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

Re: Untitled window on re-activation

Expand Messages
  • björn
    ... Michael, thanks for confirming the problem. Are you using identical .[g]vimrc files on both machines? Does the problem persist if these (and the .vim
    Message 1 of 17 , Feb 4, 2009
      2009/2/3 Michael Wookey:
      >
      > Yes, I see the same behaviour on one machine but it works fine on
      > another. Where it works, I built MacVim myself. Where it doesn't,
      > I'm using the official pre-built snapshot 42. I'm not suggesting a
      > correlation, just an observation. Both machines are Intel with
      > 10.5.6.
      >
      > [... trying a build ...]
      >
      > I just built MacVim on the machine where it doesn't work and the "new
      > window on activation" still doesn't work. This suggests some other
      > setting must be causing the failure.

      Michael, thanks for confirming the problem.

      Are you using identical .[g]vimrc files on both machines? Does the
      problem persist if these (and the .vim folder) are moved out of the
      way? Is the problem intermittent or does it fail to open a new window
      _every_ time MacVim is reactivated? (Tiago: can you answer these
      questions as well please.)

      Björn

      --~--~---------~--~----~------------~-------~--~----~
      You received this message from the "vim_mac" maillist.
      For more information, visit http://www.vim.org/maillist.php
      -~----------~----~----~----~------~----~------~--~---
    • Tiago Taveira
      ... The problem still persists after moving .vimrc/.gvimrc/.vim, and it always fails to open a new window. If there s something else I could try to track it
      Message 2 of 17 , Feb 4, 2009
        On Feb 4, 2:29 pm, björn <bjorn.winck...@...> wrote:
        > Are you using identical .[g]vimrc files on both machines?  Does the
        > problem persist if these (and the .vim folder) are moved out of the
        > way?  Is the problem intermittent or does it fail to open a new window
        > _every_ time MacVim is reactivated?  (Tiago: can you answer these
        > questions as well please.)

        The problem still persists after moving .vimrc/.gvimrc/.vim, and it
        always fails to open a new window.
        If there's something else I could try to track it down, feel free to
        ask :)

        Tiago
        --~--~---------~--~----~------------~-------~--~----~
        You received this message from the "vim_mac" maillist.
        For more information, visit http://www.vim.org/maillist.php
        -~----------~----~----~----~------~----~------~--~---
      • Michael Wookey
        ... Yes, my configuration is identical across all machines that I use. I use git to hold my configuration/dotfiles and the repo is hosted on Amazon S3; so all
        Message 3 of 17 , Feb 5, 2009
          >> Yes, I see the same behaviour on one machine but it works fine on
          >> another. Where it works, I built MacVim myself. Where it doesn't,
          >> I'm using the official pre-built snapshot 42. I'm not suggesting a
          >> correlation, just an observation. Both machines are Intel with
          >> 10.5.6.
          >>
          >> [... trying a build ...]
          >>
          >> I just built MacVim on the machine where it doesn't work and the "new
          >> window on activation" still doesn't work. This suggests some other
          >> setting must be causing the failure.
          >
          > Michael, thanks for confirming the problem.
          >
          > Are you using identical .[g]vimrc files on both machines? Does the
          > problem persist if these (and the .vim folder) are moved out of the
          > way? Is the problem intermittent or does it fail to open a new window
          > _every_ time MacVim is reactivated? (Tiago: can you answer these
          > questions as well please.)

          Yes, my configuration is identical across all machines that I use. I
          use git to hold my configuration/dotfiles and the repo is hosted on
          Amazon S3; so all my machines are easily synchronised.

          Unfortunately, the problem still occurs (each and every time) even
          when ~/.[g]vimrc and ~/.vim are removed. Something else appears
          responsible for this problem.

          Personally, this bug doesn't bother me so much. I tend to keep a
          single vim window up all the time... and it never closes.

          --~--~---------~--~----~------------~-------~--~----~
          You received this message from the "vim_mac" maillist.
          For more information, visit http://www.vim.org/maillist.php
          -~----------~----~----~----~------~----~------~--~---
        • Tobia Conforto
          I noticed this bug as well. It s not an essential feaure to me, so I didn t think of reporting it and just started hitting Cmd-N to open the first window
          Message 4 of 17 , Feb 11, 2009
            I noticed this bug as well. It's not an essential feaure to me, so I
            didn't think of reporting it and just started hitting Cmd-N to open
            the first window after reactivation, just as I would do to open the
            second and third windows in any case.

            björn wrote:
            > Does the problem persist if .[g]vimrc files (and the .vim folder)
            > are moved out of the way?

            Yes.

            > Is the problem intermittent or does it fail to open a new window
            > _every_ time MacVim is reactivated?

