[Savannah-help-public] Savannah coreutils no longer syncing CVS data

[Bob Proulx]

The savannah CVS repository for coreutils is no longer sync'ing.  Can
someone look at the problem?

Savannah has a cron task that runs at :11 minutes after the hour that
updates by rsync the repository from here:

  rsync.proulx.com::coreutils/coreutils

On my end in my logs I see the following errors repeated every hour.

  Dec 15 22:11:34 joseki rsyncd[23852]: rsync on coreutils/ from 
savannah.gnu.org (199.232.41.3) 
  Dec 15 22:11:35 joseki rsyncd[23852]: rsync: connection unexpectedly closed 
(111 bytes received so far) [sender] 
  Dec 15 22:11:35 joseki rsyncd[23852]: rsync error: error in rsync protocol 
data stream (code 12) at io.c(420) 
  Dec 15 22:11:36 joseki rsyncd[23853]: rsync on coreutils/CVSROOT/ from 
savannah.gnu.org (199.232.41.3) 
  Dec 15 22:11:36 joseki rsyncd[23853]: rsync: connection unexpectedly closed 
(125 bytes received so far) [sender] 
  Dec 15 22:11:36 joseki rsyncd[23853]: rsync error: error in rsync protocol 
data stream (code 12) at io.c(420) 

I am able to rsync those bits to other networks so I think my end is
okay.  If someone would look at the Savannah end and debug what is
happening there it would be appreciated.

Thanks
Bob

-- 
Bob Proulx <[EMAIL PROTECTED]>
http://www.proulx.com/~bob/


___
Savannah-help-public mailing list
[EMAIL PROTECTED]
http://lists.gnu.org/mailman/listinfo/savannah-hackers

[Savannah-help-public] [sr #105656] Duplicate coreutils source cvs upload scheme to gawk source cvs

[Bob Proulx]

URL:
  

 Summary: Duplicate  coreutils source cvs upload scheme to
gawk source cvs
 Project: Savannah Administration
Submitted by: rwp
Submitted on: Thursday 11/02/2006 at 23:16
Category: Developer CVS
Priority: 5 - Normal
Severity: 3 - Normal
  Status: None
 Assigned to: None
Originator Email: 
Operating System: None
 Open/Closed: Open

___

Details:

Please duplicate the coreutils source cvs upload process for the gawk cvs
source code upload process.  

Currently coreutils pulls the source cvs repository every hour by cron with
rsync from my rsync server.  Please duplicate that for gawk source cvs.  For
gawk substitute gawk-stable for coreutils and all other information is the
same.  (Trying to avoid actually saying details here where they are web
accessable.  The public location is on Savannah, not my server.  Locating the
coreutils cron will provide all of the details for setting up gawk the same
way.)

Please put these rsyncs sequentially in a script or otherwise configured so
that the rsync transfers do not happen at the same time.  Doing the rsyncs
one after the other in a script would be perfect.  Doing the rsyncs both at
the same time would be bad.

Please contact me if the terse information on this web form is not clear. 
Thanks!







___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/



___
Savannah-help-public mailing list
[EMAIL PROTECTED]
http://lists.gnu.org/mailman/listinfo/savannah-hackers

[Savannah-help-public] [sr #105656] Duplicate coreutils source cvs upload scheme to gawk source cvs

[Bob Proulx]

Follow-up Comment #2, sr #105656 (project administration):

Thank you for working to set this up.  However I think there must be a
problem.  I am not seeing any access for gawk-stable.  I can see the hourly
sync for coreutils in the syslog but nothing is seen for gawk-stable.  Can
you double check that please?

Thanks
Bob


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/



___
Savannah-help-public mailing list
[EMAIL PROTECTED]
http://lists.gnu.org/mailman/listinfo/savannah-hackers

[Savannah-help-public] [sr #105659] Dealing with SPAM

[Bob Proulx]

Follow-up Comment #1, sr #105659 (project administration):

Hi Mike,

I happened to see your note while looking through the requests for something
else.  As the author of the listhelper stuff your message caught my eye.

Listhelper is a mail processing robot that responds to mailman moderator
messages.  In each mailman moderator message contains the full poster message
along with the md5sum.  Sending the md5sum back to mailman will cause the
message to be discarded.  The listhelper robot receives these moderator
messages, unpacks the MIME attachments, runs the original message through a
conservative spamassassin configuration, if classified as spam then sends the
md5sum back to mailman where it is discarded.  This works fairly well for
keeping the spam off of the mailing list.

