You are viewing a plain text version of this content. The canonical link for it is here.
Posted to dev@commons.apache.org by "Jeffrey D. Brekke" <jb...@wi.rr.com> on 2004/02/17 06:07:40 UTC
[net] Functional Tests ( was Re: [net][vote] Release plan for
commons-net.1.2.0 )
Sorry to take so long to reply to this vote. I wanted to explore the
api a bit and did so by writing a test case that we can use for
functionally testing FTPClient wrt to listings. The test is data
driven so in the suite() method there is an array of string arrays
holding information about each site to test. For each entry the test
case is run. Currently I have a public unix and windows site being
tested. I did find one public vms ftp server at HP, but there is a
problem with it. The listings aren't quite the same as I've seen
before, the files permissions/groups aren't terminated with a closing
']'. I've tried to run this test against some of our internal vms
sites, but also ran into trouble ( although I haven't had much time to
figure it out ).
Since this test shouldn't be run at each build, I have it excluded
like the telnet functional test. I also added a test:functional
goal to our maven.xml so we can run these tests.
maven test:functional
The above command will run any test case that has a name like:
*FunctionalTest.java
I'm leaning towards a -1 until we can verify the vms case. I think it
may be my test or the public site, but I'm just not sure. I know it
is an obscure case, but we do use it and I'd like to know the parsers
do work. This functional suite, although it is slow, should provide
us with some end-to-end regression along with the currenly growing set
of programmer tests.
>>>>> On Fri, 30 Jan 2004 16:21:10 -0600, Steve Cohen <sc...@javactivity.org> said:
> I am now proposing for a formal vote, releasing commons-net.1.2.0.
> The proposal is to release the current HEAD branch, as last amended
> by me over a week ago.
> I know there are other patches that have been submitted but no one
> seems to have the time to evaluate them. The current version of the
> code contributes autodetection of system type with manual override
> capability for any ListParser and accomplishes substantial
> refactoring, reducing the confusion level in this part of the
> system. These are important changes and will enable better
> interface with Ant.
> This plan does NOT encompass the release of the 1.1.1 branch as the
> last JDK1.1 compatible version. Although the code now in that
> branch compiles and runs if jakarta-oro-2.0.1.jar is substituted for
> 2.0.7, it throws JIT errors which probably have more to do with my
> execution environment or some configuration mistake in this ancient
> JDK and I'm stumped. I recommend doing this release later (CVS tag
> is there, waiting to be handled) when more time may be available to
> look at these issues. If anyone is familiar with JIT errors in
> JDK1.1 perhaps they can weigh in on what error messages in my
> previous emails mean.
> My vote, of course: +1
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: commons-dev-help@jakarta.apache.org
--
=====================================================================
Jeffrey D. Brekke jbrekke@wi.rr.com
Wisconsin, USA brekke@apache.org
ekkerbj@yahoo.com
---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org
Re: [net] Functional Tests ( was Re: [net][vote] Release plan for commons-net.1.2.0 )
Posted by Steve Cohen <sc...@javactivity.org>.
Here's a further problem Jeff's test has uncovered:
if the code is
F=FTPClient();
...
F.listFiles("subdirectory")
it works fine.
If on the other hand, we code
F.cwd("subdirectory")
F.listFiles() it returns a smaller set of files.
In particular, it returns only those files whose names (not counting the
version number) END with a period: for example CONFIGURE., IMAKEFILE., etc.
Those files with something after the period, e.g. CONFIGURE.TXT, etc. are NOT
returned.
I think the reason for this is that when no directory name is supplied in the
call to listFiles() (or initiateListParsing(), it's the same there, too) the
code currently substitutes a dot, ".") This works for unix and windows, but
it seems as though VMS interprets this not to mean the current directory, but
instead, all files in the current directory ending in ".".
You can verify this behavior by logging into this server using another ftp
program and trying the command "ls ."
Can a VMS guru on this list confirm this?
At any rate, there is an easy solution to this. Instead of making "." the
default path expression when none is supplied, I can simply supply the empty
string. This works on unix, windows and vms.
