You are viewing a plain text version of this content. The canonical link for it is here.
Posted to user@ant.apache.org by Eric Fetzer <el...@yahoo.com> on 2009/05/29 17:45:37 UTC

Use of Sleep

I've seen a lot of folks following <exec> tasks with sleep.  Doesn't the exec task wait for a return code prior to returning to the caller?  I inherited A BUNCH of Ant build files that were just RIDDLED with 5 minutes of sleep here 10 minutes of sleep there...  The build takes half of a day to finish when all is said and done.  To me, that is rediculous.  I come from a dot net/NAnt environment where my longest build took 20 minutes and I wasn't happy with that one.


      

Re: Use of Sleep

Posted by Dale Anson <da...@grafidog.com>.
You might want to check out the AntPerformanceListener in either Ant-Contrib
(http://ant-contrib.sourceforge.net/) or Antelope (
http://antelope.tigris.org/nonav/docs/manual/bk03ch27.html).  This is a
listener that you add on the command line when you start Ant, and it will
give you a summary of where your build is spending all its time.  We had a
build that was taking anywhere from 15 minutes to an hour on dev boxes, this
listener helped me trim that down to a 5 second build.

Dale



On Mon, Jun 1, 2009 at 8:39 AM, Eric Fetzer <el...@yahoo.com> wrote:

> A good piece of the time, David, is allocated to check out of A LOT of
> code.  It also ftp's some very large files over a WAN.  As I take over these
> builds, I will be digging further in where all the time is going, but for
> now, I'm just trying to get rid of totally unnecessary aspects of the build.
>
>
>
>
> ________________________________
> From: David Weintraub <qa...@gmail.com>
> To: Ant Users List <us...@ant.apache.org>
> Sent: Saturday, May 30, 2009 10:40:37 PM
> Subject: Re: Use of Sleep
>
> What do all those <exec> tasks do? Are they really part of the build?
>
> We had a build where we brought up a Websphere server in order to run
> our JUnit tests, but that was about 2 minutes of sleep to ensure that
> the server was awake before we ran the task.
>
> But, the entire build sans tests took about 2 to 4 minutes. In my
> current work, most of our builds take between 1 to 3 minutes. One
> project (which everyone admits is an absolute mess) takes about 9
> minutes to build.
>
> There's no reason for a build to take longer than 15 minutes. After
> that, you can't use it for continuous build testing because the builds
> take too long.
>
> On Fri, May 29, 2009 at 11:45 AM, Eric Fetzer <el...@yahoo.com>
> wrote:
> > I've seen a lot of folks following <exec> tasks with sleep.  Doesn't the
> exec task wait for a return code prior to returning to the caller?  I
> inherited A BUNCH of Ant build files that were just RIDDLED with 5 minutes
> of sleep here 10 minutes of sleep there...  The build takes half of a day to
> finish when all is said and done.  To me, that is rediculous.  I come from a
> dot net/NAnt environment where my longest build took 20 minutes and I wasn't
> happy with that one.
> >
> >
> >
>
>
>
> --
> David Weintraub
> qazwart@gmail.com
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@ant.apache.org
> For additional commands, e-mail: user-help@ant.apache.org
>
>
>
>

Re: Use of Sleep

Posted by Eric Fetzer <el...@yahoo.com>.
Bill Of Materials - All of the items that make up a build (including the items necessary to perform the build).





________________________________
From: Martin Gainty <mg...@hotmail.com>
To: Ant Users List <us...@ant.apache.org>
Sent: Monday, June 1, 2009 12:05:30 PM
Subject: RE: Use of Sleep


the closest I've seen for Windows to daemon or & is START /BELOWNORMAL option:
START ["title"] [/Dpath] [/I] [/MIN] [/MAX] [/SEPARATE | /SHARED]
      [/LOW | /NORMAL | /HIGH | /REALTIME | /ABOVENORMAL | /BELOWNORMAL]
      [/WAIT] [/B] [command/program]
      [parameters]

