Loading ...
Sorry, an error occurred while loading the content.
Skip to search.
 

Re: Patch 7.3.970

Expand Messages
  • Charles
    ... Here s my benchmark result Benchmark: newengine vs oldengine Trial #1 newengine : 0.746297 oldengine : 2.197819 Trial #2 newengine : 0.742969 oldengine :
    Message 1 of 25 , May 20, 2013
      On Mon, May 20, 2013 at 9:12 AM, Ben Fritz <fritzophrenic@...> wrote:
      >
      > On Sunday, May 19, 2013 12:44:49 PM UTC-5, Bram Moolenaar wrote:
      > > I wrote:
      > >
      > >
      > >
      > > > Patch 7.3.970
      > >
      > >
      > >
      > > All the tests I could do with the new engine pass. However, it is
      > >
      > > noticeable slower than the old engine. If this bothers you, set the
      > >
      > > 'regexpengine' option to one.
      > >
      >
      > Wasn't the main purpose of the new engine to speed things up? Or are we sacrificing average speed for better worst-case speed?
      >

      Here's my benchmark result

      Benchmark: newengine vs oldengine

      Trial #1
      newengine : 0.746297
      oldengine : 2.197819

      Trial #2
      newengine : 0.742969
      oldengine : 2.195868

      Trial #3
      newengine : 0.746451
      oldengine : 2.199959

      --
      --
      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+unsubscribe@....
      For more options, visit https://groups.google.com/groups/opt_out.
    • John Marriott
      Hi All, Attempting to compile on HP-UX fails after this patch with this message: cc -c -I. -Iproto -DHAVE_CONFIG_H -O2 -o objects/regexp.o regexp.c cc:
      Message 2 of 25 , May 20, 2013
        Hi All,

        Attempting to compile on HP-UX fails after this patch with this message:
            cc -c -I. -Iproto -DHAVE_CONFIG_H     -O2       -o objects/regexp.o regexp.c

        cc: "regexp_nfa.c", line 1966: error 1521: Incorrect initialization.

        It seems that HP's cc doesn't like this initialisation:
        /*
         * Initialize Frag_T struct.
         */
            static Frag_T
        frag(start, out)
            nfa_state_T    *start;
            Ptrlist    *out;
        {
            Frag_T n = { start, out };
            return n;
        }

        After changing it to:
        /*
         * Initialize Frag_T struct.
         */
            static Frag_T
        frag(start, out)
            nfa_state_T    *start;
            Ptrlist    *out;
        {
            Frag_T n;

            n.start = start;
            n.out = out;

            return n;
        }


        It's happy again.

        Cheers
        John

        On 20-May-2013 3:44 AM, Bram Moolenaar wrote:
        I wrote:
        
        
        Patch 7.3.970
        
        All the tests I could do with the new engine pass.  However, it is
        noticeable slower than the old engine.  If this bothers you, set the
        'regexpengine' option to one.
        
        If you spot a mistake in regexp pattern matching, please send a
        reproducible example, so that we can add it to the tests.
        
        

        --
        --
        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+unsubscribe@....
        For more options, visit https://groups.google.com/groups/opt_out.
         
         
      • Bram Moolenaar
        ... So the new engine is faster? Or what do the numbers mean? My experience is that the new engine is slower. Thus we have work to do to make it faster. --
        Message 3 of 25 , May 20, 2013
          Charles Peacech wrote:

          > On Mon, May 20, 2013 at 9:12 AM, Ben Fritz <fritzophrenic@...> wrote:
          > >
          > > On Sunday, May 19, 2013 12:44:49 PM UTC-5, Bram Moolenaar wrote:
          > > > I wrote:
          > > >
          > > >
          > > >
          > > > > Patch 7.3.970
          > > >
          > > >
          > > >
          > > > All the tests I could do with the new engine pass. However, it is
          > > >
          > > > noticeable slower than the old engine. If this bothers you, set the
          > > >
          > > > 'regexpengine' option to one.
          > > >
          > >
          > > Wasn't the main purpose of the new engine to speed things up? Or are we sacrificing average speed for better worst-case speed?
          > >
          >
          > Here's my benchmark result
          >
          > Benchmark: newengine vs oldengine
          >
          > Trial #1
          > newengine : 0.746297
          > oldengine : 2.197819
          >
          > Trial #2
          > newengine : 0.742969
          > oldengine : 2.195868
          >
          > Trial #3
          > newengine : 0.746451
          > oldengine : 2.199959

          So the new engine is faster? Or what do the numbers mean?

          My experience is that the new engine is slower. Thus we have work to do
          to make it faster.