However, there were still some test failures. For one thing, the searched for
file, XTERM.EXE, does not fall in the first 25 files on the server (the
listing is alphabetical and there are 127 entries). This one I solved by
changing the test to look for a file that will be among the first returned by
the server. This is not great testing practice as we have no control over
the files that will be placed there, but I don't know what else to do.
Finally other tests were failing with null pointer exceptions. When
initiateListParsing attempts to parse a non-existent directory on the FTP
server, the routine returns was returning null, again, only for VMS.
Evidently
_openDataConnection_(FTPCommand.LIST, pathname)
returns null on the VMS FTP server when pathname is not a path it can find but
on unix and windows, it will return a valid socket.
Some simple coding changes fixed this.
So that was that. Thanks for writing that functional test, Jeff. It found a
lot of problems I hadn't suspected were there.
On Wednesday 18 February 2004 6:27 pm, Steve Cohen wrote:
> Home sick, so I have a little time.
>
> I ftp'd to the hp site that you have commented out and I can see the
> following problem right away:
>
> Here is the output from issuing the FTP command:
> Directory FREEWARE:[FREEWARE]
>
> FREEWARE40.DIR;1 27/36 16-FEB-1999 10:01:46
> [AP_HTTPD,APACHE$WWW (RWE,RWE,RE,RE)
> FREEWARE50.BCK;1 1658475/1658484 19-JUN-2003 15:22:42
> [AP_HTTPD,APACHE$WW W (RE,RE,RE,RE)
> FREEWARE50.DIR;1 13/18 30-APR-2001 10:23:09
> [AP_HTTPD,APACHE$WWW (RWE,RWE,RE,RE)
> FREEWARE50.ZIP;1 800159/800172 19-JUN-2003 16:17:44
> [AP_HTTPD,APACHE$WW W (RE,RE,RE,RE)
> FREEWARE60.DIR;1 8/18 1-DEC-2003 13:58:03
> [AP_HTTPD,APACHE$WWW (RWE,RWE,RE,E)
> FREEWARE60_1.ZIP;1 895547/895554 1-DEC-2003 12:02:18
> [AP_HTTPD,APACHE$WW W (RWED,RWED,RE,)
> FREEWARE60_2.ZIP;1 439061/439074 1-DEC-2003 12:15:51
> [AP_HTTPD,APACHE$WW W (RWED,RWED,RE,)
> TTI_INTOUCH.DIR;1 1/18 2-JAN-2002 14:00:06
> [AP_HTTPD,APACHE$WWW (RWE,RWE,RE,RE)
>
> Notice that the [USER,GROUP] section of each entry does not terminate with
> a closing square bracket. The regex expects this and all of the original
> unit tests are coded that way. If this output is correct, then our regex
> has always been wrong. It appears that the final ']' should be optional.
> I hate the FTP spec. It's so fricking vague.
>
> Still, by making the closing ']' optional, I get around this first problem.
>
> However, that is not the end of it. There are more problems. I'm
> investigating further.
>
> On Wednesday 18 February 2004 2:22 pm, Jeffrey D. Brekke wrote:
> > >>>>> On Tue, 17 Feb 2004 06:36:23 -0600, Steve Cohen
> > >>>>> <sc...@javactivity.org> said:
> > >
> > > Good, Jeff. I've long thought I was hampered by not having actual
> > > sites to test against. It's quite possible something slipped as VMS
> > > underwent heavier changes than any of the others.
> > >
> > > Would it be possible to run the functional test against an earlier
> > > version of the package? That might help to quickly verify where the
> > > problem lies.
> >
> > Maybe, skipping methods that were added. I really wanted to use the
> > new methods you added and just that exploring in the form of
> > functional testing. You maybe on to something, it may have never
> > worked.