    "title"    Title to display in  window title bar.
    path        Starting directory
    B          Start application without creating a new window. The
                application has ^C handling ignored. Unless the application
                enables ^C processing, ^Break is the only way to interrupt
                the application
    I          The new environment will be the original environment passed
                to the cmd.exe and not the current environment.
    MIN        Start window minimized
    MAX        Start window maximized
    SEPARATE    Start 16-bit Windows program in separate memory space
    SHARED      Start 16-bit Windows program in shared memory space
    LOW        Start application in the IDLE priority class
    NORMAL      Start application in the NORMAL priority class
    HIGH        Start application in the HIGH priority class
    REALTIME    Start application in the REALTIME priority class
    ABOVENORMAL Start application in the ABOVENORMAL priority class
    BELOWNORMAL Start application in the BELOWNORMAL priority class
    WAIT        Start application and wait for it to terminate

could you specify the definition of BOM?
Bill of Ordered Materials
Byte Order Mark
?
thanks
Martin 
______________________________________________ 
Verzicht und Vertraulichkeitanmerkung/Note de déni et de confidentialité

Diese Nachricht ist vertraulich. Sollten Sie nicht der vorgesehene Empfaenger sein, so bitten wir hoeflich um eine Mitteilung. Jede unbefugte Weiterleitung oder Fertigung einer Kopie ist unzulaessig. Diese Nachricht dient lediglich dem Austausch von Informationen und entfaltet keine rechtliche Bindungswirkung. Aufgrund der leichten Manipulierbarkeit von E-Mails koennen wir keine Haftung fuer den Inhalt uebernehmen.
Ce message est confidentiel et peut être privilégié. Si vous n'êtes pas le destinataire prévu, nous te demandons avec bonté que pour satisfaire informez l'expéditeur. N'importe quelle diffusion non autorisée ou la copie de ceci est interdite. Ce message sert à l'information seulement et n'aura pas n'importe quel effet légalement obligatoire. Étant donné que les email peuvent facilement être sujets à la manipulation, nous ne pouvons accepter aucune responsabilité pour le contenu fourni.




> Date: Mon, 1 Jun 2009 09:48:53 -0700
> From: elstonkers@yahoo.com
> Subject: Re: Use of Sleep
> To: user@ant.apache.org
> 
> Will "updates" get rid of files that got deleted from the repository?  I don't like not having an accurate BOM.  I want to know every file and what version of that file makes up my build.  When I remove the directories and re-check my files out every time, I'm afforded the abiility to keep an accurate BOM and know that the BOM represents the product exactly.  Otherwise, what confidence can I lend the customer that I know what I delivered to them?
> 
> 
> 
> ________________________________
> From: David Weintraub <qa...@gmail.com>
> To: Ant Users List <us...@ant.apache.org>
> Sent: Monday, June 1, 2009 9:19:16 AM
> Subject: Re: Use of Sleep
> 
> On Mon, Jun 1, 2009 at 10:39 AM, Eric Fetzer <el...@yahoo.com> wrote:
> 
> > A good piece of the time, David, is allocated to check out of A LOT of
> > code.  It also ftp's some very large files over a WAN.  As I take over these
> > builds, I will be digging further in where all the time is going, but for
> > now, I'm just trying to get rid of totally unnecessary aspects of the build.
> 
> 
> I normally don't have Ant do my checkouts since you first have to checkout
> the build.xml file to do the initial checkout. We use Hudson for our
> continuous integration, and Hudson handles the checkouts for us. That way, I
> don't even have to define a checkout task in our build.xml file. (See <
> http://hudson.dev.java.net/>). You might want to switch from checkouts to
> updates since updates only update changed files, and don't checkout
> everything from scratch again.
> 
> There are specific Ant checkout tasks for various version control systems,
> but I usually don't find them any faster than <exec>. Unless the <exec> task
> runs in the background (spawn="true"), it should wait for the <exec> task to
> complete. That means you shouldn't be having <sleep> tasks to wait for the
> checkout to complete.
> 
> You normally run an <exec> task in the background when you're doing things
> like starting up a server, so you can run your tests. Otherwise, the default
> is to wait for the <exec> command to complete. BTW, there is a "timeout"
> parameter for the <exec> task that can help kill an otherwise lollying
> program, but it has a bit of trouble on Windows sytems.
> 
> Best of luck speeding up your build. When I first got to my current job,
> builds took 30+ minutes to complete. I rewrote the build.xml file from
> scratch, and speeded it up to only "12 minutes". Meanwhile, we've pulled
> projects out of our main build, and each one of these build in under 2 or 3
> minutes. The "big" project still takes about 9 minutes to build. Not where
> I'd like it, but much better than it was before.
> 
> My experience shows that when builds take forever, developers simply don't
> test as well, and your code gets sloppier.
> 
> -- 
> David Weintraub
> qazwart@gmail.com
> 
> 
> 
>      