Karl posted a nice summary a while back.  Let me paraphrase.
To set this up for your mailing lists:
 * On the General Options page:
   - add the listhelper addr as a moderator (not administrator).
   - ensure admin_immed_notify=yes
   - ensure respond_to_post_requests=no
 * On the Privacy Options > Sender Filters page:
   - ensure generic_nonmember_action=hold

That is it.  Without going into every detail, the result is that message that
remain in the hold queue will be piped through a
very-conservatively-configured spamassassin, and 95+% of spam will be
automatically deleted after a short delay (usually measured in minutes,
unless things get backlogged).

Real messages from non-members (or non-approved addresses) will remain in the
hold queue, and need to be approved by a human.  (I recommend adding real
addresses to the "accept in the future" list when approving them.)  Thus, you
should pay attention to the once-a-day mailman notification when there are
"pending message(s)", but ignore the once-per-message mailman notification.

The process does not affect messages from members and approved addresses,
which continue go through without delay.

Futures: GNU sysadmin continues to work on an integrated mail handling system
that will be an improvement in many ways over this.  But this is what we could
do now, to try to keep from drowning in the junk in the meanwhile.



___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/



___
Savannah-help-public mailing list
[EMAIL PROTECTED]
http://lists.gnu.org/mailman/listinfo/savannah-hackers

[Savannah-help-public] [sr #105656] Duplicate coreutils source cvs upload scheme to gawk source cvs

[Bob Proulx]

Follow-up Comment #4, sr #105656 (project administration):

I can see the updates running in the syslog now.  I think things are working.
Thanks!

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/



___
Savannah-help-public mailing list
[EMAIL PROTECTED]
http://lists.gnu.org/mailman/listinfo/savannah-hackers

[Savannah-help-public] [sr #106304] Bug spam from logged in spammers?

[Bob Proulx]

URL:
  <http://savannah.gnu.org/support/?106304>

 Summary: Bug spam from logged in spammers?
 Project: Savannah Administration
Submitted by: rwp
Submitted on: Friday 03/14/2008 at 11:27
Category: Trackers (bugs, support, tasks...)
Priority: 5 - Normal
Severity: 3 - Normal
  Status: None
 Assigned to: None
Originator Email: [EMAIL PROTECTED]
Operating System: None
 Open/Closed: Open
 Discussion Lock: Any

___

Details:

Is there a way to lock accounts from known spammers and to find and clean up
the spam they have left behind?  See these bugs for examples of pharmacy spam
from two different accounts.

  http://savannah.gnu.org/bugs/?16179
  http://savannah.gnu.org/bugs/?15280
  http://savannah.gnu.org/task/?3491
  http://savannah.gnu.org/bugs/?1212

Bob Proulx





___

Reply to this item at:

  <http://savannah.gnu.org/support/?106304>

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[Savannah-help-public] [sr #106304] Bug spam from logged in spammers?

[Bob Proulx]

Follow-up Comment #5, sr #106304 (project administration):

> Bob, for your initial question, maybe see
https://savannah.gnu.org/maintenance/SavaneTasks

The SavaneTasks page describes a process to clean spam through direct SQL
access.  I am happy to do this.  Where can I log in to be able to run the SQL
commands documented on that page?


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[Savannah-help-public] [sr #106304] Bug spam from logged in spammers?

[Bob Proulx]

Follow-up Comment #12, sr #106304 (project administration):

It seems poetic to me that this bug reporting a problem with bug spam is also
getting bug spam.  :-)


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[Savannah-help-public] [sr #106304] Bug spam from logged in spammers?

[Bob Proulx]

Follow-up Comment #25, sr #106304 (project administration):

> Meanwhile, do you by any chance know about
https://savannah.gnu.org/siteadmin/spamlist.php ?
> Reported spammers can be banned in a single click.

I did not know about that page but that page gives me a permission denied
error.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[Savannah-help-public] [sr #106304] Bug spam from logged in spammers?

