all the salmon caught in Canada in one year were laid
end to end across the Sahara Desert, the smell would be absolutely awful.

 /// Bram Moolenaar -- b...@moolenaar.net -- http://www.Moolenaar.net   \\\
///sponsor Vim, vote for features -- http://www.Vim.org/sponsor/ \\\
\\\  an exciting new programming language -- http://www.Zimbu.org///
 \\\help me help AIDS victims -- http://ICCF-Holland.org///

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: Test failure on OS X

[Manuel Ortega]

On Wed, Jan 28, 2015 at 3:07 AM, Kazunobu Kuriyama <
kazunobu.kuriy...@nifty.com> wrote:

> Hi,
>
> On Jan 28, 2015, at 11:33, Manuel Ortega  wrote:
>
> Sorry if this is a repeat of something already known.  On OS X 10.10.1,
> Vim 7.4.608 will have "make test" fail:
>
> Test results:
> test_eval FAILED
> TEST FAILURE
> make[2]: *** [report] Error 1
> make[1]: *** [test] Error 2
> make: *** [test] Error 2
>
>
> Do you have src/testdir/test_eval_func.vim in your source tree?
>

Yes I do.

-Manny

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: Test failure on OS X

[Jun T.]

On 2015/01/28, at 11:33, Manuel Ortega  wrote:
> Test results:
> test_eval FAILED

The failure is in the section '{{{1 System clipboard'.

There is only one clipboard on Mac OS X, and it seems
both "* and "+ refer to it. For example both
"*yyand "+yy
yank the current line to the clipboard. But
let &clipboard='unnamedplus'
does not work on Mac; even with this option set,
:5y
yanks the 5th line to the default register "".

If I replace
if has('clipboard') (test_eval.in, line 131)
by
if has('unnamedplus')
then the test succeeds (or is skipped) on my Mac.
But I don't know whether this causes the test to be
skipped even on Linux etc.

Jun

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Re: Test failure on OS X

[Kazunobu Kuriyama]

Hi,On Jan 28, 2015, at 11:33, Manuel Ortega  wrote:Sorry if this is a repeat of something already known.  On OS X 10.10.1, Vim 7.4.608 will have "make test" fail:Test results:test_eval FAILEDTEST FAILUREmake[2]: *** [report] Error 1make[1]: *** [test] Error 2make: *** [test] Error 2
Do you have src/testdir/test_eval_func.vim in your source tree?While that file has been added to mine at an update via Mercurial after Patch 7.4.608, this particular patch itself shows no indication of including that newly added file.I didn’t noticed that until I read your mail, that’s why I guess you’ve missed that file.For your convenience, test_eval_func.vim is attached to this mail.Regards,Kazunobu KuriyamaP.S. I confirm Patch 7.4.608 fixes the issues related to Patch 7.4.598.



-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups "vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.


test_eval_func.vim
Description: Binary data




-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups "vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.

Test failure on OS X

[Manuel Ortega]

Sorry if this is a repeat of something already known.  On OS X 10.10.1, Vim
7.4.608 will have "make test" fail:

Test results:
test_eval FAILED
TEST FAILURE
make[2]: *** [report] Error 1
make[1]: *** [test] Error 2
make: *** [test] Error 2

The portion of that file that seems to indicate what happened is this:

Vim(call):E119: Not enough arguments for function: setreg
Executing call setreg(1, 2, 3, 4)
Vim(call):E118: Too many arguments for function: setreg
Executing call setreg([], 2)
Vim(call):E730: using List as a String
Executing call setreg(1, {})
Vim(call):E731: using Dictionary as a String
Executing call setreg(1, 2, [])
Vim(call):E730: using List as a String
Executing call setreg("/", ["1", "2"])
Vim(call):E883: search pattern and expression register may not contain two
or more lines
Executing call setreg("=", ["1", "2"])
Vim(call):E883: search pattern and expression register may not contain two
or more lines
Executing call setreg(1, ["", "", [], ""])
Vim(call):E730: using List as a String
Vim(function):E128: Function name must start with a capital or "s:":
g:test()
Vim(function):E128: Function name must start with a capital or "s:":
b:test()
Vim(function):E128: Function name must start with a capital or "s:":
test2() "#
called Foo(1)
called Foo(2)
called Foo(3)
s:Testje exists: 0
func s:Testje exists: 1
Bar exists: 1
func Bar exists: 1
Vim(call):E116: Invalid arguments for function append
{{{1 getcurpos/setpos
6

Let me know if you need to see the whole file.

-Manny

-- 
-- 
You received this message from the "vim_dev" maillist.
Do not top-post! Type your reply below the text you are replying to.
For more information, visit http://www.vim.org/maillist.php

--- 
You received this message because you are subscribed to the Google Groups 
"vim_dev" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to vim_dev+unsubscr...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.