You are viewing a plain text version of this content. The canonical link for it is here.
Posted to general@gump.apache.org by Ceki Gülcü <ce...@qos.ch> on 2004/12/14 15:57:24 UTC
Re: cvs commit: logging-log4j/src/java/org/apache/log4j/net
SMTPAppender.java
Niclas et al.,
IMHO, the kind of urgent pleas as the one by Niclas (quoted below) are
quite inappropriate. We, the log4j developers, have the right to
occasionally fuck up like every one else. The fact that a small typo
affects 200 projects should tell you that something is wrong with
Gump's current approach.
When a totally silly problem attains such cataclysmic proportions, it
puts us (log4j developers) under unnecessary and useless
pressure. Please revise your model so that a silly mistake can be
sidestepped without affecting 200 projects. For example, just alerting
log4j-dev and having the 200 affected projects to use yesterday's
version of log4j would have been much better.
I am pleased to see that the problem on our side was corrected
promptly. I am much less so with nature of the social interaction that
led to it.
At 07:37 AM 12/14/2004, Niclas Hedhman wrote:
>May I suggest a revert of the change first, and then think about how it
>should
>be done?? :o) (only 200 projects are affected...)
>
>Cheers
>Niclas
>--
--
Ceki Gülcü
The complete log4j manual:
https://www.qos.ch/shop/products/log4j/log4j-Manual.jsp
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@gump.apache.org
For additional commands, e-mail: general-help@gump.apache.org
RE: cvs commit: logging-log4j/src/java/org/apache/log4j/net
SMTPAppender.java
Posted by Ceki Gülcü <ce...@qos.ch>.
At 07:41 PM 12/14/2004, Mark Womack wrote:
>The fact that 200 projects were affected in Gump is a Gump problem, not a
>log4j problem.
Gump is reportedly being improved to avoid the recent observed massive
breakdown. However, when 200 projects start blaming log4j, then things
become much less obvious.
The point I am trying to make is that when projects like log4j, with
many dependents, have compilation issues, Gump currently raises way
too many alarms. Again, this is being reportedly worked upon.
If you go back and read Niclas' first message, you should notice the
urgent tone, where, as I am trying to explain, there should have been
no urgency.
>I'm more concerned that there might be something up with the main log4j
>build script. Yoav did a recompile before submitting. And Jake had that
>strange compile behavior recently too.
Are you going to check?
>-Mark
--
Ceki Gülcü
The complete log4j manual:
https://www.qos.ch/shop/products/log4j/log4j-Manual.jsp
---------------------------------------------------------------------
To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
For additional commands, e-mail: log4j-dev-help@logging.apache.org
Re: cvs commit: logging-log4j/src/java/org/apache/log4j/net
SMTPAppender.java
Posted by Ceki Gülcü <ce...@qos.ch>.
At 04:48 PM 12/14/2004, Niclas Hedhman wrote:
>On Tuesday 14 December 2004 22:57, Ceki Gülcü wrote:
>
>My sincerest apologies. Of course, Log4J are entitled to "fuck up" from time
>to time. That is not something I want to undermine, and if it came across
>like that, then I am truly sorry.
Niclas, there is no need to apologize. No offense was taken and my
remarks were not directed at you personally.
>This is actually what the current implementation tries to do, but it has not
>been solved properly. But Stefano is now on a mission to re-do a lot of
>Gump's behaviour, and hopefully there will be improvements in both this area,
>as well as the notification system.
Glad to hear it. You can view my previous post as a request for the
type of enhancement worked on by Stephano.
>Looking back at my posts, I have a hard time seeing them as coming on hard.
>The first post was a plain "FYI" in case you guys haven't noticed.
I know. Again, my remarks were not directed at you personally but at
the way in which a small mistake can be perceived by the community as
something as important as a nuclear meltdown.
>I would like to know, if it is some old grudges you have against me, or how
>humble does one have to be, not to offend you, Ceki... I am confused.
It's not a question of politeness or humility but of alarm. A typo in
a log4j class should be a non event, not a cause for 200 projects to
fail.