> >
> > > On Monday 16 February 2004 11:07 pm, you wrote:
> > >> Sorry to take so long to reply to this vote. I wanted to explore
> > >> the api a bit and did so by writing a test case that we can use for
> > >> functionally testing FTPClient wrt to listings. The test is data
> > >> driven so in the suite() method there is an array of string arrays
> > >> holding information about each site to test. For each entry the
> > >> test case is run. Currently I have a public unix and windows site
> > >> being tested. I did find one public vms ftp server at HP, but
> > >> there is a problem with it. The listings aren't quite the same as
> > >> I've seen before, the files permissions/groups aren't terminated
> > >> with a closing ']'. I've tried to run this test against some of
> > >> our internal vms sites, but also ran into trouble ( although I
> > >> haven't had much time to figure it out ).
> > >>
> > >> Since this test shouldn't be run at each build, I have it excluded
> > >> like the telnet functional test. I also added a test:functional
> > >> goal to our maven.xml so we can run these tests.
> > >>
> > >> maven test:functional
> > >>
> > >> The above command will run any test case that has a name like:
> > >> *FunctionalTest.java
> > >>
> > >> I'm leaning towards a -1 until we can verify the vms case. I think
> > >> it may be my test or the public site, but I'm just not sure. I
> > >> know it is an obscure case, but we do use it and I'd like to know
> > >> the parsers do work. This functional suite, although it is slow,
> > >> should provide us with some end-to-end regression along with the
> > >> currenly growing set of programmer tests.
> > >>
> > >> >>>>> On Fri, 30 Jan 2004 16:21:10 -0600, Steve Cohen >>>>>
> > >>
> > >> <sc...@javactivity.org> said:
> > >> > I am now proposing for a formal vote, releasing
> > >>
> > >> commons-net.1.2.0. > The proposal is to release the current HEAD
> > >> branch, as last amended > by me over a week ago.
> > >>
> > >> > I know there are other patches that have been submitted but no
> > >>
> > >> one > seems to have the time to evaluate them. The current version
> > >> of the > code contributes autodetection of system type with manual
> > >> override > capability for any ListParser and accomplishes
> > >> substantial > refactoring, reducing the confusion level in this
> > >> part of the > system. These are important changes and will enable
> > >> better > interface with Ant.
> > >>
> > >> > This plan does NOT encompass the release of the 1.1.1 branch as
> > >>
> > >> the > last JDK1.1 compatible version. Although the code now in
> > >> that > branch compiles and runs if jakarta-oro-2.0.1.jar is
> > >> substituted for > 2.0.7, it throws JIT errors which probably have
> > >> more to do with my > execution environment or some configuration
> > >> mistake in this ancient > JDK and I'm stumped. I recommend doing
> > >> this release later (CVS tag > is there, waiting to be handled) when
> > >> more time may be available to > look at these issues. If anyone is
> > >> familiar with JIT errors in > JDK1.1 perhaps they can weigh in on
> > >> what error messages in my > previous emails mean.
> > >>
> > >> > My vote, of course: +1
> > >>
> > >> ---------------------------------------------------------------------
> > >>
> > >> > To unsubscribe, e-mail:
> > >>
> > >> commons-dev-unsubscribe@jakarta.apache.org > For additional
> > >> commands, e-mail: commons-dev-help@jakarta.apache.org
> > >
> > > ---------------------------------------------------------------------
> > > To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
> > > For additional commands, e-mail: commons-dev-help@jakarta.apache.org
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: commons-dev-help@jakarta.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org
Re: [net] Functional Tests ( was Re: [net][vote] Release plan for commons-net.1.2.0 )
Posted by Steve Cohen <sc...@javactivity.org>.
Home sick, so I have a little time.
