t5801-remote-helpers.sh fails

Hej,
on peff/pu t5801 fails, the error is in git-remote-testgit, please see below.

That's on my Mac OS X box.

I haven't digged further into the test case, but it looks as if 
"[-+]A 	make NAMEs associative arrays"
is not supported on this version of bash.
/Torsten


/Users/tb/projects/git/git.peff/git-remote-testgit: line 64: declare: -A: invalid option
declare: usage: declare [-afFirtx] [-p] [name[=value] ...]
/Users/tb/projects/git/git.peff/git-remote-testgit: line 66: refs/heads/master: division by 0 (error token is "/master")
error: fast-export died of signal 13
fatal: Error while running fast-export

Hi,

On Sat, Nov 10, 2012 at 2:48 PM, Torsten Bögershausen [off-list ref] wrote:

on peff/pu t5801 fails, the error is in git-remote-testgit, please see below.

That's on my Mac OS X box.

I haven't digged further into the test case, but it looks as if
"[-+]A  make NAMEs associative arrays"
is not supported on this version of bash.
/Torsten


/Users/tb/projects/git/git.peff/git-remote-testgit: line 64: declare: -A: invalid option
declare: usage: declare [-afFirtx] [-p] [name[=value] ...]
/Users/tb/projects/git/git.peff/git-remote-testgit: line 66: refs/heads/master: division by 0 (error token is "/master")
error: fast-export died of signal 13
fatal: Error while running fast-export

What is your bash --version?

-- 
Felipe Contreras

On 11/10/2012 08:15 PM, Felipe Contreras wrote:

Hi,

On Sat, Nov 10, 2012 at 2:48 PM, Torsten Bögershausen [off-list ref] wrote:

quoted

on peff/pu t5801 fails, the error is in git-remote-testgit, please see below.

That's on my Mac OS X box.

I haven't digged further into the test case, but it looks as if
"[-+]A  make NAMEs associative arrays"
is not supported on this version of bash.
/Torsten


/Users/tb/projects/git/git.peff/git-remote-testgit: line 64: declare: -A: invalid option
declare: usage: declare [-afFirtx] [-p] [name[=value] ...]
/Users/tb/projects/git/git.peff/git-remote-testgit: line 66: refs/heads/master: division by 0 (error token is "/master")
error: fast-export died of signal 13
fatal: Error while running fast-export

What is your bash --version?

  bash --version
GNU bash, version 3.2.48(1)-release (x86_64-apple-darwin10.0)
Copyright (C) 2007 Free Software Foundation, Inc.

On the other hand, Documentation/CodingGuidelines says:
   - No shell arrays.

Could we use perl to have arrays?

/Torsten

On Sat, Nov 10, 2012 at 8:20 PM, Torsten Bögershausen [off-list ref] wrote:

On 11/10/2012 08:15 PM, Felipe Contreras wrote:

quoted

Hi,

On Sat, Nov 10, 2012 at 2:48 PM, Torsten Bögershausen [off-list ref]
wrote:

quoted

on peff/pu t5801 fails, the error is in git-remote-testgit, please see
below.

That's on my Mac OS X box.

I haven't digged further into the test case, but it looks as if
"[-+]A  make NAMEs associative arrays"
is not supported on this version of bash.
/Torsten


/Users/tb/projects/git/git.peff/git-remote-testgit: line 64: declare: -A:
invalid option
declare: usage: declare [-afFirtx] [-p] [name[=value] ...]
/Users/tb/projects/git/git.peff/git-remote-testgit: line 66:
refs/heads/master: division by 0 (error token is "/master")
error: fast-export died of signal 13
fatal: Error while running fast-export

What is your bash --version?

 bash --version
GNU bash, version 3.2.48(1)-release (x86_64-apple-darwin10.0)
Copyright (C) 2007 Free Software Foundation, Inc.

I see, that version indeed doesn't have associative arrays.

On the other hand, Documentation/CodingGuidelines says:
  - No shell arrays.

Yeah, for shell code I guess, but this is bash code.

Could we use perl to have arrays?

I think the code in perl would be harder to follow, and this is meant
not only as a test, but also as a reference.

I'm not exactly sure what we should do here:

a) remove the code (would not be so good as a reference)
b) enable the code conditionally based on the version of bash (harder to read)
c) replace the code without associative arrays (will be much more
complicated and ugly)
d) add a check for the bash version to the top of the test in t/

I'm leaning towards d), followed by b).

If only there was a clean way to do this, so c) would not be so ugly.

After investigating different optins this seems to be the best:

	join -e empty -o '0 1.2 2.2' -a 2 <(echo "$before") <(echo "$after")
| while read e a b; do
		test "$a" == "$b" && continue
		echo "changed $e"
	done

But to me seems a bit harder to grasp. Not sure.

Cheers.