Old grudges? What old grudges? I don't know what you are talking
about. :-)
>Cheers
>Niclas
>--
--
Ceki Gülcü
The complete log4j manual:
https://www.qos.ch/shop/products/log4j/log4j-Manual.jsp
---------------------------------------------------------------------
To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
For additional commands, e-mail: log4j-dev-help@logging.apache.org
Re: cvs commit: logging-log4j/src/java/org/apache/log4j/net SMTPAppender.java
Posted by Niclas Hedhman <ni...@hedhman.org>.
On Wednesday 15 December 2004 02:41, Mark Womack wrote:
> If there are that many projects in there, all dependent on
> each other in a multitude of ways, it amazes me that Gump ever completes a
> nightly build of all the projects. I like Gump, I support it, but I am not
> going to take special action to make sure everything is go for Gump every
> night.
Noone really expects that to happen. We have been hoovering around 90-95% for
a couple of weeks now, and only occassionally does a 'fundamental' project
fail like yesterday. But hey, that happens, we are just happy if it is
resolved quickly.
This level of success rate for Gump is a testamony to the pillar projects of
the ASF. Log4J is up there, but far from the heaviest dependencies. Most
failures happens in the smaller projects, which is reassurring.
Ceki, keep up the good work. Gump Meisters will attend to the problems of
Gump. But we will continue to 'inform' you guys when builds breaks. It's a
service, not a police.
Cheers
Niclas
--
+------//-------------------+
/ http://www.dpml.net /
/ http://niclas.hedhman.org /
+------//-------------------+
---------------------------------------------------------------------
To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
For additional commands, e-mail: log4j-dev-help@logging.apache.org
RE: cvs commit: logging-log4j/src/java/org/apache/log4j/net
SMTPAppender.java
Posted by Mark Womack <wo...@adobe.com>.
I agree with Niclas. Unless there is something else going on here, his
suggestion was appropriate. I just did not take any action because I
figured Yoav would fix it the right way when he had a chance. It didn't
look like there were any other dependencies, but hey, I haven't been keeping
track of what everyone is doing. I was in the middle of something else I
wanted to do for the project.
The fact that 200 projects were affected in Gump is a Gump problem, not a
log4j problem. If there are that many projects in there, all dependent on
each other in a multitude of ways, it amazes me that Gump ever completes a
nightly build of all the projects. I like Gump, I support it, but I am not
going to take special action to make sure everything is go for Gump every
night.
I'm more concerned that there might be something up with the main log4j
build script. Yoav did a recompile before submitting. And Jake had that
strange compile behavior recently too.
-Mark
-----Original Message-----
From: Niclas Hedhman [mailto:niclas@hedhman.org]
Sent: Tuesday, December 14, 2004 7:48 AM
To: Log4J Developers List
Subject: Re: cvs commit: logging-log4j/src/java/org/apache/log4j/net
SMTPAppender.java
On Tuesday 14 December 2004 22:57, Ceki Gülcü wrote:
> When a totally silly problem attains such cataclysmic proportions, it
> puts us (log4j developers) under unnecessary and useless
> pressure.
My sincerest apologies. Of course, Log4J are entitled to "fuck up" from time
to time. That is not something I want to undermine, and if it came across
like that, then I am truly sorry.
> Please revise your model so that a silly mistake can be
> sidestepped without affecting 200 projects. For example, just alerting
> log4j-dev and having the 200 affected projects to use yesterday's
> version of log4j would have been much better.
This is actually what the current implementation tries to do, but it has not
been solved properly. But Stefano is now on a mission to re-do a lot of
Gump's behaviour, and hopefully there will be improvements in both this
area,
as well as the notification system.
> I am pleased to see that the problem on our side was corrected
> promptly. I am much less so with nature of the social interaction that
> led to it.
Looking back at my posts, I have a hard time seeing them as coming on hard.
The first post was a plain "FYI" in case you guys haven't noticed.
The second was a response to Mark noticing it as well, where I have both a
"May I suggest..." and a smiley...
I would like to know, if it is some old grudges you have against me, or how
humble does one have to be, not to offend you, Ceki... I am confused.