I ftp'd to the hp site that you have commented out and I can see the following
problem right away:
Here is the output from issuing the FTP command:
Directory FREEWARE:[FREEWARE]
FREEWARE40.DIR;1 27/36 16-FEB-1999 10:01:46
[AP_HTTPD,APACHE$WWW (RWE,RWE,RE,RE)
FREEWARE50.BCK;1 1658475/1658484 19-JUN-2003 15:22:42
[AP_HTTPD,APACHE$WW W (RE,RE,RE,RE)
FREEWARE50.DIR;1 13/18 30-APR-2001 10:23:09
[AP_HTTPD,APACHE$WWW (RWE,RWE,RE,RE)
FREEWARE50.ZIP;1 800159/800172 19-JUN-2003 16:17:44
[AP_HTTPD,APACHE$WW W (RE,RE,RE,RE)
FREEWARE60.DIR;1 8/18 1-DEC-2003 13:58:03
[AP_HTTPD,APACHE$WWW (RWE,RWE,RE,E)
FREEWARE60_1.ZIP;1 895547/895554 1-DEC-2003 12:02:18
[AP_HTTPD,APACHE$WW W (RWED,RWED,RE,)
FREEWARE60_2.ZIP;1 439061/439074 1-DEC-2003 12:15:51
[AP_HTTPD,APACHE$WW W (RWED,RWED,RE,)
TTI_INTOUCH.DIR;1 1/18 2-JAN-2002 14:00:06
[AP_HTTPD,APACHE$WWW (RWE,RWE,RE,RE)
Notice that the [USER,GROUP] section of each entry does not terminate with a
closing square bracket. The regex expects this and all of the original unit
tests are coded that way. If this output is correct, then our regex has
always been wrong. It appears that the final ']' should be optional. I hate
the FTP spec. It's so fricking vague.
Still, by making the closing ']' optional, I get around this first problem.
However, that is not the end of it. There are more problems. I'm
investigating further.
On Wednesday 18 February 2004 2:22 pm, Jeffrey D. Brekke wrote:
> >>>>> On Tue, 17 Feb 2004 06:36:23 -0600, Steve Cohen
> >>>>> <sc...@javactivity.org> said:
> >
> > Good, Jeff. I've long thought I was hampered by not having actual
> > sites to test against. It's quite possible something slipped as VMS
> > underwent heavier changes than any of the others.
> >
> > Would it be possible to run the functional test against an earlier
> > version of the package? That might help to quickly verify where the
> > problem lies.
>
> Maybe, skipping methods that were added. I really wanted to use the
> new methods you added and just that exploring in the form of
> functional testing. You maybe on to something, it may have never
> worked.
>
> > On Monday 16 February 2004 11:07 pm, you wrote:
> >> Sorry to take so long to reply to this vote. I wanted to explore
> >> the api a bit and did so by writing a test case that we can use for
> >> functionally testing FTPClient wrt to listings. The test is data
> >> driven so in the suite() method there is an array of string arrays
> >> holding information about each site to test. For each entry the
> >> test case is run. Currently I have a public unix and windows site
> >> being tested. I did find one public vms ftp server at HP, but
> >> there is a problem with it. The listings aren't quite the same as
> >> I've seen before, the files permissions/groups aren't terminated
> >> with a closing ']'. I've tried to run this test against some of
> >> our internal vms sites, but also ran into trouble ( although I
> >> haven't had much time to figure it out ).
> >>
> >> Since this test shouldn't be run at each build, I have it excluded
> >> like the telnet functional test. I also added a test:functional
> >> goal to our maven.xml so we can run these tests.
> >>
> >> maven test:functional
> >>
> >> The above command will run any test case that has a name like:
> >> *FunctionalTest.java
> >>
> >> I'm leaning towards a -1 until we can verify the vms case. I think
> >> it may be my test or the public site, but I'm just not sure. I
> >> know it is an obscure case, but we do use it and I'd like to know
> >> the parsers do work. This functional suite, although it is slow,
> >> should provide us with some end-to-end regression along with the
> >> currenly growing set of programmer tests.
> >>
> >> >>>>> On Fri, 30 Jan 2004 16:21:10 -0600, Steve Cohen >>>>>
> >>
> >> <sc...@javactivity.org> said:
> >> > I am now proposing for a formal vote, releasing
> >>
> >> commons-net.1.2.0. > The proposal is to release the current HEAD
> >> branch, as last amended > by me over a week ago.
> >>
> >> > I know there are other patches that have been submitted but no
> >>
> >> one > seems to have the time to evaluate them. The current version
> >> of the > code contributes autodetection of system type with manual
> >> override > capability for any ListParser and accomplishes
> >> substantial > refactoring, reducing the confusion level in this
> >> part of the > system. These are important changes and will enable
> >> better > interface with Ant.
