On Feb 6, 10:04am, chris.hega...@oracle.com (Chris Hegarty) wrote:
-- Subject: Re: RFR: race with nested repos in hgforest.sh
| Yes this is better, but the quotes are incorrect. It should simply be
|for rc in ${tmp}/*.pid.rc ; do
This will break if there is no match.
christos
On 6 Feb 2013, at 16:26, chris...@zoulas.com wrote:
> On Feb 6, 10:04am, chris.hega...@oracle.com (Chris Hegarty) wrote:
> -- Subject: Re: RFR: race with nested repos in hgforest.sh
>
> | Yes this is better, but the quotes are incorrect. It should simply be
> |for rc in ${tmp
Neil,
I think a wider discussion needs to be had around how this script should
operate, but I didn't want this specific issue to get side tracked by that.
The hgforest script already clones/pulls in parallel, the problem here
is that it just doesn't work with older versions of hg. And worse t
fwiw, I've always found the attempts by hg_forest to process things in
parallel cause me more pain than it's worth.
Marching through the repo locations serially in order might take a few
moments longer elapsed time, but creates instantly understandable
status / output / log and points for returnin
On 06/02/2013 10:24, David Holmes wrote:
Thanks Chris. I missed a couple of obvious things there. :(
One thing related to the repo name in the $(rc) variable. Somewhere the
repo name gets mangled:
+ echo jdk/src/closed
+ sed -e s@./@@ -e s@/@_@g
+ repopidfile=jdsrc_closed
the / to _ doesn't
Thanks Chris. I missed a couple of obvious things there. :(
One thing related to the repo name in the $(rc) variable. Somewhere the
repo name gets mangled:
+ echo jdk/src/closed
+ sed -e s@./@@ -e s@/@_@g
+ repopidfile=jdsrc_closed
the / to _ doesn't work for the first / (not part of your fi
Thank for the review. comments inline...
On 06/02/2013 01:53, David Holmes wrote:
204 # Terminate with exit 0 only if all subprocesses were successful
205 ec=0
206 if [ -d ${tmp} ]; then
207 for rc in `ls ${tmp}/*.pid.rc` ; do
I don't think you need to subshell an invocation of ls, simply do
f
Hi Chris,
I don't speak fluent sh but ...
204 # Terminate with exit 0 only if all subprocesses were successful
205 ec=0
206 if [ -d ${tmp} ]; then
207 for rc in `ls ${tmp}/*.pid.rc` ; do
I don't think you need to subshell an invocation of ls, simply do
for rc in "${tmp}/*.pid.rc" ; do
On 02/05/2013 06:21 AM, David Holmes wrote:
Chris,
When these failures occur does the failure get reflected in an error
exit code?
Not currently, good point.
I'm seeing hudson builds merilly buildinh OpenJDK instead of Oracle JDK
because the closed repos were skipped due to this bug :(
Thi
Fixing get_source.sh problem:
We can hardcode pre-creation of the closed repo locations as mkdir -p calls.
We can do the openjdk clone first (in parllell) and when that is
finished do the closedjdk clone.
We can drop parallell clone altogether because it does not give the
speedup we are looking fo
Chris,
When these failures occur does the failure get reflected in an error
exit code?
I'm seeing hudson builds merilly buildinh OpenJDK instead of Oracle JDK
because the closed repos were skipped due to this bug :(
David
On 2/02/2013 12:40 AM, Chris Hegarty wrote:
[ to build-dev and core
How about adding a message to indicate the sleep, just in case the
directory is never created for whatever reason. If you don't like too
many such messages, you can print the message after the initial wait.
177 while [ ! -d "$path" ] ## nested repo, ensure containing dir
exists
1
On 4 Feb 2013, at 19:36, Ioi Lam wrote:
> How about adding a message to indicate the sleep, just in case the directory
> is never created for whatever reason. If you don't like too many such
> messages, you can print the message after the initial wait.
I understand the reason for your suggest
Looks fine to me.
-kto
On Feb 1, 2013, at 6:40 AM, Chris Hegarty wrote:
> [ to build-dev and core-libs-dev, expect reviewer from either, but will
> integrate through jdk8/tl ]
>
> This issue is mainly of interest to Oracle engineers, but it effects the
> public hgforest script.
>
> When hgfo
On 04/02/2013 10:04, Fredrik Öhrström wrote:
2013-02-04 10:48, Chris Hegarty skrev:
There has been some discussion on this with David, but no other takers.
Fredrik, am I right that you added the support for parallel
clone/update? I believe the changes I have are correct, but maybe you
had other
2013-02-04 10:48, Chris Hegarty skrev:
There has been some discussion on this with David, but no other takers.
Fredrik, am I right that you added the support for parallel
clone/update? I believe the changes I have are correct, but maybe you
had other ideas about how to solve this issue?
-Chr
There has been some discussion on this with David, but no other takers.
Fredrik, am I right that you added the support for parallel
clone/update? I believe the changes I have are correct, but maybe you
had other ideas about how to solve this issue?
-Chris.
On 01/02/2013 14:40, Chris Hegarty
On 02/02/2013 12:37 AM, David Holmes wrote:
...
I can see that happening if the sleeps are too short. But what I noticed
from the logging was that the closed repos were being processed before
the open ones - which shouldn't happen.
I believe the output is misleading. I think the output from
On 2/02/2013 9:06 AM, Chris Hegarty wrote:
On 1 Feb 2013, at 22:37, David Holmes wrote:
Hi Chris,
On 2/02/2013 12:40 AM, Chris Hegarty wrote:
[ to build-dev and core-libs-dev, expect reviewer from either, but will
integrate through jdk8/tl ]
This issue is mainly of interest to Oracle engin
On 1 Feb 2013, at 22:37, David Holmes wrote:
> Hi Chris,
>
> On 2/02/2013 12:40 AM, Chris Hegarty wrote:
>> [ to build-dev and core-libs-dev, expect reviewer from either, but will
>> integrate through jdk8/tl ]
>>
>> This issue is mainly of interest to Oracle engineers, but it effects the
>> p
Hi Chris,
On 2/02/2013 12:40 AM, Chris Hegarty wrote:
[ to build-dev and core-libs-dev, expect reviewer from either, but will
integrate through jdk8/tl ]
This issue is mainly of interest to Oracle engineers, but it effects the
public hgforest script.
When hgforest.sh is run with an addition ar
[ to build-dev and core-libs-dev, expect reviewer from either, but will
integrate through jdk8/tl ]
This issue is mainly of interest to Oracle engineers, but it effects the
public hgforest script.
When hgforest.sh is run with an addition argument to specify a closed
server, there is a proble
22 matches
Mail list logo