[Bob Proulx]

Follow-up Comment #28, sr #106304 (project administration):

I figured the spamlist.php url wasn't available to normal mortals.  But since
I submitted this bug in the first place *and* that was offered as an option I
felt justified in saying that it didn't work for me.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[Savannah-help-public] [sr #106304] Bug spam from logged in spammers?

[Bob Proulx]

Follow-up Comment #31, sr #106304 (project administration):

Re: "Well I put the TextCHA 9 days ago and the spamlist.php stayed empty
since then - so you should have an idea on whether it works now :)"

These got through it since then:

  http://savannah.nongnu.org/task/?4755

But I am sure it is an improvement in spite of that counter example.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[Savannah-help-public] [sr #106892] gawk cvs cron updater needs recovery

[Bob Proulx]

URL:
  

 Summary: gawk cvs cron updater needs recovery
 Project: Savannah Administration
Submitted by: rwp
Submitted on: Wed 10 Jun 2009 12:56:32 PM MDT
Category: RECOVERY: CVS/SVN
Priority: 5 - Normal
Severity: 3 - Normal
  Status: None
 Assigned to: None
Originator Email: b...@proulx.com
Operating System: None
 Open/Closed: Open
 Discussion Lock: Any

___

Details:

On the old Savannah a cron rsync pull synchronized the gawk cvs source code
repository with Arnold's private copy.  The last successful contact was on Jun
2 00:21:02 -0600.  The cron task apparently ran hourly at 21 minutes after the
hour according to my logs.

Please recover this task and enable the source repository sync for GNU awk
again.

Notes:
  rsync://rsync.proulx.com/gawk/
  Arnold maintains two branches in two individual repositories.

Thanks!





___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[Savannah-help-public] [sr #106892] gawk cvs cron updater needs recovery

[Bob Proulx]

Follow-up Comment #2, sr #106892 (project administration):

I now see the connections in the logs again.

Thanks!


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[Savannah-help-public] [sr #107262] gawk CVS rsync fix needed

[Bob Proulx]

URL:
  

 Summary: gawk CVS rsync fix needed
 Project: Savannah Administration
Submitted by: rwp
Submitted on: Tue 09 Feb 2010 09:15:49 AM MST
Category: Developer CVS
Priority: 5 - Normal
Severity: 3 - Normal
  Status: None
 Assigned to: None
Originator Email: 
Operating System: None
 Open/Closed: Open
 Discussion Lock: Any

___

Details:

The Savannah rsync has stopped pulling gawk cvs updates.
Could someone look into it?  Previously this was set up
for sr #105656 for reference.  Here is a hint from that
previous exchange.

  > 21 * * * * gawkrsync rsync_external_cvs_repository.sh
rsync://rsync.xxx.com/gawk/ gawk

Some additional information: I had a breakage on my end
that may have been the start of the problem.  The
rsync-data user got dropped here and it broke the outgoing
gawk rsync.  It is fixed here now but Savannah isn't
pulling updates anymore.

Thanks
Bob





___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[Savannah-help-public] [sr #107262] gawk CVS rsync fix needed

[Bob Proulx]

Follow-up Comment #2, sr #107262 (project administration):

Thanks for fixing it.  It seems to be pulling now.  In the future
if you notice breakage like that if you could send me a note I
would know that there was a problem on my end sooner.

Thanks,
Bob


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[Savannah-help-public] [sr #107262] gawk CVS rsync fix needed

[Bob Proulx]

Follow-up Comment #4, sr #107262 (project administration):

Thanks!
Bob


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[Savannah-help-public] [sr #108063] Update gnetic.git/description

[Bob Proulx]

Follow-up Comment #1, sr #108063 (project administration):

Done.  I have changed the description to "Tool to clone GNU/Linux systems over
a network."

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

Re: [Savannah-help-public] request for file name change...

[Bob Proulx]

Hello Rüdiger,

Rüdiger Asche  wrote:
> may I humbly suggest that src/core/timers.c be renamed to tcp_timers.c ?
> 
> The rationale is that we evaluate lwip in conjunction with FreeRTOS (a
> rather widely used combination in the field), and FreeRTOS also has a file
> named timers.c in its code base. Our IDE (iSystem WinIDEA) does not allow
> multiple files with the same name in the same code base. In any case, it is
> considered good coding practice to tag file names with its context, so it
> would be a useful thing to do in either  case.