> >>
> >> > This plan does NOT encompass the release of the 1.1.1 branch as
> >>
> >> the > last JDK1.1 compatible version. Although the code now in
> >> that > branch compiles and runs if jakarta-oro-2.0.1.jar is
> >> substituted for > 2.0.7, it throws JIT errors which probably have
> >> more to do with my > execution environment or some configuration
> >> mistake in this ancient > JDK and I'm stumped. I recommend doing
> >> this release later (CVS tag > is there, waiting to be handled) when
> >> more time may be available to > look at these issues. If anyone is
> >> familiar with JIT errors in > JDK1.1 perhaps they can weigh in on
> >> what error messages in my > previous emails mean.
> >>
> >> > My vote, of course: +1
> >>
> >> ---------------------------------------------------------------------
> >>
> >> > To unsubscribe, e-mail:
> >>
> >> commons-dev-unsubscribe@jakarta.apache.org > For additional
> >> commands, e-mail: commons-dev-help@jakarta.apache.org
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
> > For additional commands, e-mail: commons-dev-help@jakarta.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org
Re: [net] Functional Tests ( was Re: [net][vote] Release plan for
commons-net.1.2.0 )
Posted by "Jeffrey D. Brekke" <jb...@wi.rr.com>.
>>>>> On Tue, 17 Feb 2004 06:36:23 -0600, Steve Cohen <sc...@javactivity.org> said:
> Good, Jeff. I've long thought I was hampered by not having actual
> sites to test against. It's quite possible something slipped as VMS
> underwent heavier changes than any of the others.
> Would it be possible to run the functional test against an earlier
> version of the package? That might help to quickly verify where the
> problem lies.
Maybe, skipping methods that were added. I really wanted to use the
new methods you added and just that exploring in the form of
functional testing. You maybe on to something, it may have never
worked.
> On Monday 16 February 2004 11:07 pm, you wrote:
>> Sorry to take so long to reply to this vote. I wanted to explore
>> the api a bit and did so by writing a test case that we can use for
>> functionally testing FTPClient wrt to listings. The test is data
>> driven so in the suite() method there is an array of string arrays
>> holding information about each site to test. For each entry the
>> test case is run. Currently I have a public unix and windows site
>> being tested. I did find one public vms ftp server at HP, but
>> there is a problem with it. The listings aren't quite the same as
>> I've seen before, the files permissions/groups aren't terminated
>> with a closing ']'. I've tried to run this test against some of
>> our internal vms sites, but also ran into trouble ( although I
>> haven't had much time to figure it out ).
>>
>> Since this test shouldn't be run at each build, I have it excluded
>> like the telnet functional test. I also added a test:functional
>> goal to our maven.xml so we can run these tests.
>>
>> maven test:functional
>>
>> The above command will run any test case that has a name like:
>> *FunctionalTest.java
>>
>> I'm leaning towards a -1 until we can verify the vms case. I think
>> it may be my test or the public site, but I'm just not sure. I
>> know it is an obscure case, but we do use it and I'd like to know
>> the parsers do work. This functional suite, although it is slow,
>> should provide us with some end-to-end regression along with the
>> currenly growing set of programmer tests.
>>
>> >>>>> On Fri, 30 Jan 2004 16:21:10 -0600, Steve Cohen >>>>>
>> <sc...@javactivity.org> said:
>> >
>> > I am now proposing for a formal vote, releasing
>> commons-net.1.2.0. > The proposal is to release the current HEAD
>> branch, as last amended > by me over a week ago.
>> >
>> > I know there are other patches that have been submitted but no
>> one > seems to have the time to evaluate them. The current version
>> of the > code contributes autodetection of system type with manual
>> override > capability for any ListParser and accomplishes
>> substantial > refactoring, reducing the confusion level in this
>> part of the > system. These are important changes and will enable
>> better > interface with Ant.
