Loading ...
Sorry, an error occurred while loading the content.

Re: [perlsemny] Any interest in hearing about and giving feedback on a new Perl Debugger I've been working on?

Expand Messages
  • Steven Lembark
    On Tue, 27 Dec 2011 16:08:34 -0500 ... Please don t. 5.8 is no longer supported, we have enough people who won t change just because they use 5.8. I have
    Message 1 of 8 , Jan 3, 2012
    • 0 Attachment
      On Tue, 27 Dec 2011 16:08:34 -0500
      Rocky Bernstein <rocky@...> wrote:

      > > Q: What version(s) of Perl have you tested it with?
      > >
      >
      > It requires Perl 5.10.0 or greater. The code could be rewritten for earlier
      > versions but I'd have to remove uses of "state" and "switch".

      Please don't. 5.8 is no longer supported, we have
      enough people who won't change "just because" they
      use 5.8.

      I have 5.14 here.

      > For OS's and Perl versions, I use GNU/Linux but I've also tried out on
      > Strawberry Perl, ActiveState Perl and cygwin.

      I'll be testing it on Gentoo (if I can sit in one place long
      enough to try it).

      Q: Do you have any screen shots of using it on anything
      other than linux?

      I can try to show people what it looks like at our next
      Perl Monger's meeting.

      > And of course CPANTS can be consulted for other Unixish servers that this
      > works on:
      >
      > http://www.cpantesters.org/distro/D/Devel-Trepan.html#Devel-Trepan-v0.1.4
      >
      > STO's smokers seem to fail in a weird way in the integration tests.
      > Possibly this iI because in some of them I fork, run the debugger and
      > compare output. Something in his threaded Perl kernels (and possibly
      > Module::Build which version 0.39_01 -- which I hadn't known existed and
      > can't generally find) cause Test::More to run in an unexpected way. I've
      > sent email to him to no avail.
      >
      > The next release I will reduce the dependencies and also skip those failing
      > tests on his smokers.

      I'll try to give it a try this week.
      Hopefully I can get enough data together to give a
      short presentation of it on the 18th :-)

      enjoi

      --
      Steven Lembark 3646 Flora Pl
      Workhorse Computing St Louis, MO 63110
      lembark@... +1 888 359 3508
    • Rocky Bernstein
      ... Too late. I wrote that a week ago. Over the New Year, I have made this 5.8.8 compatible. At work we have RHEL 5.x servers and 5.8.8 is what comes shipped
      Message 2 of 8 , Jan 3, 2012
      • 0 Attachment
        On Tue, Jan 3, 2012 at 4:08 PM, Steven Lembark <lembark@...> wrote:
         

        On Tue, 27 Dec 2011 16:08:34 -0500
        Rocky Bernstein <rocky@...> wrote:

        > > Q: What version(s) of Perl have you tested it with?
        > >
        >
        > It requires Perl 5.10.0 or greater. The code could be rewritten for earlier
        > versions but I'd have to remove uses of "state" and "switch".

        Please don't. 5.8 is no longer supported, we have
        enough people who won't change "just because" they
        use 5.8.

        Too late. I wrote that a week ago. Over the New Year, I have made this 5.8.8 compatible. At work we have RHEL 5.x servers and 5.8.8 is what comes shipped with that.  (And to date, I'm probably the only one I know of who has seriously used this so I may as well accommodate myself ;-)

        But that said, this works better on later Perls. JJORE's Enbugger doesn't run on 5.8 to my knowledge. And RHEL did something to break Eval::WithLexicals.



        I have 5.14 here.

        Since Eval::WithLexicals + PadWalker doesn't handle local variables, right now I'm considering working up a patch to Perl5 to do this much simpler, more reliably and cleaner.

        If that happens, even 5.14 will be too passe.


        > For OS's and Perl versions, I use GNU/Linux but I've also tried out on
        > Strawberry Perl, ActiveState Perl and cygwin.

        I'll be testing it on Gentoo (if I can sit in one place long
        enough to try it).

        Q: Do you have any screen shots of using it on anything
        other than linux?

        Since you asked I created  a sample session in a cygwin terminal: https://github.com/rocky/Perl-Devel-Trepan/wiki/Sample-Cygwin-Session

        Strawberry Perl's terminal doesn't seem to be able to handle colors. ActiveState Perl would be about the same because all one has to do there is set PATH in a cygwin window. There are other terminals in cygwin/mingw that understand escape sequences such as rxvt and I've used those in the past.

        And of course one doesn't *have* to use syntax coloring and highlighting. Personally, I think it is still a very good debugger** even without that or the command completion.

        **good in terms of design and testing and user experience. Any debugger though has very wacky and Rube-Goldberg stuff in it that is foisted on it by the runtime. So in this respect, perl5db is the more battle worn. But I've noticed that some of the whacky things in this debugger are equally whacky in perl5db


        I can try to show people what it looks like at our next
        Perl Monger's meeting.

        Great! I'd be happy to go over any material you put together and if there's a video of that made or want to make the presentation made available I can either add it to the github wiki page or add a link to it if it is otherwise publicly available.
         


        > And of course CPANTS can be consulted for other Unixish servers that this
        > works on:
        >
        > http://www.cpantesters.org/distro/D/Devel-Trepan.html#Devel-Trepan-v0.1.4

        An update I- the current results which pass every smoker that the current version has been tried on:

         http://www.cpantesters.org/distro/D/Devel-Trepan.html

        This is 16 smokers which includes OS's GNU/Linux, OpenBSD, FreeBSD, Windows (Win32). Furthermore, at present any OS listed in the matrix (so that includes OSX and NetBSD) has passed on some version of Devel::Trepan.

