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

Re: [fusebox5] Problem with Load

Expand Messages
  • Damien McKenna
    ... OMG. That explains *so* *much* of the weirdness I see day to day on our server, I keep thinking it is the SFTP server not unlocking the files I upload or
    Message 1 of 14 , Aug 1 6:58 AM
    • 0 Attachment
      On Jul 30, 2006, at 5:23 PM, Sean Corfield wrote:
      However I am not sure that is the full story as I tried waiting and it
      didn't help.

      Right, if you reload before the minute changes, it thinks the core is  
      up to date.

      OMG.  That explains *so* *much* of the weirdness I see day to day on our server, I keep thinking it is the SFTP server not unlocking the files I upload or something.  Gah!  Thanks for clearing that up!

      -- 
      Damien McKenna - Web Developer - dmckenna@...
      The Limu Company - http://www.thelimucompany.com/ - 407-804-1014
      #include <stdjoke.h>


    • Sean Corfield
      ... This should only affect servers where returns the last modified timestamp in minutes rather than seconds... I used the exact same test that
      Message 2 of 14 , Aug 1 8:00 AM
      • 0 Attachment
        On Aug 1, 2006, at 6:58 AM, Damien McKenna wrote:
        > On Jul 30, 2006, at 5:23 PM, Sean Corfield wrote:
        >> Right, if you reload before the minute changes, it thinks the core is
        >> up to date.
        > OMG. That explains *so* *much* of the weirdness I see day to day
        > on our server, I keep thinking it is the SFTP server not unlocking
        > the files I upload or something. Gah! Thanks for clearing that up!

        This should only affect servers where <cfdirectory> returns the last
        modified timestamp in minutes rather than seconds...

        I used the exact same test that Fusebox 4 uses so folks should be
        seeing the same behavior there too.

        I think a possible workaround for now would be to change the 'gt'
        operator to 'gte' in the following places:
        - fuseboxApplication.cfc line 163
        - fuseboxCircuit.cfc line 110

        This wouldn't be optimal because it would repeatedly reload files on
        some O/S for the first minute after a change but it would probably
        solve your problems for now...

        A full fix is scheduled for Fusebox 5.1:

        http://projects.simb.net/fusebox/trac.cgi/ticket/99

        It will need testing on every platform to ensure relying on Java does
        not introduce cross-platform issues.

        Sean A Corfield -- (904) 302-SEAN
        An Architect's View -- http://corfield.org/

        "If you're not annoying somebody, you're not really alive."
        -- Margaret Atwood
      • Barney Boisvert
        ... Can you make that change conditional? Do a test CFDIRECTORY, check the granularity of the timestamps, and then use that info to do the checks in the
        Message 3 of 14 , Aug 1 8:23 AM
        • 0 Attachment
          On 8/1/06, Sean Corfield <sean@...> wrote:
          > I think a possible workaround for now would be to change the 'gt'
          > operator to 'gte' in the following places:
          > - fuseboxApplication.cfc line 163
          > - fuseboxCircuit.cfc line 110
          >
          > This wouldn't be optimal because it would repeatedly reload files on
          > some O/S for the first minute after a change but it would probably
          > solve your problems for now...

          Can you make that change conditional? Do a test CFDIRECTORY, check
          the granularity of the timestamps, and then use that info to do the
          checks in the optimal way. That'd also give you a fallback from using
          the Java solution.

          cheers,
          barneyb
          --
          Barney Boisvert
          bboisvert@...
          360.319.6145
          http://www.barneyb.com/

          Got Gmail? I have 100 invites.
        • Kevin Roche
          ... I have never noticed this problem with Fusebox 4 which I have been using for some time. It leads me to believe that somthing else is happening here. I will
          Message 4 of 14 , Aug 1 2:46 PM
          • 0 Attachment
            > On Tue 01/08/2006 at 16:01, Sean Corfield wrote:
            >This should only affect servers where <cfdirectory> returns the last
            >modified timestamp in minutes rather than seconds...
            >
            >I used the exact same test that Fusebox 4 uses so folks should be
            >seeing the same behavior there too.

            I have never noticed this problem with Fusebox 4 which I have been using for
            some time. It leads me to believe that somthing else is happening here. I
            will have to spend more time trying it out and checking Fusebox 4 too.