>> >
>> > This plan does NOT encompass the release of the 1.1.1 branch as
>> the > last JDK1.1 compatible version. Although the code now in
>> that > branch compiles and runs if jakarta-oro-2.0.1.jar is
>> substituted for > 2.0.7, it throws JIT errors which probably have
>> more to do with my > execution environment or some configuration
>> mistake in this ancient > JDK and I'm stumped. I recommend doing
>> this release later (CVS tag > is there, waiting to be handled) when
>> more time may be available to > look at these issues. If anyone is
>> familiar with JIT errors in > JDK1.1 perhaps they can weigh in on
>> what error messages in my > previous emails mean.
>> >
>> > My vote, of course: +1
>> >
>> >
>> >
>> ---------------------------------------------------------------------
>> > To unsubscribe, e-mail:
>> commons-dev-unsubscribe@jakarta.apache.org > For additional
>> commands, e-mail: commons-dev-help@jakarta.apache.org
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: commons-dev-help@jakarta.apache.org
--
=====================================================================
Jeffrey D. Brekke jbrekke@wi.rr.com
Wisconsin, USA brekke@apache.org
ekkerbj@yahoo.com
---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org
Re: [net] Unit and functional tests (resend)
Posted by Steve Cohen <sc...@javactivity.org>.
On Wednesday 26 May 2004 5:48 am, Mike George wrote:
> Thanks. I think my ISP strips the .java files. I attached them as
> *.java.txt this time. If they get stripped, I will add them to Bugzilla.
>
> Mike
Putting them in bugzilla is better even if your ISP doesn't strip them.
---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org
[net] Unit and functional tests (resend)
Posted by Mike George <mi...@comcast.net>.
Thanks. I think my ISP strips the .java files. I attached them as
*.java.txt this time. If they get stripped, I will add them to Bugzilla.
Mike
::>-----Original Message-----
::>From: Jeffrey D. Brekke [mailto:jbrekke@wi.rr.com]
::>Sent: Monday, April 26, 2004 1:54 AM
::>To: Jakarta Commons Developers List
::>Subject: Re: [net] Unit and functional tests
::>
::>
::>
::>Sounds great, thanks for the effort! I don't think your attachments
::>came through. Can you put them somewhere we can get at them, or add
::>them to bugzilla I guess.
::>
::>>>>>> On Sun, 25 Apr 2004 13:27:05 -0400, Mike George
::><mi...@comcast.net> said:
::>
::>> Hello, Attached please find 5 java source files designed to test the
::>> classes in the package org.apache.commons.net.pop3 package.
::>
::>> In response to a comment from Jeffrey Brekke, below, that
::>> suggested that it would be good to be able to run the tests locally,
::>> I designed these tests to run against a local, default installation
::>> of James. The file TestSetupParameters.java contains parameters
::>> that can be changed if you prefer to run the tests against another
::>> target.
::>
::>> This is my first submission to this project. I appreciate any
::>> feedback or advice you may have. I am also willing to make any
::>> changes necessary to make this code suitable for inclusion in this
::>> project.
::>
::>> Thanks,
::>
::>> Mike
::>
::>>> -----Original Message----- From: Jeffrey D. Brekke
::>>> [mailto:jbrekke@wi.rr.com] Sent: Wednesday, February 18, 2004 3:25
::>>> PM To: Jakarta Commons Developers List Subject: Re: [net] Unit and
::>>> functional tests
::>>>
::>
::>> <snip>
::>
::>>> I'd prefer that you post the code on the -dev list and not in
::>>> bugzilla. Save that for bugs. ( that is of course only *my*
::>>> opinion ).
::>
::>> <snip>
::>
::>>> Sure. Providing a way to test locally ( i.e. not requiring the
::>>> internet ) would be nice.
::>
::>> ---------------------------------------------------------------------
::>> To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
::>> For additional commands, e-mail: commons-dev-help@jakarta.apache.org
::>
::>--
::>=====================================================================
::>Jeffrey D. Brekke jbrekke@wi.rr.com
::>Wisconsin, USA brekke@apache.org
::> ekkerbj@yahoo.com
::>
::>
::>---------------------------------------------------------------------
::>To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
::>For additional commands, e-mail: commons-dev-help@jakarta.apache.org
::>
Re: [net] Unit and functional tests
Posted by "Jeffrey D. Brekke" <jb...@wi.rr.com>.