_________________________________________________________________
Hotmail® has a new way to see what's up with your friends.
http://windowslive.com/Tutorial/Hotmail/WhatsNew?ocid=TXT_TAGLM_WL_HM_Tutorial_WhatsNew1_052009


      

RE: Use of Sleep

Posted by Martin Gainty <mg...@hotmail.com>.
the closest I've seen for Windows to daemon or & is START /BELOWNORMAL option:
START ["title"] [/Dpath] [/I] [/MIN] [/MAX] [/SEPARATE | /SHARED]
      [/LOW | /NORMAL | /HIGH | /REALTIME | /ABOVENORMAL | /BELOWNORMAL]
      [/WAIT] [/B] [command/program]
      [parameters]

    "title"     Title to display in  window title bar.
    path        Starting directory
    B           Start application without creating a new window. The
                application has ^C handling ignored. Unless the application
                enables ^C processing, ^Break is the only way to interrupt
                the application
    I           The new environment will be the original environment passed
                to the cmd.exe and not the current environment.
    MIN         Start window minimized
    MAX         Start window maximized
    SEPARATE    Start 16-bit Windows program in separate memory space
    SHARED      Start 16-bit Windows program in shared memory space
    LOW         Start application in the IDLE priority class
    NORMAL      Start application in the NORMAL priority class
    HIGH        Start application in the HIGH priority class
    REALTIME    Start application in the REALTIME priority class
    ABOVENORMAL Start application in the ABOVENORMAL priority class
    BELOWNORMAL Start application in the BELOWNORMAL priority class
    WAIT        Start application and wait for it to terminate

could you specify the definition of BOM?
Bill of Ordered Materials
Byte Order Mark
?
thanks
Martin 
______________________________________________ 
Verzicht und Vertraulichkeitanmerkung/Note de déni et de confidentialité
 
Diese Nachricht ist vertraulich. Sollten Sie nicht der vorgesehene Empfaenger sein, so bitten wir hoeflich um eine Mitteilung. Jede unbefugte Weiterleitung oder Fertigung einer Kopie ist unzulaessig. Diese Nachricht dient lediglich dem Austausch von Informationen und entfaltet keine rechtliche Bindungswirkung. Aufgrund der leichten Manipulierbarkeit von E-Mails koennen wir keine Haftung fuer den Inhalt uebernehmen.
Ce message est confidentiel et peut être privilégié. Si vous n'êtes pas le destinataire prévu, nous te demandons avec bonté que pour satisfaire informez l'expéditeur. N'importe quelle diffusion non autorisée ou la copie de ceci est interdite. Ce message sert à l'information seulement et n'aura pas n'importe quel effet légalement obligatoire. Étant donné que les email peuvent facilement être sujets à la manipulation, nous ne pouvons accepter aucune responsabilité pour le contenu fourni.




