Hey, I used to have a TRS-80. 32k of RAM...yup, those were the good 'ol days.
I agree that shortening the DT artifacts is only moving the bubble around. The fact is that for all
of its issues there are a lot more Windows users out there than the *nix OS's.
One area to improve things is to eliminate this redundancy:
daytrader-derby-tomcat_streamer.jar\1.2-SNAPSHOT\daytrader-derby-tomcat_streamer.jar-1.2-SNAPSHOT.car
At a minimum we burn up 47 characters with
daytrader-derby-tomcat_streamer.jar\1.2-SNAPSHOT where the actual artifact name contains exactly the
same information for the most part. I'm not sure what the right solution is here but if we had a
creative way to remove the first part we'd get some serious relief.
Is the reason we have this so we can use the Maven component directly? If we decided to modify our
repository code to do some introspection would that help? Dain, I know you wrote most of the code
with David, what are your thoughts on this?
Jason Dillon wrote:
I think that in general... the m2 repository using the m2 style of
artifactId and groupId names is not compatible with Windows.
Short of changing the way m2 works or altering the m2 naming standard, I
do not see any simple way around this problem.
And we are getting hit by this in two places too... building the server,
and then again when running the server.
I think we need to find a real solution to this issue and not just
truncate the names to get around problem... until it pops up again...
which is will.
And as much as I would love to just give all windows users the shaft and
force them to snap out of their microsoft induced blue screen of death,
reboot every day, annoying dancing paper clip, stupid file locking, file
names can't be longer than 256 insanity... and I would really love to do
that... we can't :-(
Lucky we don't have to support TRS-80, CPM or Altair users...
--jason
On Jul 19, 2006, at 7:38 AM, anita kulshreshtha wrote:
Jeff,
We also need to shorten the artifactId. It makes a good puzzle -
Come up with sensible groupId and artifactId such that this does not
exceed 256!
Example 2 -
D:\geronimo-1.2\configs\daytrader-tomcat\target\reposito
ry\org\apache\geronimo\configs\daytrader-derby-tomcat_streamer.jar\1.2-SNAPSHOT\daytrader-derby-tomc
at_streamer.jar-1.2-SNAPSHOT.car\org.apache.geronimo.modules\ge-activemq-rar\1.2-SNAPSHOT\rar\
activemq-3.2.4-SNAPSHOT.jar
This example is taken from
http://www.nabble.com/M2-%3A-build-on-Windows-p4914787.html
is 256+30 long!
Cheers
Anita
--- Jeff Genender <[EMAIL PROTECTED]> wrote:
I am a big fan of org.apache.geronimo.samples.daytrader. But...are
we
leading ourselves down the path of a Windows dir size too big
nightmare?
Jeff
Jason Dillon wrote:
Why not give it is own:
org.apache.geronimo.daytrader
--jason
On Jul 18, 2006, at 11:59 AM, Prasad Kashyap wrote:
The DayTrader project needs a consistent groupId for all it's
artifacts. Currently, some of it's artifacts are under the
"geronimo"
groupid while others are under the "org.apache.geronimo" groupid.
I discussed this with Matt and we decided to pose this Q to the
community. One option is to have a groupId called
"org.apache.geronimo.samples" which will be home to daytrader
artifacts, other samples we currently have and ones to come.
One other thing to consider is the very long artifactId name of
some
of daytrader's artifacts. An example is something like
daytrader-derby-jetty-streamer. When this is coupled with an
equally
long groupId name, we'll be close to hitting the path limit on
windows.
Cheers
Prasad
__________________________________________________
Do You Yahoo!?
Tired of spam? Yahoo! Mail has the best spam protection around
http://mail.yahoo.com