Cheers
Niclas
--
+------//-------------------+
/ http://www.dpml.net /
/ http://niclas.hedhman.org /
+------//-------------------+
---------------------------------------------------------------------
To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
For additional commands, e-mail: log4j-dev-help@logging.apache.org
---------------------------------------------------------------------
To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
For additional commands, e-mail: log4j-dev-help@logging.apache.org
Re: cvs commit: logging-log4j/src/java/org/apache/log4j/net SMTPAppender.java
Posted by Niclas Hedhman <ni...@hedhman.org>.
On Tuesday 14 December 2004 22:57, Ceki Gülcü wrote:
> When a totally silly problem attains such cataclysmic proportions, it
> puts us (log4j developers) under unnecessary and useless
> pressure.
My sincerest apologies. Of course, Log4J are entitled to "fuck up" from time
to time. That is not something I want to undermine, and if it came across
like that, then I am truly sorry.
> Please revise your model so that a silly mistake can be
> sidestepped without affecting 200 projects. For example, just alerting
> log4j-dev and having the 200 affected projects to use yesterday's
> version of log4j would have been much better.
This is actually what the current implementation tries to do, but it has not
been solved properly. But Stefano is now on a mission to re-do a lot of
Gump's behaviour, and hopefully there will be improvements in both this area,
as well as the notification system.
> I am pleased to see that the problem on our side was corrected
> promptly. I am much less so with nature of the social interaction that
> led to it.
Looking back at my posts, I have a hard time seeing them as coming on hard.
The first post was a plain "FYI" in case you guys haven't noticed.
The second was a response to Mark noticing it as well, where I have both a
"May I suggest..." and a smiley...
I would like to know, if it is some old grudges you have against me, or how
humble does one have to be, not to offend you, Ceki... I am confused.
Cheers
Niclas
--
+------//-------------------+
/ http://www.dpml.net /
/ http://niclas.hedhman.org /
+------//-------------------+
---------------------------------------------------------------------
To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
For additional commands, e-mail: log4j-dev-help@logging.apache.org
Re: cvs commit: logging-log4j/src/java/org/apache/log4j/net
SMTPAppender.java
Posted by Stefan Bodewig <bo...@apache.org>.
On Tue, 14 Dec 2004, Ceki Gülc <ce...@qos.ch> wrote:
> The fact that a small typo affects 200 projects should tell you that
> something is wrong with Gump's current approach.
Absolutely agreed. We know we could do better, we just haven't had
the time/skills to change it yet.
> For example, just alerting log4j-dev and having the 200 affected
> projects to use yesterday's version of log4j would have been much
> better.
Which is what we intend to do one day.
> I am pleased to see that the problem on our side was corrected
> promptly.
Many thanks.
> I am much less so with nature of the social interaction that led to
> it.
I understand that and agree. I also understand Niclas, since he'd
certainly would love to see Gump build more stuff than it did and is
frustrated by glitches like this one. We all know that the typo could
have been caught before the commit, but we also know that things like
this have happened to all of us - probably more than once.
Cheers
Stefan
---------------------------------------------------------------------
To unsubscribe, e-mail: general-unsubscribe@gump.apache.org
For additional commands, e-mail: general-help@gump.apache.org
Re: cvs commit: logging-log4j/src/java/org/apache/log4j/net
SMTPAppender.java
Posted by Stefan Bodewig <bo...@apache.org>.
On Tue, 14 Dec 2004, Ceki Gülc <ce...@qos.ch> wrote:
> The fact that a small typo affects 200 projects should tell you that
> something is wrong with Gump's current approach.
Absolutely agreed. We know we could do better, we just haven't had
the time/skills to change it yet.
> For example, just alerting log4j-dev and having the 200 affected
> projects to use yesterday's version of log4j would have been much
> better.
Which is what we intend to do one day.
> I am pleased to see that the problem on our side was corrected
> promptly.
Many thanks.
> I am much less so with nature of the social interaction that led to
> it.
I understand that and agree. I also understand Niclas, since he'd
certainly would love to see Gump build more stuff than it did and is
frustrated by glitches like this one. We all know that the typo could
have been caught before the commit, but we also know that things like
this have happened to all of us - probably more than once.
Cheers
Stefan
---------------------------------------------------------------------
To unsubscribe, e-mail: log4j-dev-unsubscribe@logging.apache.org
For additional commands, e-mail: log4j-dev-help@logging.apache.org