> Date: Mon, 1 Jun 2009 09:48:53 -0700
> From: elstonkers@yahoo.com
> Subject: Re: Use of Sleep
> To: user@ant.apache.org
> 
> Will "updates" get rid of files that got deleted from the repository?  I don't like not having an accurate BOM.  I want to know every file and what version of that file makes up my build.  When I remove the directories and re-check my files out every time, I'm afforded the abiility to keep an accurate BOM and know that the BOM represents the product exactly.  Otherwise, what confidence can I lend the customer that I know what I delivered to them?
> 
> 
> 
> ________________________________
> From: David Weintraub <qa...@gmail.com>
> To: Ant Users List <us...@ant.apache.org>
> Sent: Monday, June 1, 2009 9:19:16 AM
> Subject: Re: Use of Sleep
> 
> On Mon, Jun 1, 2009 at 10:39 AM, Eric Fetzer <el...@yahoo.com> wrote:
> 
> > A good piece of the time, David, is allocated to check out of A LOT of
> > code.  It also ftp's some very large files over a WAN.  As I take over these
> > builds, I will be digging further in where all the time is going, but for
> > now, I'm just trying to get rid of totally unnecessary aspects of the build.
> 
> 
> I normally don't have Ant do my checkouts since you first have to checkout
> the build.xml file to do the initial checkout. We use Hudson for our
> continuous integration, and Hudson handles the checkouts for us. That way, I
> don't even have to define a checkout task in our build.xml file. (See <
> http://hudson.dev.java.net/>). You might want to switch from checkouts to
> updates since updates only update changed files, and don't checkout
> everything from scratch again.
> 
> There are specific Ant checkout tasks for various version control systems,
> but I usually don't find them any faster than <exec>. Unless the <exec> task
> runs in the background (spawn="true"), it should wait for the <exec> task to
> complete. That means you shouldn't be having <sleep> tasks to wait for the
> checkout to complete.
> 
> You normally run an <exec> task in the background when you're doing things
> like starting up a server, so you can run your tests. Otherwise, the default
> is to wait for the <exec> command to complete. BTW, there is a "timeout"
> parameter for the <exec> task that can help kill an otherwise lollying
> program, but it has a bit of trouble on Windows sytems.
> 
> Best of luck speeding up your build. When I first got to my current job,
> builds took 30+ minutes to complete. I rewrote the build.xml file from
> scratch, and speeded it up to only "12 minutes". Meanwhile, we've pulled
> projects out of our main build, and each one of these build in under 2 or 3
> minutes. The "big" project still takes about 9 minutes to build. Not where
> I'd like it, but much better than it was before.
> 
> My experience shows that when builds take forever, developers simply don't
> test as well, and your code gets sloppier.
> 
> -- 
> David Weintraub
> qazwart@gmail.com
> 
> 
> 
>       

_________________________________________________________________
Hotmail® has a new way to see what's up with your friends.
http://windowslive.com/Tutorial/Hotmail/WhatsNew?ocid=TXT_TAGLM_WL_HM_Tutorial_WhatsNew1_052009

Re: Use of Sleep

Posted by David Weintraub <qa...@gmail.com>.
On Mon, Jun 1, 2009 at 3:42 PM, Eric Fetzer <el...@yahoo.com> wrote:

> This is the weekly build I'm referring to rather than the CC builds.  I
> really need to be sure it's clean.  If one file gets removed from the code
> repository (*StarTeam *for us),


Oh, StarTeam...

That explains everything. Yeah, you can't do just an update in StarTeam and
know that your build artifacts are in order. StarTeam does not remove
deleted files or directories -- at least on a consistent basis.

I've used Perforce, Subversion, and even CVS, and haven't had a problem with
using an "update" vs. a "checkout" as long as I follow a few basic rules
(don't modify files stored in the revision control system during the build
process, do a good clean before a build, and try not to mix your built
objects from your source code). However, StarTeam gets me every time.

-- 
David Weintraub
qazwart@gmail.com

Re: Use of Sleep

Posted by Eric Fetzer <el...@yahoo.com>.
This is the weekly build I'm referring to rather than the CC builds.  I really need to be sure it's clean.  If one file gets removed from the code repository (StarTeam for us), but it doesn't get cleaned out of the build directory, the repercussions could be humongous.  Maybe the file got deleted, but references to it in the application didn't get removed.  The build continues to work because we aren't doing a clean.  6 months down the line when we finally do a clean, the build stops working but the file is no where to be found.  I've been bit like this before and changed my evil ways ;0)




________________________________
From: David Weintraub <qa...@gmail.com>
To: Ant Users List <us...@ant.apache.org>
Sent: Monday, June 1, 2009 12:46:19 PM
Subject: Re: Use of Sleep

On Mon, Jun 1, 2009 at 12:48 PM, Eric Fetzer <el...@yahoo.com> wrote:

> Will "updates" get rid of files that got deleted from the repository?


It all really depends upon the revision control system you're using. If the
revision control system tracks directories (like Subversion and ClearCase),
the answer is "yes", it will remove directories when they have been deleted.
If you use something like CVS, there's a special parameter on the "update"
command that removes empty directories. A few revision control systems might
not.

I find that if you do a few things, you'll be good about 99% of the time:
Don't update source files. Try to keep your built object files (like
classfiles) and your source files in separate directories. And, have a good
"clean" target. I usually do a clean before each build to eliminate rebuild
issues.