-- 
Felipe Contreras

On 10.11.12 23:05, Felipe Contreras wrote:

On Sat, Nov 10, 2012 at 8:20 PM, Torsten Bögershausen [off-list ref] wrote:

quoted

On 11/10/2012 08:15 PM, Felipe Contreras wrote:

quoted

Hi,

On Sat, Nov 10, 2012 at 2:48 PM, Torsten Bögershausen [off-list ref]
wrote:

quoted

on peff/pu t5801 fails, the error is in git-remote-testgit, please see
below.

That's on my Mac OS X box.

I haven't digged further into the test case, but it looks as if
"[-+]A  make NAMEs associative arrays"
is not supported on this version of bash.
/Torsten


/Users/tb/projects/git/git.peff/git-remote-testgit: line 64: declare: -A:
invalid option
declare: usage: declare [-afFirtx] [-p] [name[=value] ...]
/Users/tb/projects/git/git.peff/git-remote-testgit: line 66:
refs/heads/master: division by 0 (error token is "/master")
error: fast-export died of signal 13
fatal: Error while running fast-export

What is your bash --version?

 bash --version
GNU bash, version 3.2.48(1)-release (x86_64-apple-darwin10.0)
Copyright (C) 2007 Free Software Foundation, Inc.

I see, that version indeed doesn't have associative arrays.

quoted

On the other hand, Documentation/CodingGuidelines says:
  - No shell arrays.

Yeah, for shell code I guess, but this is bash code.

quoted

Could we use perl to have arrays?

I think the code in perl would be harder to follow, and this is meant
not only as a test, but also as a reference.

I'm not exactly sure what we should do here:

a) remove the code (would not be so good as a reference)
b) enable the code conditionally based on the version of bash (harder to read)
c) replace the code without associative arrays (will be much more
complicated and ugly)
d) add a check for the bash version to the top of the test in t/

I'm leaning towards d), followed by b).

If only there was a clean way to do this, so c) would not be so ugly.

After investigating different optins this seems to be the best:

	join -e empty -o '0 1.2 2.2' -a 2 <(echo "$before") <(echo "$after")
| while read e a b; do
		test "$a" == "$b" && continue
		echo "changed $e"
	done

But to me seems a bit harder to grasp. Not sure.

Cheers.

Hi again,
I managed to have a working solution for
"d) add a check for the bash version to the top of the test in t/"
Please see diff below.

This unbreaks the test suite here.
Is this a good way forward?

Filipe, does the code line you mention above work for you?

If yes: I can test it here, if you send it as a patch.

/Torsten

diff --git a/t/t5801-remote-helpers.sh b/t/t5801-remote-helpers.sh
old mode 100644
new mode 100755
index 6e4e078..ea3d0f3
--- a/t/t5801-remote-helpers.sh
+++ b/t/t5801-remote-helpers.sh

@@ -13,6 +13,15 @@ compare_refs() {
        test_cmp expect actual
 }
 
+cat >"testbashArray" <<-EOF
+  declare -A assa
+EOF
+
+/bin/bash testbashArray || {
+       skip_all='t5801. /bin/bash does not handle associative arrays'
+       test_done
+}
+
 test_expect_success 'setup repository' '
        git init server &&
        (cd server &&

Hi,

On Sun, Nov 18, 2012 at 6:50 AM, Torsten Bögershausen [off-list ref] wrote:

I managed to have a working solution for
"d) add a check for the bash version to the top of the test in t/"
Please see diff below.

This unbreaks the test suite here.
Is this a good way forward?

Filipe, does the code line you mention above work for you?

If yes: I can test it here, if you send it as a patch.

It's already sent:
http://article.gmane.org/gmane.comp.version-control.git/209364

-- 
Felipe Contreras

On 18.11.12 09:23, Felipe Contreras wrote:

Hi,

On Sun, Nov 18, 2012 at 6:50 AM, Torsten Bögershausen [off-list ref] wrote:

quoted

I managed to have a working solution for
"d) add a check for the bash version to the top of the test in t/"
Please see diff below.

This unbreaks the test suite here.
Is this a good way forward?

Filipe, does the code line you mention above work for you?

If yes: I can test it here, if you send it as a patch.

It's already sent:
http://article.gmane.org/gmane.comp.version-control.git/209364

Thanks for working on it, and sorry that you need to remind me:
I had the V5 series already on my harddisk, but forgot to test it.
It applied on v.1.8.0 with the following results:

ok 1 - setup repository
ok 2 - cloning from local repo
ok 3 - create new commit on remote
ok 4 - pulling from local repo
ok 5 - pushing to local repo
ok 6 - fetch new branch
ok 7 - fetch multiple branches
ok 8 - push when remote has extra refs
ok 9 - push new branch by name
not ok 10 - push new branch with old:new refspec # TODO known breakage
ok 11 - cloning without refspec
ok 12 - pulling without refspecs
not ok 13 - pushing without refspecs # TODO known breakage
not ok 14 - pulling with straight refspec # TODO known breakage
not ok 15 - pushing with straight refspec # TODO known breakage
not ok 16 - pulling without marks # TODO known breakage
not ok 17 - pushing without marks # TODO known breakage
ok 18 - push ref with existing object
# still have 6 known breakage(s)
/Torsten