            Every time.

            In fact, the behaviour is so consistent that I was able to do a quick
            test on past snapshots (freshly dowloaded) and repeat the test a few
            times. I'm positive the bug was introduced in snapshot 41 (meaning
            that 40 works, 41 and 42 don't.)

            This is an Intel 10.5.6 with loads of input managers, plugins,
            logitech control panels and other random hacks installed ;-P I'll
            disable the ones I can find and let you know if it makes any difference.

            If there's any patch you'd like me to test, feel free to send it my way.

            -Tobia
            --~--~---------~--~----~------------~-------~--~----~
            You received this message from the "vim_mac" maillist.
            For more information, visit http://www.vim.org/maillist.php
            -~----------~----~----~----~------~----~------~--~---
          • björn
            ... If it is that consistent then it would be great if somebody could do a bisection with Git to find the exact commit where it happened. I had a look at the
            Message 5 of 17 , Feb 11, 2009
              2009/2/11 Tobia Conforto:
              >
              > I noticed this bug as well. It's not an essential feaure to me, so I
              > didn't think of reporting it and just started hitting Cmd-N to open
              > the first window after reactivation, just as I would do to open the
              > second and third windows in any case.
              >
              > björn wrote:
              >> Does the problem persist if .[g]vimrc files (and the .vim folder)
              >> are moved out of the way?
              >
              > Yes.
              >
              >> Is the problem intermittent or does it fail to open a new window
              >> _every_ time MacVim is reactivated?
              >
              > Every time.
              >
              > In fact, the behaviour is so consistent that I was able to do a quick
              > test on past snapshots (freshly dowloaded) and repeat the test a few
              > times. I'm positive the bug was introduced in snapshot 41 (meaning
              > that 40 works, 41 and 42 don't.)

              If it is that consistent then it would be great if somebody could do a
              bisection with Git to find the exact commit where it happened. I had
              a look at the commits between 40 and 41 but I can't "guess" which one
              might be the offending commit. Does anybody with this problem know
              how to do this and have time? (I still can't reproduce reliably but
              it has happened once or twice.)

              Thanks,
              Björn

              --~--~---------~--~----~------------~-------~--~----~
              You received this message from the "vim_mac" maillist.
              For more information, visit http://www.vim.org/maillist.php
              -~----------~----~----~----~------~----~------~--~---
            • Keith Hubbard
              ... It looks like this is the culprit: 662aa32b399c1149daa1cd3a2b5bccc8e75f589a is first bad commit commit 662aa32b399c1149daa1cd3a2b5bccc8e75f589a Author:
              Message 6 of 17 , Feb 11, 2009
                björn wrote:
                > If it is that consistent then it would be great if somebody could do a
                > bisection with Git to find the exact commit where it happened. I had
                > a look at the commits between 40 and 41 but I can't "guess" which one
                > might be the offending commit. Does anybody with this problem know
                > how to do this and have time? (I still can't reproduce reliably but
                > it has happened once or twice.)
                >

                It looks like this is the culprit:

                662aa32b399c1149daa1cd3a2b5bccc8e75f589a is first bad commit
                commit 662aa32b399c1149daa1cd3a2b5bccc8e75f589a
                Author: Bjorn Winckler <bjorn.winckler@...>
                Date: Wed Jan 7 20:19:44 2009 +0100

                Prune unnecessary code, clarify comment

                :040000 040000 0e451b9611fb96374a76c4319e86aa0d56fdc19e
                d36cd9b7ad35c238218df462a9233614e4a0d480 M src

                662aa32b399c1149daa1cd3a2b5bccc8e75f589a


                --~--~---------~--~----~------------~-------~--~----~
                You received this message from the "vim_mac" maillist.
                For more information, visit http://www.vim.org/maillist.php
                -~----------~----~----~----~------~----~------~--~---
              • björn
                ... Thanks Keith, that gives me somewhere to start looking. I don t see that this commit is directly to blame...there seems to be some other problem related
                Message 7 of 17 , Feb 11, 2009
                  2009/2/11 Keith Hubbard:
                  >
                  > björn wrote:
                  >> If it is that consistent then it would be great if somebody could do a
                  >> bisection with Git to find the exact commit where it happened. I had
                  >> a look at the commits between 40 and 41 but I can't "guess" which one
                  >> might be the offending commit. Does anybody with this problem know
                  >> how to do this and have time? (I still can't reproduce reliably but
                  >> it has happened once or twice.)
                  >>
                  >
                  > It looks like this is the culprit:
                  >
                  > 662aa32b399c1149daa1cd3a2b5bccc8e75f589a is first bad commit
                  > commit 662aa32b399c1149daa1cd3a2b5bccc8e75f589a
                  > Author: Bjorn Winckler <bjorn.winckler@...>
                  > Date: Wed Jan 7 20:19:44 2009 +0100
                  >
                  > Prune unnecessary code, clarify comment

                  Thanks Keith, that gives me somewhere to start looking. I don't see
                  that this commit is directly to blame...there seems to be some other
                  problem related to detection of "untitled windows" and launching
                  processes so I'll have to dig deeper. It is weird too...right now I
                  have one instance of MacVim open where reactivation fails every time,
                  but when I build a new version and run it from Xcode then I cannot for
                  the life of me reproduce the bug. Hmmm...I should be able to figure
                  it out now though.