Where things mess up is when there's a major revamping of the whole project
structure where directories and files are  renamed, deleted, and moved. In
those cases, I am usually forced to do a checkout instead of an update.

Of course, the only way to be really, really sure is to do a checkout, but
if checkouts take hours, and your build process drags out, you'll end up
doing more harm than good.

A good compromise is to do an update on nomal continuous test builds for
speed, and when you actually build the release, use a clean checkout. That
will give you the development speed, but make sure your customers are
getting exactly everything in the build.

-- 
David Weintraub
qazwart@gmail.com



      

Re: Use of Sleep

Posted by David Weintraub <qa...@gmail.com>.
On Mon, Jun 1, 2009 at 12:48 PM, Eric Fetzer <el...@yahoo.com> wrote:

> Will "updates" get rid of files that got deleted from the repository?


It all really depends upon the revision control system you're using. If the
revision control system tracks directories (like Subversion and ClearCase),
the answer is "yes", it will remove directories when they have been deleted.
If you use something like CVS, there's a special parameter on the "update"
command that removes empty directories. A few revision control systems might
not.

I find that if you do a few things, you'll be good about 99% of the time:
Don't update source files. Try to keep your built object files (like
classfiles) and your source files in separate directories. And, have a good
"clean" target. I usually do a clean before each build to eliminate rebuild
issues.

Where things mess up is when there's a major revamping of the whole project
structure where directories and files are  renamed, deleted, and moved. In
those cases, I am usually forced to do a checkout instead of an update.

Of course, the only way to be really, really sure is to do a checkout, but
if checkouts take hours, and your build process drags out, you'll end up
doing more harm than good.

A good compromise is to do an update on nomal continuous test builds for
speed, and when you actually build the release, use a clean checkout. That
will give you the development speed, but make sure your customers are
getting exactly everything in the build.

-- 
David Weintraub
qazwart@gmail.com

Re: Use of Sleep

Posted by Eric Fetzer <el...@yahoo.com>.
Will "updates" get rid of files that got deleted from the repository?  I don't like not having an accurate BOM.  I want to know every file and what version of that file makes up my build.  When I remove the directories and re-check my files out every time, I'm afforded the abiility to keep an accurate BOM and know that the BOM represents the product exactly.  Otherwise, what confidence can I lend the customer that I know what I delivered to them?



________________________________
From: David Weintraub <qa...@gmail.com>
To: Ant Users List <us...@ant.apache.org>
Sent: Monday, June 1, 2009 9:19:16 AM
Subject: Re: Use of Sleep

On Mon, Jun 1, 2009 at 10:39 AM, Eric Fetzer <el...@yahoo.com> wrote:

> A good piece of the time, David, is allocated to check out of A LOT of
> code.  It also ftp's some very large files over a WAN.  As I take over these
> builds, I will be digging further in where all the time is going, but for
> now, I'm just trying to get rid of totally unnecessary aspects of the build.


I normally don't have Ant do my checkouts since you first have to checkout
the build.xml file to do the initial checkout. We use Hudson for our
continuous integration, and Hudson handles the checkouts for us. That way, I
don't even have to define a checkout task in our build.xml file. (See <
http://hudson.dev.java.net/>). You might want to switch from checkouts to
updates since updates only update changed files, and don't checkout
everything from scratch again.

There are specific Ant checkout tasks for various version control systems,
but I usually don't find them any faster than <exec>. Unless the <exec> task
runs in the background (spawn="true"), it should wait for the <exec> task to
complete. That means you shouldn't be having <sleep> tasks to wait for the
checkout to complete.

You normally run an <exec> task in the background when you're doing things
like starting up a server, so you can run your tests. Otherwise, the default
is to wait for the <exec> command to complete. BTW, there is a "timeout"
parameter for the <exec> task that can help kill an otherwise lollying
program, but it has a bit of trouble on Windows sytems.

Best of luck speeding up your build. When I first got to my current job,
builds took 30+ minutes to complete. I rewrote the build.xml file from
scratch, and speeded it up to only "12 minutes". Meanwhile, we've pulled
projects out of our main build, and each one of these build in under 2 or 3
minutes. The "big" project still takes about 9 minutes to build. Not where
I'd like it, but much better than it was before.