        >
        > STO's smokers seem to fail in a weird way in the integration tests.
        > Possibly this iI because in some of them I fork, run the debugger and
        > compare output. Something in his threaded Perl kernels (and possibly
        > Module::Build which version 0.39_01 -- which I hadn't known existed and
        > can't generally find) cause Test::More to run in an unexpected way. I've
        > sent email to him to no avail.
        >
        > The next release I will reduce the dependencies and also skip those failing
        > tests on his smokers.

        I'll try to give it a try this week.
        Hopefully I can get enough data together to give a
        short presentation of it on the 18th :-)

        enjoi


        --
        Steven Lembark 3646 Flora Pl
        Workhorse Computing St Louis, MO 63110
        lembark@... +1 888 359 3508


      • Morris M. Siegel
        I haven t played with the new debugger yet, but I d be interested in attending a presentation about it. By the way, Rocky, you may perhaps be interested in
        Message 3 of 8 , Jan 3, 2012
        • 0 Attachment
          I haven't played with the new debugger yet, but I'd be interested in attending a presentation about it.

          By the way, Rocky, you may perhaps be interested in seeing whether any of the ideas or code of Gregor Purdy's Perl Shell (psh [1,2]) are relevant to your work.

          [1] http://www.focusresearch.com/gregor/sw/psh/
          [2] http://sourceforge.net/projects/psh/

          -- Morrie Siegel
             mmsiegel@...
          ______________________________________________________________________________

          On Wed, Jan 4, 2012 at 4:09 AM, Rocky Bernstein <rocky@...> wrote:

          On Tue, Jan 3, 2012 at 4:08 PM, Steven Lembark <lembark@...> wrote:
           

          On Tue, 27 Dec 2011 16:08:34 -0500
          Rocky Bernstein <rocky@...> wrote:

          > > Q: What version(s) of Perl have you tested it with?
          > >
          >
          > It requires Perl 5.10.0 or greater. The code could be rewritten for earlier
          > versions but I'd have to remove uses of "state" and "switch".

          Please don't. 5.8 is no longer supported, we have
          enough people who won't change "just because" they
          use 5.8.

          Too late. I wrote that a week ago. Over the New Year, I have made this 5.8.8 compatible. At work we have RHEL 5.x servers and 5.8.8 is what comes shipped with that.  (And to date, I'm probably the only one I know of who has seriously used this so I may as well accommodate myself ;-)

          But that said, this works better on later Perls. JJORE's Enbugger doesn't run on 5.8 to my knowledge. And RHEL did something to break Eval::WithLexicals.



          I have 5.14 here.

          Since Eval::WithLexicals + PadWalker doesn't handle local variables, right now I'm considering working up a patch to Perl5 to do this much simpler, more reliably and cleaner.

          If that happens, even 5.14 will be too passe.


          > For OS's and Perl versions, I use GNU/Linux but I've also tried out on
          > Strawberry Perl, ActiveState Perl and cygwin.

          I'll be testing it on Gentoo (if I can sit in one place long
          enough to try it).

          Q: Do you have any screen shots of using it on anything
          other than linux?

          Since you asked I created  a sample session in a cygwin terminal: https://github.com/rocky/Perl-Devel-Trepan/wiki/Sample-Cygwin-Session

          Strawberry Perl's terminal doesn't seem to be able to handle colors. ActiveState Perl would be about the same because all one has to do there is set PATH in a cygwin window. There are other terminals in cygwin/mingw that understand escape sequences such as rxvt and I've used those in the past.

          And of course one doesn't *have* to use syntax coloring and highlighting. Personally, I think it is still a very good debugger** even without that or the command completion.

          **good in terms of design and testing and user experience. Any debugger though has very wacky and Rube-Goldberg stuff in it that is foisted on it by the runtime. So in this respect, perl5db is the more battle worn. But I've noticed that some of the whacky things in this debugger are equally whacky in perl5db


          I can try to show people what it looks like at our next
          Perl Monger's meeting.

          Great! I'd be happy to go over any material you put together and if there's a video of that made or want to make the presentation made available I can either add it to the github wiki page or add a link to it if it is otherwise publicly available.
           


          > And of course CPANTS can be consulted for other Unixish servers that this
          > works on:
          >
          > http://www.cpantesters.org/distro/D/Devel-Trepan.html#Devel-Trepan-v0.1.4

          An update I- the current results which pass every smoker that the current version has been tried on:

           http://www.cpantesters.org/distro/D/Devel-Trepan.html

          This is 16 smokers which includes OS's GNU/Linux, OpenBSD, FreeBSD, Windows (Win32). Furthermore, at present any OS listed in the matrix (so that includes OSX and NetBSD) has passed on some version of Devel::Trepan.

          >
          > STO's smokers seem to fail in a weird way in the integration tests.
          > Possibly this iI because in some of them I fork, run the debugger and
          > compare output. Something in his threaded Perl kernels (and possibly
          > Module::Build which version 0.39_01 -- which I hadn't known existed and
          > can't generally find) cause Test::More to run in an unexpected way. I've
          > sent email to him to no avail.
          >
          > The next release I will reduce the dependencies and also skip those failing
          > tests on his smokers.

          I'll try to give it a try this week.
          Hopefully I can get enough data together to give a
          short presentation of it on the 18th :-)

          enjoi


          --
          Steven Lembark 3646 Flora Pl
          Workhorse Computing St Louis, MO 63110
          lembark@... +1 888 359 3508





        Your message has been successfully submitted and would be delivered to recipients shortly.