Sounds great, thanks for the effort! I don't think your attachments
came through. Can you put them somewhere we can get at them, or add
them to bugzilla I guess.
>>>>> On Sun, 25 Apr 2004 13:27:05 -0400, Mike George <mi...@comcast.net> said:
> Hello, Attached please find 5 java source files designed to test the
> classes in the package org.apache.commons.net.pop3 package.
> In response to a comment from Jeffrey Brekke, below, that
> suggested that it would be good to be able to run the tests locally,
> I designed these tests to run against a local, default installation
> of James. The file TestSetupParameters.java contains parameters
> that can be changed if you prefer to run the tests against another
> target.
> This is my first submission to this project. I appreciate any
> feedback or advice you may have. I am also willing to make any
> changes necessary to make this code suitable for inclusion in this
> project.
> Thanks,
> Mike
>> -----Original Message----- From: Jeffrey D. Brekke
>> [mailto:jbrekke@wi.rr.com] Sent: Wednesday, February 18, 2004 3:25
>> PM To: Jakarta Commons Developers List Subject: Re: [net] Unit and
>> functional tests
>>
> <snip>
>> I'd prefer that you post the code on the -dev list and not in
>> bugzilla. Save that for bugs. ( that is of course only *my*
>> opinion ).
> <snip>
>> Sure. Providing a way to test locally ( i.e. not requiring the
>> internet ) would be nice.
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
> For additional commands, e-mail: commons-dev-help@jakarta.apache.org
--
=====================================================================
Jeffrey D. Brekke jbrekke@wi.rr.com
Wisconsin, USA brekke@apache.org
ekkerbj@yahoo.com
---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org
RE: [net] Unit and functional tests
Posted by Mike George <mi...@comcast.net>.
Hello,
Attached please find 5 java source files designed to
test the classes in the package org.apache.commons.net.pop3
package.
In response to a comment from Jeffrey Brekke, below,
that suggested that it would be good to be able to run the tests
locally, I designed these tests to run against a local, default
installation of James.
The file TestSetupParameters.java contains parameters that
can be changed if you prefer to run the tests against another
target.
This is my first submission to this project. I appreciate
any feedback or advice you may have. I am also willing to make
any changes necessary to make this code suitable for inclusion in
this project.
Thanks,
Mike
>-----Original Message-----
>From: Jeffrey D. Brekke [mailto:jbrekke@wi.rr.com]
>Sent: Wednesday, February 18, 2004 3:25 PM
>To: Jakarta Commons Developers List
>Subject: Re: [net] Unit and functional tests
>
<snip>
>I'd prefer that you post the code on the -dev list and not in
>bugzilla. Save that for bugs. ( that is of course only *my* opinion
>).
<snip>
>Sure. Providing a way to test locally ( i.e. not requiring the
>internet ) would be nice.
Re: [net] Unit and functional tests
Posted by "Jeffrey D. Brekke" <jb...@wi.rr.com>.
>>>>> On Tue, 17 Feb 2004 20:33:06 -0500, Mike George <mi...@comcast.net> said:
> Hello, As part of project for a college class I'm taking, I am
> writing a series of unit tests and functional tests for two of the
> components of the Commons Net project. I intend to submit any tests
> I write to Bugzilla. Hopefully someone will have the time to review
> them and submit them if they are acceptable.
Welcome. The commons/net project could use some functional tests.
I'd prefer that you post the code on the -dev list and not in
bugzilla. Save that for bugs. ( that is of course only *my* opinion
).
> Question: 1. I would like to work on tests for the POP3 and SMTP
> components but I do not want to duplicate someone else's work. Is
> anyone working on tests for these components?
Not that I know of, have at it!
> 2. Assuming that I do work on these components, should I use RFC 821
> (SMTP) and RFC 1939 (POP3) as guidelines?
Sure. Providing a way to test locally ( i.e. not requiring the
internet ) would be nice.
--
=====================================================================
Jeffrey D. Brekke jbrekke@wi.rr.com
Wisconsin, USA brekke@apache.org
ekkerbj@yahoo.com
---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org
[net] Unit and functional tests
Posted by Mike George <mi...@comcast.net>.