My experience shows that when builds take forever, developers simply don't
test as well, and your code gets sloppier.

-- 
David Weintraub
qazwart@gmail.com



      

Re: Use of Sleep

Posted by David Weintraub <qa...@gmail.com>.
On Mon, Jun 1, 2009 at 10:39 AM, Eric Fetzer <el...@yahoo.com> wrote:

> A good piece of the time, David, is allocated to check out of A LOT of
> code.  It also ftp's some very large files over a WAN.  As I take over these
> builds, I will be digging further in where all the time is going, but for
> now, I'm just trying to get rid of totally unnecessary aspects of the build.


I normally don't have Ant do my checkouts since you first have to checkout
the build.xml file to do the initial checkout. We use Hudson for our
continuous integration, and Hudson handles the checkouts for us. That way, I
don't even have to define a checkout task in our build.xml file. (See <
http://hudson.dev.java.net/>). You might want to switch from checkouts to
updates since updates only update changed files, and don't checkout
everything from scratch again.

There are specific Ant checkout tasks for various version control systems,
but I usually don't find them any faster than <exec>. Unless the <exec> task
runs in the background (spawn="true"), it should wait for the <exec> task to
complete. That means you shouldn't be having <sleep> tasks to wait for the
checkout to complete.

You normally run an <exec> task in the background when you're doing things
like starting up a server, so you can run your tests. Otherwise, the default
is to wait for the <exec> command to complete. BTW, there is a "timeout"
parameter for the <exec> task that can help kill an otherwise lollying
program, but it has a bit of trouble on Windows sytems.

Best of luck speeding up your build. When I first got to my current job,
builds took 30+ minutes to complete. I rewrote the build.xml file from
scratch, and speeded it up to only "12 minutes". Meanwhile, we've pulled
projects out of our main build, and each one of these build in under 2 or 3
minutes. The "big" project still takes about 9 minutes to build. Not where
I'd like it, but much better than it was before.

My experience shows that when builds take forever, developers simply don't
test as well, and your code gets sloppier.

-- 
David Weintraub
qazwart@gmail.com

Re: Use of Sleep

Posted by Eric Fetzer <el...@yahoo.com>.
A good piece of the time, David, is allocated to check out of A LOT of code.  It also ftp's some very large files over a WAN.  As I take over these builds, I will be digging further in where all the time is going, but for now, I'm just trying to get rid of totally unnecessary aspects of the build.




________________________________
From: David Weintraub <qa...@gmail.com>
To: Ant Users List <us...@ant.apache.org>
Sent: Saturday, May 30, 2009 10:40:37 PM
Subject: Re: Use of Sleep

What do all those <exec> tasks do? Are they really part of the build?

We had a build where we brought up a Websphere server in order to run
our JUnit tests, but that was about 2 minutes of sleep to ensure that
the server was awake before we ran the task.

But, the entire build sans tests took about 2 to 4 minutes. In my
current work, most of our builds take between 1 to 3 minutes. One
project (which everyone admits is an absolute mess) takes about 9
minutes to build.

There's no reason for a build to take longer than 15 minutes. After
that, you can't use it for continuous build testing because the builds
take too long.

On Fri, May 29, 2009 at 11:45 AM, Eric Fetzer <el...@yahoo.com> wrote:
> I've seen a lot of folks following <exec> tasks with sleep.  Doesn't the exec task wait for a return code prior to returning to the caller?  I inherited A BUNCH of Ant build files that were just RIDDLED with 5 minutes of sleep here 10 minutes of sleep there...  The build takes half of a day to finish when all is said and done.  To me, that is rediculous.  I come from a dot net/NAnt environment where my longest build took 20 minutes and I wasn't happy with that one.
>
>
>



-- 
David Weintraub
qazwart@gmail.com

---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@ant.apache.org
For additional commands, e-mail: user-help@ant.apache.org


      

RE: Use of Sleep

Posted by Martin Gainty <mg...@hotmail.com>.
Good Morning David-

cruisecontrol will help you with
scheduling
handle dependencies

http://confluence.public.thoughtworks.org/display/CC/CruiseControl+Scheduling+Scenarios