                  Thanks,
                  Björn

                  --~--~---------~--~----~------------~-------~--~----~
                  You received this message from the "vim_mac" maillist.
                  For more information, visit http://www.vim.org/maillist.php
                  -~----------~----~----~----~------~----~------~--~---
                • björn
                  ... Aaah. If MacVim is started from the command line then the bug appears (i.e. Cmd-q if MacVim is open, then type mvim in Terminal), if it is started from
                  Message 8 of 17 , Feb 11, 2009
                    2009/2/12 björn:
                    > It is weird too...right now I
                    > have one instance of MacVim open where reactivation fails every time,
                    > but when I build a new version and run it from Xcode then I cannot for
                    > the life of me reproduce the bug. Hmmm...I should be able to figure
                    > it out now though.

                    Aaah. If MacVim is started from the command line then the bug appears
                    (i.e. Cmd-q if MacVim is open, then type 'mvim' in Terminal), if it is
                    started from Finder it does not. This is why I could never see the
                    bug when I was trying to debug in Xcode! Can somebody confirm this?

                    Björn

                    --~--~---------~--~----~------------~-------~--~----~
                    You received this message from the "vim_mac" maillist.
                    For more information, visit http://www.vim.org/maillist.php
                    -~----------~----~----~----~------~----~------~--~---
                  • björn
                    ... The attached patch fixes the problem I ve noticed on my machine. I d appreciate if I could get some confirmation that it works since I m not sure if there
                    Message 9 of 17 , Feb 11, 2009
                      2009/2/12 björn:
                      > 2009/2/12 björn:
                      >> It is weird too...right now I
                      >> have one instance of MacVim open where reactivation fails every time,
                      >> but when I build a new version and run it from Xcode then I cannot for
                      >> the life of me reproduce the bug. Hmmm...I should be able to figure
                      >> it out now though.
                      >
                      > Aaah. If MacVim is started from the command line then the bug appears
                      > (i.e. Cmd-q if MacVim is open, then type 'mvim' in Terminal), if it is
                      > started from Finder it does not. This is why I could never see the
                      > bug when I was trying to debug in Xcode! Can somebody confirm this?

                      The attached patch fixes the problem I've noticed on my machine. I'd
                      appreciate if I could get some confirmation that it works since I'm
                      not sure if there is more than one bug.

                      Thanks!
                      Björn

                      --~--~---------~--~----~------------~-------~--~----~
                      You received this message from the "vim_mac" maillist.
                      For more information, visit http://www.vim.org/maillist.php
                      -~----------~----~----~----~------~----~------~--~---
                    • Keith Hubbard
                      ... I think there are multiple bugs -- your patch did not fix the problem on my machine. I did more experimenting, and I only see the problem when Quickstart
                      Message 10 of 17 , Feb 11, 2009
                        björn wrote:
                        > The attached patch fixes the problem I've noticed on my machine. I'd
                        > appreciate if I could get some confirmation that it works since I'm
                        > not sure if there is more than one bug.
                        >
                        I think there are multiple bugs -- your patch did not fix the problem on
                        my machine.

                        I did more experimenting, and I only see the problem when Quickstart is
                        enabled. I've been testing by launching MacVim.app using open ("open
                        build/Release/MacVim.app"). I'm also pretty confident that commit
                        662aa32 introduced the problem.

                        Keith


                        --~--~---------~--~----~------------~-------~--~----~
                        You received this message from the "vim_mac" maillist.
                        For more information, visit http://www.vim.org/maillist.php
                        -~----------~----~----~----~------~----~------~--~---
                      • Tobia Conforto
                        ... I can confirm this. Also, I see the problem independent of the way I start MacVim. I ll try to do some Git testing later, if I have time. Tobia
                        Message 11 of 17 , Feb 12, 2009
                          Keith Hubbard wrote:
                          > I did more experimenting, and I only see the problem when Quickstart
                          > is
                          > enabled.

                          I can confirm this.
                          Also, I see the problem independent of the way I start MacVim.

