[
Date Prev][
Date Next][
Thread Prev][
Thread Next][
Date Index][
Thread Index]
[
List Home]
|
Re: [buckminster-dev] help with CVS
|
Hi Todd,
By all means, enter a bugzilla. We will resolve this one way or another.
I'm more or less a CVS server illiterate and I have always thought that
the server concealed the actual physical location of folders on disk,
whether or not they were symlinked etc. The CVSNT server I use does that
(there's a 'location' which is the physical location on disc and a
'name' which is the cvsroot as it appears to the clients). Apparently,
Linux/Unix systems work differently.
Thanks Todd for spending time tracking this down.
Regards,
Thomas Hallgren
Todd_Lee@xxxxxxxx wrote:
Okay, so just to follow up on-list with the cause/resolution for this (in
case anyone besides myself has worn a hole in their desk from banging their
head :) ) With Thomas' help we were able to determine that the cvs provider
(reader) built into buckminster does not support CVS repositories whose
CVSROOT contains symbolic links.
ie -
physical absolute path:
/path/to/cvsroot
path used externally:
/path/to/cvsroot/with/symlink -> /path/to/cvsroot
As a result, the provider will fail to get metadata for any files in the
repository, making for an unsuccessful resolution. The reason is that,
while the the provider uses the external path in the rmap, when collecting
metadata it gets the file location (physical absolute path) from the RCS
file info gathered via the 'rlog' request. A validation check that matches
the file paths fails, causing the file to be skipped - ignoring its
metadata.
As a workaround, CVS can be reconfigured to add the additional path as an
alias for the CVSROOT (which we've done) but I wonder if this should be
entered as a bugzilla issue to be handled inside Buckminster?
Thanks,
Todd
Todd_Lee@xxxxxxxx
Sent by:
buckminster-dev-b To
ounces@xxxxxxxxxx Buckminster developer discussions
g <buckminster-dev@xxxxxxxxxxx>
cc
02/15/2008 09:07 Subject
AM Re: [buckminster-dev] help with CVS
Please respond to
Buckminster
developer
discussions
<buckminster-dev@
eclipse.org>
Hi Thomas,
Thanks again for all your help in trying to get us up and running...
I'm actually working from home today, so I'm on our vpn and I just grabbed
the latest buckminster from the update site.
I was able to successfully run your public cquery and everything
resolved/materialized without incident.
Feeling bold, I tried the same helloworld test that I set up on our local
cvs yesterday and I still got the same error. I'm left scratching my head!?
Oh, and so as not to create any wild goose chases, I did find docs last
night that confirmed that eclipse cvs should be fine with cvs going all the
way back to 1.11.1, so I'm assuming our 1.11.3 shouldn't be a problem.
If you have any other suggestions, please feel free to fire them off.
One question - is there any way to get more debug info out of the
resolution stage? I have all three log settings in the prefs set to full
DEBUG and there's still not a whole lot of information in the console or
the .metadata/.log file.
It would probably be helpful if the debug statements actually printed out
checks for the files inside the cvs path that buckminster uses to poll for
metadata for a given component type. Could give us a better idea of where
it's failing?
Thanks
Todd
Todd Lee
Software Tools
AMI Semiconductor
Tel: +1.519.884.9696 ext 2242
Fax: +1.519.884.0228
E-mail: todd_lee@xxxxxxxx
-----buckminster-dev-bounces@xxxxxxxxxxx wrote: -----
To: Buckminster developer discussions <buckminster-dev@xxxxxxxxxxx>
From: Thomas Hallgren <thomas@xxxxxxx>
Sent by: buckminster-dev-bounces@xxxxxxxxxxx
Date: 02/15/2008 05:06AM
Subject: Re: [buckminster-dev] help with CVS
Hi Todd,
>
> Thanks for the effort - sorry to report it's a no go...
> After spinning it's wheels for a while (pressed resolve to wizard) I get
a
> dialog popup the details of which are:
>
> "Could not connect to
:pserver:anonymous@xxxxxxxxxxx:/cvsroot/test:
> I/O has been interrupted.
> Timeout while reading from input stream"
>
A difficult one to crack, this is...
Last night I tried using machines that were on the same LAN as the CVSNT
server. So today I tried
running the resolution from one machine in our office in Pilsen Czech
Republic (running Vista) and
another machine here in Sweden (running Linux). No luck. I just can't
provoke this error. It works
like a charm every time.
I made the cquery public. You can find it here:
http://www.tada.se/pub/queries/demo.cquery
This query contains relative URL's so it will really require the latest
version of Buckminster.
- thomas
_______________________________________________
buckminster-dev mailing list
buckminster-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/buckminster-dev
AMI Semiconductor - "Silicon Solutions for the Real World"
NOTICE:
This electronic message contains information that may be confidential or
privileged. The information is intended for the use of the individual or
entity named above. If you are not the intended recipient, please be aware
that any disclosure, copying, distribution or use of the contents of this
information is prohibited. If you received this electronic message in
error, please notify the sender and delete the copy you received.
_______________________________________________
buckminster-dev mailing list
buckminster-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/buckminster-dev
AMI Semiconductor - "Silicon Solutions for the Real World"
NOTICE:
This electronic message contains information that may be confidential or privileged. The information is intended for the use of the individual or entity named above. If you are not the intended recipient, please be aware that any disclosure, copying, distribution or use of the contents of this information is prohibited. If you received this electronic message in error, please notify the sender and delete the copy you received.
_______________________________________________
buckminster-dev mailing list
buckminster-dev@xxxxxxxxxxx
https://dev.eclipse.org/mailman/listinfo/buckminster-dev