Martin Gainty 
______________________________________________ 
Verzicht und Vertraulichkeitanmerkung/Note de déni et de confidentialité
 
Diese Nachricht ist vertraulich. Sollten Sie nicht der vorgesehene Empfaenger sein, so bitten wir hoeflich um eine Mitteilung. Jede unbefugte Weiterleitung oder Fertigung einer Kopie ist unzulaessig. Diese Nachricht dient lediglich dem Austausch von Informationen und entfaltet keine rechtliche Bindungswirkung. Aufgrund der leichten Manipulierbarkeit von E-Mails koennen wir keine Haftung fuer den Inhalt uebernehmen.
Ce message est confidentiel et peut être privilégié. Si vous n'êtes pas le destinataire prévu, nous te demandons avec bonté que pour satisfaire informez l'expéditeur. N'importe quelle diffusion non autorisée ou la copie de ceci est interdite. Ce message sert à l'information seulement et n'aura pas n'importe quel effet légalement obligatoire. Étant donné que les email peuvent facilement être sujets à la manipulation, nous ne pouvons accepter aucune responsabilité pour le contenu fourni.




> Date: Sun, 31 May 2009 00:40:37 -0400
> Subject: Re: Use of Sleep
> From: qazwart@gmail.com
> To: user@ant.apache.org
> 
> What do all those <exec> tasks do? Are they really part of the build?
> 
> We had a build where we brought up a Websphere server in order to run
> our JUnit tests, but that was about 2 minutes of sleep to ensure that
> the server was awake before we ran the task.
> 
> But, the entire build sans tests took about 2 to 4 minutes. In my
> current work, most of our builds take between 1 to 3 minutes. One
> project (which everyone admits is an absolute mess) takes about 9
> minutes to build.
> 
> There's no reason for a build to take longer than 15 minutes. After
> that, you can't use it for continuous build testing because the builds
> take too long.
> 
> On Fri, May 29, 2009 at 11:45 AM, Eric Fetzer <el...@yahoo.com> wrote:
> > I've seen a lot of folks following <exec> tasks with sleep.  Doesn't the exec task wait for a return code prior to returning to the caller?  I inherited A BUNCH of Ant build files that were just RIDDLED with 5 minutes of sleep here 10 minutes of sleep there...  The build takes half of a day to finish when all is said and done.  To me, that is rediculous.  I come from a dot net/NAnt environment where my longest build took 20 minutes and I wasn't happy with that one.
> >
> >
> >
> 
> 
> 
> -- 
> David Weintraub
> qazwart@gmail.com
> 
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: user-unsubscribe@ant.apache.org
> For additional commands, e-mail: user-help@ant.apache.org
> 

_________________________________________________________________
Hotmail® has ever-growing storage! Don’t worry about storage limits.
http://windowslive.com/Tutorial/Hotmail/Storage?ocid=TXT_TAGLM_WL_HM_Tutorial_Storage1_052009

Re: Use of Sleep

Posted by David Weintraub <qa...@gmail.com>.
What do all those <exec> tasks do? Are they really part of the build?

We had a build where we brought up a Websphere server in order to run
our JUnit tests, but that was about 2 minutes of sleep to ensure that
the server was awake before we ran the task.

But, the entire build sans tests took about 2 to 4 minutes. In my
current work, most of our builds take between 1 to 3 minutes. One
project (which everyone admits is an absolute mess) takes about 9
minutes to build.

There's no reason for a build to take longer than 15 minutes. After
that, you can't use it for continuous build testing because the builds
take too long.

On Fri, May 29, 2009 at 11:45 AM, Eric Fetzer <el...@yahoo.com> wrote:
> I've seen a lot of folks following <exec> tasks with sleep.  Doesn't the exec task wait for a return code prior to returning to the caller?  I inherited A BUNCH of Ant build files that were just RIDDLED with 5 minutes of sleep here 10 minutes of sleep there...  The build takes half of a day to finish when all is said and done.  To me, that is rediculous.  I come from a dot net/NAnt environment where my longest build took 20 minutes and I wasn't happy with that one.
>
>
>



-- 
David Weintraub
qazwart@gmail.com

---------------------------------------------------------------------
To unsubscribe, e-mail: user-unsubscribe@ant.apache.org
For additional commands, e-mail: user-help@ant.apache.org