          --
          Close your shells, or I'll kill -9 you
          Tomorrow I'll quota you
          Remember the disks'll always be full
          And then while I'm away
          I'll write ~ everyday
          And I'll send-pr all my buggings to you.
          [ CVS log "Beatles style" for FreeBSD ports/INDEX, Satoshi Asami ]

          /// Bram Moolenaar -- Bram@... -- 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+unsubscribe@....
          For more options, visit https://groups.google.com/groups/opt_out.
        • Bram Moolenaar
          ... Thanks, I can reproduce that. It s very helpful to get a good reproducible example. -- Snoring is prohibited unless all bedroom windows are closed and
          Message 4 of 25 , May 20, 2013
            Taro Muraoka wrote:

            > I got a report that 970 causes crash when do this:
            >
            > :echo search("[^A-z]")
            >
            > # reported at here in Japanese -> https://github.com/vim-jp/issues/issues/386
            >
            > Yasuhiro Matsumoto try to write a patch to fix this,
            > but it seems that there are some other similar problems.

            Thanks, I can reproduce that. It's very helpful to get a good
            reproducible example.

            --
            Snoring is prohibited unless all bedroom windows are closed and securely
            locked.
            [real standing law in Massachusetts, United States of America]

            /// Bram Moolenaar -- Bram@... -- 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+unsubscribe@....
            For more options, visit https://groups.google.com/groups/opt_out.
          • Bram Moolenaar
            ... You can uncomment this line in regexp.c: /* #undef DEBUG */ -- Zen Microsystems: we re the om in .commmmmmmmm /// Bram Moolenaar -- Bram@Moolenaar.net --
            Message 5 of 25 , May 20, 2013
              Taro Muraoka wrote:

              > This makes vim (debug build) slow to startup.
              > New NFA engine seems to write huge log files in debug build.
              > And it causes this terrible slow down.
              >
              > Added files are generated like this:
              >
              > gvim -u NONE -U NONE --noplugin --startuptime startuptime-7.3.XXX.txt
              >
              > I'll look into source codes later, but reported at first.

              You can uncomment this line in regexp.c:

              /* #undef DEBUG */


              --
              Zen Microsystems: we're the om in .commmmmmmmm

              /// Bram Moolenaar -- Bram@... -- 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+unsubscribe@....
              For more options, visit https://groups.google.com/groups/opt_out.
            • Bram Moolenaar
              ... The new engine is supposed to be faster. But we are not there yet. It might be that the new engine is only faster for repeating. It has more overhead
              Message 6 of 25 , May 20, 2013
                Ben Fritz wrote:

                > On Sunday, May 19, 2013 12:44:49 PM UTC-5, Bram Moolenaar wrote:
                > > I wrote:
                > >
                > > > Patch 7.3.970
                > >
                > > All the tests I could do with the new engine pass. However, it is
                > >
                > > noticeable slower than the old engine. If this bothers you, set the
                > >
                > > 'regexpengine' option to one.
                > >
                >
                > Wasn't the main purpose of the new engine to speed things up? Or are
                > we sacrificing average speed for better worst-case speed?

                The new engine is supposed to be faster. But we are not there yet.

                It might be that the new engine is only faster for repeating. It has
                more overhead setting up the regexp. If we can't reduce the overhead we
                might have to select the engine when compiling the regexp.

                --
                So when I saw the post to comp.editors, I rushed over to the FTP site to
                grab it. So I yank apart the tarball, light x candles, where x= the
                vim version multiplied by the md5sum of the source divided by the MAC of
                my NIC (8A3FA78155A8A1D346C3C4A), put on black robes, dim the lights,
                wave a dead chicken over the hard drive, and summon the power of GNU GCC
                with the magic words "make config ; make!".
                [Jason Spence, compiling Vim 5.0]

                /// Bram Moolenaar -- Bram@... -- 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+unsubscribe@....
                For more options, visit https://groups.google.com/groups/opt_out.
              • Ron Aaron
                ... Just doing gvim --startuptime a.txt with regexpengine=0 and ... b.txt with it =1, shows about a second less time (e.g. faster) with =1 over =0 on my
                Message 7 of 25 , May 20, 2013
                  On Monday, May 20, 2013 2:34:14 PM UTC+3, Bram Moolenaar wrote:

                  >
                  > So the new engine is faster? Or what do the numbers mean?
                  >
                  >
                  >
                  > My experience is that the new engine is slower. Thus we have work to do
                  >
                  > to make it faster.

                  Just doing "gvim --startuptime a.txt" with regexpengine=0 and "... b.txt" with it =1, shows about a second less time (e.g. faster) with =1 over =0 on my work box (64bit linux)

                  --
                  --
                  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+unsubscribe@....
                  For more options, visit https://groups.google.com/groups/opt_out.
                • Bram Moolenaar
                  John Marriott wrote: [html removed] ... Thanks. That is a modern initialization construct, not supported by older compilers. -- Violators can be fined,
                  Message 8 of 25 , May 20, 2013
                    John Marriott wrote:

                    [html removed]