            Kevin
          • Terry L Schmitt
            I m going to chime in again... I brought this up a week or so ago, as I was having trouble with FB5 reloading the circuits when they change. I dropped the
            Message 5 of 14 , Aug 2 7:15 AM
            • 0 Attachment
              I'm going to chime in again...
              I brought this up a week or so ago, as I was having trouble with FB5 reloading the circuits when they change. I dropped the issue, because I couldn't really capture some meaningful symptoms and a simple test app reloaded the circuits just fine while in development-circuit-load. Go figure.

              The app where I'm having the trouble is not huge, but does have some complexity with a fair number of plugins and globalfuseactions firing.
              FB5 seems to recognize changes made to an existing fuseaction, but adding a new fuseaction to a circuit is not recognized. The work-around is to simply reload manually using the URL parameters.

              I am using English Windows 2003 Server. The cfdirectory granularity is to the second, so that should not be the concern in my case.

              I will keep this in mind as I develop, but unfortunately I am so stacked up at work that I can't put serious time into researching this more.

              Terry





              "Kevin Roche" <kevin@...>
              Sent by: fusebox5@yahoogroups.com

              08/01/2006 05:46 PM

              Please respond to
              fusebox5@yahoogroups.com

              To
              <fusebox5@yahoogroups.com>
              cc
              Subject
              RE: [fusebox5] Problem with Load






              > On Tue 01/08/2006 at 16:01, Sean Corfield wrote:
              >This should only affect servers where <cfdirectory> returns the
              last
              >modified timestamp in minutes rather than seconds...
              >
              >I used the exact same test that Fusebox 4 uses so folks should be
              >seeing the same behavior there too.

              I have never noticed this problem with Fusebox 4 which I have been using for
              some time. It leads me to believe that somthing else is happening here. I
              will have to spend more time trying it out and checking Fusebox 4 too.

              Kevin


            • Adam Reynolds
              Terry, Can you make a test where you change the XML logic inside an existing fuseaction and see if this is reflected. If you are changing a template that is
              Message 6 of 14 , Aug 3 9:25 AM
              • 0 Attachment
                Terry,
                Can you make a test where you change the XML logic inside an existing
                fuseaction and see if this is reflected. If you are changing a template
                that is executed then this will show, as you are not changing the
                fusebox app, just the template it is including.

                Adam


                Terry L Schmitt wrote:
                >
                >
                > I'm going to chime in again...
                > I brought this up a week or so ago, as I was having trouble with FB5
                > reloading the circuits when they change. I dropped the issue, because
                > I couldn't really capture some meaningful symptoms and a simple test
                > app reloaded the circuits just fine while in development-circuit-load.
                > Go figure.
                >
                > The app where I'm having the trouble is not huge, but does have some
                > complexity with a fair number of plugins and globalfuseactions firing.
                > FB5 seems to recognize changes made to an existing fuseaction, but
                > adding a new fuseaction to a circuit is not recognized. The
                > work-around is to simply reload manually using the URL parameters.
                >
                > I am using English Windows 2003 Server. The cfdirectory granularity is
                > to the second, so that should not be the concern in my case.
                >
                > I will keep this in mind as I develop, but unfortunately I am so
                > stacked up at work that I can't put serious time into researching this
                > more.
                >
                > Terry
                >
                >
                >
                >
                >
                > *"Kevin Roche" <kevin@...>*
                > Sent by: fusebox5@yahoogroups.com
                >
                > 08/01/2006 05:46 PM
                > Please respond to
                > fusebox5@yahoogroups.com
                >
                >
                >
                > To
                > <fusebox5@yahoogroups.com>
                > cc
                >
                > Subject
                > RE: [fusebox5] Problem with Load
                >
                >
                >
                >
                >
                >
                >
                >
                >
                >
                > > On Tue 01/08/2006 at 16:01, Sean Corfield wrote:
                > >This should only affect servers where <cfdirectory> returns the last
                > >modified timestamp in minutes rather than seconds...
                > >
                > >I used the exact same test that Fusebox 4 uses so folks should be
                > >seeing the same behavior there too.
                >
                > I have never noticed this problem with Fusebox 4 which I have been
                > using for
                > some time. It leads me to believe that somthing else is happening here. I
                > will have to spend more time trying it out and checking Fusebox 4 too.
                >
                > Kevin
                >
                >
                >
              • Daniel Schmid
                I can t see the reason to make it conditional? whats the problem with the java solution? It uses simple java classes which are available in jdk1.4 and
                Message 7 of 14 , Aug 4 10:58 AM
                • 0 Attachment
                  I can't see the reason to make it conditional? whats the problem with the java solution? It uses simple java classes which are available in jdk1.4 and before...