Hello,
As part of project for a college class I'm taking, I am writing a series of
unit tests and functional tests for two of the components of the Commons Net
project. I intend to submit any tests I write to Bugzilla. Hopefully
someone will have the time to review them and submit them if they are
acceptable.
Question:
1. I would like to work on tests for the POP3 and SMTP components but I do
not want to duplicate someone else's work. Is anyone working on tests for
these components?
2. Assuming that I do work on these components, should I use RFC 821 (SMTP)
and RFC 1939 (POP3) as guidelines?
Thanks,
Mike
-----------------------------------------------------------------
"However, it is by our works and not our words that
we would be judged. These, we hope, will sustain us..."
-Robert B. Thomas ed. (Old Farmer's Almanac)
-----------------------------------------------------------------
---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org
Re: [net] Functional Tests ( was Re: [net][vote] Release plan for commons-net.1.2.0 )
Posted by Steve Cohen <sc...@javactivity.org>.
Good, Jeff. I've long thought I was hampered by not having actual sites to
test against. It's quite possible something slipped as VMS underwent heavier
changes than any of the others.
Would it be possible to run the functional test against an earlier version of
the package? That might help to quickly verify where the problem lies.
On Monday 16 February 2004 11:07 pm, you wrote:
> Sorry to take so long to reply to this vote. I wanted to explore the
> api a bit and did so by writing a test case that we can use for
> functionally testing FTPClient wrt to listings. The test is data
> driven so in the suite() method there is an array of string arrays
> holding information about each site to test. For each entry the test
> case is run. Currently I have a public unix and windows site being
> tested. I did find one public vms ftp server at HP, but there is a
> problem with it. The listings aren't quite the same as I've seen
> before, the files permissions/groups aren't terminated with a closing
> ']'. I've tried to run this test against some of our internal vms
> sites, but also ran into trouble ( although I haven't had much time to
> figure it out ).
>
> Since this test shouldn't be run at each build, I have it excluded
> like the telnet functional test. I also added a test:functional
> goal to our maven.xml so we can run these tests.
>
> maven test:functional
>
> The above command will run any test case that has a name like:
> *FunctionalTest.java
>
> I'm leaning towards a -1 until we can verify the vms case. I think it
> may be my test or the public site, but I'm just not sure. I know it
> is an obscure case, but we do use it and I'd like to know the parsers
> do work. This functional suite, although it is slow, should provide
> us with some end-to-end regression along with the currenly growing set
> of programmer tests.
>
> >>>>> On Fri, 30 Jan 2004 16:21:10 -0600, Steve Cohen
> >>>>> <sc...@javactivity.org> said:
> >
> > I am now proposing for a formal vote, releasing commons-net.1.2.0.
> > The proposal is to release the current HEAD branch, as last amended
> > by me over a week ago.
> >
> > I know there are other patches that have been submitted but no one
> > seems to have the time to evaluate them. The current version of the
> > code contributes autodetection of system type with manual override
> > capability for any ListParser and accomplishes substantial
> > refactoring, reducing the confusion level in this part of the
> > system. These are important changes and will enable better
> > interface with Ant.
> >
> > This plan does NOT encompass the release of the 1.1.1 branch as the
> > last JDK1.1 compatible version. Although the code now in that
> > branch compiles and runs if jakarta-oro-2.0.1.jar is substituted for
> > 2.0.7, it throws JIT errors which probably have more to do with my
> > execution environment or some configuration mistake in this ancient
> > JDK and I'm stumped. I recommend doing this release later (CVS tag
> > is there, waiting to be handled) when more time may be available to
> > look at these issues. If anyone is familiar with JIT errors in
> > JDK1.1 perhaps they can weigh in on what error messages in my
> > previous emails mean.
> >
> > My vote, of course: +1
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
> > For additional commands, e-mail: commons-dev-help@jakarta.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: commons-dev-unsubscribe@jakarta.apache.org
For additional commands, e-mail: commons-dev-help@jakarta.apache.org