                          I'll try to do some Git testing later, if I have time.

                          Tobia

                          --~--~---------~--~----~------------~-------~--~----~
                          You received this message from the "vim_mac" maillist.
                          For more information, visit http://www.vim.org/maillist.php
                          -~----------~----~----~----~------~----~------~--~---
                        • björn
                          ... Ok, I can also reproduce this reliably now. The key is indeed to turn on Quickstart -- as soon as this is done the reactivation doesn t work, even if
                          Message 12 of 17 , Feb 12, 2009
                            2009/2/12 Keith Hubbard:
                            >
                            > björn wrote:
                            >> The attached patch fixes the problem I've noticed on my machine. I'd
                            >> appreciate if I could get some confirmation that it works since I'm
                            >> not sure if there is more than one bug.
                            >>
                            > I think there are multiple bugs -- your patch did not fix the problem on
                            > my machine.
                            >
                            > I did more experimenting, and I only see the problem when Quickstart is
                            > enabled. I've been testing by launching MacVim.app using open ("open
                            > build/Release/MacVim.app"). I'm also pretty confident that commit
                            > 662aa32 introduced the problem.

                            Ok, I can also reproduce this reliably now. The key is indeed to turn
                            on Quickstart -- as soon as this is done the reactivation doesn't
                            work, even if Quickstart is disabled again. If I restart without
                            Quickstart reactivation works.

                            So there are (at least) two bugs. It should be simple to fix this now
                            that I can reproduce it...I'll post again as soon as I have a patch
                            ready.

                            /Björn

                            --~--~---------~--~----~------------~-------~--~----~
                            You received this message from the "vim_mac" maillist.
                            For more information, visit http://www.vim.org/maillist.php
                            -~----------~----~----~----~------~----~------~--~---
                          • björn
                            ... Well, I ended up reverting the commit Keith pointed at. In itself that commit seemed fine but there are some nasty side-effects when quickstart is
                            Message 13 of 17 , Feb 13, 2009
                              2009/2/12 björn:
                              > 2009/2/12 Keith Hubbard:
                              >>
                              >> björn wrote:
                              >>> The attached patch fixes the problem I've noticed on my machine. I'd
                              >>> appreciate if I could get some confirmation that it works since I'm
                              >>> not sure if there is more than one bug.
                              >>>
                              >> I think there are multiple bugs -- your patch did not fix the problem on
                              >> my machine.
                              >>
                              >> I did more experimenting, and I only see the problem when Quickstart is
                              >> enabled. I've been testing by launching MacVim.app using open ("open
                              >> build/Release/MacVim.app"). I'm also pretty confident that commit
                              >> 662aa32 introduced the problem.
                              >
                              > Ok, I can also reproduce this reliably now. The key is indeed to turn
                              > on Quickstart -- as soon as this is done the reactivation doesn't
                              > work, even if Quickstart is disabled again. If I restart without
                              > Quickstart reactivation works.
                              >
                              > So there are (at least) two bugs. It should be simple to fix this now
                              > that I can reproduce it...I'll post again as soon as I have a patch
                              > ready.

                              Well, I ended up reverting the commit Keith pointed at. In itself
                              that commit seemed fine but there are some nasty side-effects when
                              quickstart is enabled. At first I thought it would be best to rewrite
                              the affected code to clarify what's going on but decided that it is
                              easier to just revert that commit and comment the behaviour better.
                              I'm not very satisfied with this situation, but I might refactor later
                              if I get the time.

                              Thanks to everybody who reported the problem and in particular to
                              Keith for finding the offending commit. Some confirmation that things
                              work again would be nice, but I am pretty confident that all is well
                              now.

                              Björn

                              --~--~---------~--~----~------------~-------~--~----~
                              You received this message from the "vim_mac" maillist.
                              For more information, visit http://www.vim.org/maillist.php
                              -~----------~----~----~----~------~----~------~--~---
                            • Tiago Taveira
                              ... Works nicely :) Thanks for fixing it! Tiago --~--~---------~--~----~------------~-------~--~----~ You received this message from the vim_mac maillist.
                              Message 14 of 17 , Feb 23, 2009
                                On Feb 13, 2009, at 9:48 PM, björn wrote:
                                > Some confirmation that things work again
                                > would be nice, but I am pretty confident
                                > that all is well now.
                                >
                                > Björn

                                Works nicely :)
                                Thanks for fixing it!

                                Tiago
                                --~--~---------~--~----~------------~-------~--~----~
                                You received this message from the "vim_mac" maillist.
                                For more information, visit http://www.vim.org/maillist.php
                                -~----------~----~----~----~------~----~------~--~---
                              Your message has been successfully submitted and would be delivered to recipients shortly.