                  On my opinion it would be interesting to know what CF is doing by reading the timestamp from the files. Since in java you get seconds  I don't believe is OS or language dependent, I guess it depends maybe on the cf implementation for different OS?

                  Daniel

                  On 8/1/06, Barney Boisvert <bboisvert@...> wrote:
                  On 8/1/06, Sean Corfield <sean@...> wrote:
                  > I think a possible workaround for now would be to change the 'gt'
                  > operator to 'gte' in the following places:
                  > - fuseboxApplication.cfc line 163
                  > - fuseboxCircuit.cfc line 110
                  >
                  > This wouldn't be optimal because it would repeatedly reload files on
                  > some O/S for the first minute after a change but it would probably
                  > solve your problems for now...

                  Can you make that change conditional?  Do a test CFDIRECTORY, check
                  the granularity of the timestamps, and then use that info to do the
                  checks in the optimal way.  That'd also give you a fallback from using
                  the Java solution.

                  cheers,
                  barneyb
                  --
                  Barney Boisvert
                  bboisvert@...
                  360.319.6145
                  http://www.barneyb.com/

                  Got Gmail? I have 100 invites.



                  Yahoo! Groups Links

                  <*> To visit your group on the web, go to:
                      http://groups.yahoo.com/group/fusebox5/

                  <*> To unsubscribe from this group, send an email to:
                      fusebox5-unsubscribe@yahoogroups.com

                  <*> Your use of Yahoo! Groups is subject to:
                      http://docs.yahoo.com/info/terms/





                • Barney Boisvert
                  Not everyone has Java (BD.NET, for example), and there have been issues with certain Java operations on the Java-based BD as well. I ve no experience with
                  Message 8 of 14 , Aug 4 11:06 AM
                  • 0 Attachment
                    Not everyone has Java (BD.NET, for example), and there have been
                    issues with certain Java operations on the Java-based BD as well.
                    I've no experience with Ralio, but it's entirely possible a similar
                    problems could crop up. ColdFusion is not the single target platform,
                    CFML runtimes in general are. Therefore there's a advantage to using
                    only CFML constructs.

                    cheers,
                    barneyb

                    On 8/4/06, Daniel Schmid <daniel.schmid@...> wrote:
                    >
                    >
                    > I can't see the reason to make it conditional? whats the problem with the java solution? It uses simple java classes which are available in jdk1.4 and before...
                    >
                    > On my opinion it would be interesting to know what CF is doing by reading the timestamp from the files. Since in java you get seconds I don't believe is OS or language dependent, I guess it depends maybe on the cf implementation for different OS?
                    >
                    >
                    > Daniel
                    >
                    >
                    >
                    >
                    > On 8/1/06, Barney Boisvert <bboisvert@...> wrote:
                    >
                    > >
                    >
                    > On 8/1/06, Sean Corfield <sean@...> wrote:
                    > > I think a possible workaround for now would be to change the 'gt'
                    > > operator to 'gte' in the following places:
                    > > - fuseboxApplication.cfc line 163
                    > > - fuseboxCircuit.cfc line 110
                    > >
                    > > This wouldn't be optimal because it would repeatedly reload files on
                    > > some O/S for the first minute after a change but it would probably
                    > > solve your problems for now...
                    >
                    > Can you make that change conditional? Do a test CFDIRECTORY, check
                    > the granularity of the timestamps, and then use that info to do the
                    > checks in the optimal way. That'd also give you a fallback from using
                    > the Java solution.
                    >
                    > cheers,
                    > barneyb
                    > --
                    > Barney Boisvert
                    > bboisvert@...
                    > 360.319.6145
                    > http://www.barneyb.com/
                    >
                    > Got Gmail? I have 100 invites.
                    >
                    >


                    --
                    Barney Boisvert
                    bboisvert@...
                    360.319.6145
                    http://www.barneyb.com/

                    Got Gmail? I have 100 invites.
                  Your message has been successfully submitted and would be delivered to recipients shortly.