                    > Attempting to compile on HP-UX fails after this patch with this
                    > message:
                    >
                    > cc -c -I. -Iproto
                    > -DHAVE_CONFIG_H -O2 -o objects/regexp.o regexp.c
                    >
                    > cc: "regexp_nfa.c",
                    > line 1966: error 1521: Incorrect initialization.
                    >
                    >
                    > It
                    > seems that HP's cc doesn't like this initialisation:
                    >
                    > /*
                    > * Initialize Frag_T struct.
                    > */
                    > static Frag_T
                    > frag(start, out)
                    > nfa_state_T *start;
                    > Ptrlist *out;
                    > {
                    > Frag_T n = { start, out };
                    > return n;
                    > }
                    >
                    > After
                    > changing it to:
                    >
                    > /*
                    > * Initialize Frag_T struct.
                    > */
                    >
                    > static Frag_T
                    > frag(start, out)
                    > nfa_state_T *start;
                    > Ptrlist *out;
                    > {
                    > Frag_T n;
                    >
                    > n.start = start;
                    > n.out = out;
                    > return n;
                    >
                    > }
                    >
                    > It's happy again.

                    Thanks. That is a modern initialization construct, not supported
                    by older compilers.

                    --
                    Violators can be fined, arrested or jailed for making ugly faces at a dog.
                    [real standing law in Oklahoma, United States of America]

                    /// Bram Moolenaar -- Bram@... -- 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+unsubscribe@....
                    For more options, visit https://groups.google.com/groups/opt_out.
                  • Karsten Hopp
                    ... The patch in this mail and the one on the ftp site is broken. It contains the patch for src/testdir/Makefile twice. The mercurial patch 4444:ccecb03e5e8b
                    Message 9 of 25 , May 21, 2013
                      Am 19.05.2013 19:15, schrieb Bram Moolenaar:
                      > Patch 7.3.970
                      > Problem: Syntax highlighting can be slow.
                      > Solution: Include the NFA regexp engine. Add the 'regexpengine' option to
                      > select which one is used. (various authors, including Ken Takata,
                      > Andrei Aiordachioaie, Russ Cox, Xiaozhou Liua, Ian Young)
                      > Files: src/Make_cyg.mak, src/Make_ming.mak, src/Make_mvc.mak,
                      > src/Makefile, src/regexp.c, src/regexp.h, src/regexp_nfa.c,
                      > src/structs.h, src/testdir/Makefile, src/testdir/test64.in,
                      > src/testdir/test64.ok, Filelist, runtime/doc/pattern.txt,
                      > runtime/doc/option.txt, src/option.c, src/option.h,
                      > src/testdir/test95.in, src/testdir/test95.ok,
                      > src/testdir/Make_amiga.mak, src/testdir/Make_dos.mak,
                      > src/testdir/Make_ming.mak, src/testdir/Make_os2.mak,
                      > src/testdir/Make_vms.mms, src/testdir/Makefile
                      >
                      >

                      The patch in this mail and the one on the ftp site is broken. It contains the patch for src/testdir/Makefile twice.
                      The mercurial patch 4444:ccecb03e5e8b (7.3.970) is ok.

                      Karsten

                      --
                      --
                      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+unsubscribe@....
                      For more options, visit https://groups.google.com/groups/opt_out.
                    • Bram Moolenaar
                      ... Oops. I corrected the ftp site. Can t correct the email.. Obviously using Mercurial is the way to go these days. -- A fool must search for a greater fool
                      Message 10 of 25 , May 21, 2013
                        Karsten Hopp wrote:

                        > Am 19.05.2013 19:15, schrieb Bram Moolenaar:
                        > > Patch 7.3.970
                        > > Problem: Syntax highlighting can be slow.
                        > > Solution: Include the NFA regexp engine. Add the 'regexpengine' option to
                        > > select which one is used. (various authors, including Ken Takata,
                        > > Andrei Aiordachioaie, Russ Cox, Xiaozhou Liua, Ian Young)
                        > > Files: src/Make_cyg.mak, src/Make_ming.mak, src/Make_mvc.mak,
                        > > src/Makefile, src/regexp.c, src/regexp.h, src/regexp_nfa.c,
                        > > src/structs.h, src/testdir/Makefile, src/testdir/test64.in,
                        > > src/testdir/test64.ok, Filelist, runtime/doc/pattern.txt,
                        > > runtime/doc/option.txt, src/option.c, src/option.h,
                        > > src/testdir/test95.in, src/testdir/test95.ok,
                        > > src/testdir/Make_amiga.mak, src/testdir/Make_dos.mak,
                        > > src/testdir/Make_ming.mak, src/testdir/Make_os2.mak,
                        > > src/testdir/Make_vms.mms, src/testdir/Makefile
                        > >
                        > >
                        >
                        > The patch in this mail and the one on the ftp site is broken. It
                        > contains the patch for src/testdir/Makefile twice.
                        > The mercurial patch 4444:ccecb03e5e8b (7.3.970) is ok.

                        Oops. I corrected the ftp site. Can't correct the email..

                        Obviously using Mercurial is the way to go these days.

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

                        /// Bram Moolenaar -- Bram@... -- 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+unsubscribe@....
                        For more options, visit https://groups.google.com/groups/opt_out.
                      Your message has been successfully submitted and would be delivered to recipients shortly.