You have sent this request to the savannah-help-public mailing list.
This mailing list is for help with the Savannah GNU software forge
which is committed to free software.  You can learn more about
Savannah here at the home page:

  http://savannah.gnu.org/

But you are asking about some specific file from some unknown program
or project.  I am sorry but this is the wrong place to ask about it.
At the least you would need to locate the people responsible for
whatever program you are asking about.

Perhaps you were wanting to contact the lwip-users mailing list?
Or perhaps lwip-devel?

  https://lists.nongnu.org/mailman/listinfo/lwip-devel

Bob

> Diese E-Mail enthält vertrauliche und/oder rechtlich geschützte
> Informationen. Wenn Sie nicht der richtige Adressat sind oder diese E-Mail
> irrtümlich erhalten haben, informieren Sie bitte sofort den Absender und
> vernichten Sie diese Mail. Das unerlaubte Kopieren sowie die unbefugte
> Weitergabe dieser Mail ist nicht gestattet.
>  
> 
> This e-mail may contain confidential and/or privileged information. If you
> are not the intended recipient (or have received this e-mail in error)
> please notify the sender immediately and destroy this e-mail. Any
> unauthorised copying, disclosure or distribution of the material in this
> e-mail is strictly forbidden.

READ CAREFULLY.  By reading this email, you agree, on behalf of your
employer, to release me from all obligations and waivers arising from
any and all NON-NEGOTIATED agreements, licences, terms-of-service,
shrinkwrap, clickwrap, browsewrap, confidentiality, non-disclosure,
non-compete and acceptable use policies ("BOGUS AGREEMENTS") that I
have entered into with your employer, its partners, licensors, agents
and assigns, in perpetuity, without prejudice to my ongoing rights and
privileges.  You further represent that you have the authority to
release me from any BOGUS AGREEMENTS on behalf of your employer.

[savannah-help-public] [sr #108258] Git repo description

[Bob Proulx]

Follow-up Comment #1, sr #108258 (project administration):

I have made the requested change.  Do to various caching you may need to force
a browser refresh to see it.  I verified it in my browser and the new
description is now visible.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108261] bino help

[Bob Proulx]

Follow-up Comment #1, sr #108261 (project administration):

You have submitted a bug against Savannah Administration.  But you are talking
about the "bino" project.  Since Savannah Administration is not bino could you
close this bug and submit one against the Bino project?  Thanks.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108294] bzr access to Emacs source code is down

[Bob Proulx]

Follow-up Comment #2, sr #108294 (project administration):

This was due to problems that occurred during an upgrade of the vcs savannah
system.  I did the upgrade and this broke during it.  The bzr path is a custom
one for Savannah.  There were various problems associated with python and bzr.
 Michael debugged this and fixed it.  I believe that it should be working
again now.  Please try it now.  If so would you be so kind as to let us know
and close this bug report?  Thanks!

Sorry for the disruption.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

Re: [savannah-help-public] Web interface doesn't work at 4am

[Bob Proulx]

Richard Stallman wrote:
> The web interface to Savannah is failing now (4am Boston time) due to
> timeouts.  I'm told this happens every day at this time.  I guess
> Savannah is doing some sort of lengthy cron job that hogs the machine
> so much the web interface times out.
> 
> Access using cvs is also slow, but it does work.

Thank you for the report.  I know the site uses Nagios monitoring.
The problem hasn't been bad enough to trip the monitoring to report a
problem.  I haven't seen any automated failure reports.  That is
probably why it has continued unnoticed and unfixed.

> I suppose this is the best time to do the cron job.

That is very likely the "daily" cron task that runs all of the
housekeeping tasks.  It runs a directory of scripts one after the
other until they have all completed.  It is sequential so that the
machine isn't swamped with an impulse spike of many concurrent tasks.

At 4am most of the US folks would not notice the problem.  (Because
our machines are also running housekeeping tasks at that time.  And so
I avoid trying to overload my machines during that time period too.)

But Savannah is a VM and the underlying host also has daily cron tasks
to run for housekeeping.  It is possible that the problem isn't on the
Savannah VM itself but on the underyling host that is hosting it.
That would trickle down and cause the VM to be slow even though
nothing on the VM is a direct effect of it.

> Is there a way to make it hog the machine less?
> 
> Limit what fraction of the memory it can have?
> Stop it for 10 seconds after 20 seconds of running?

I will sign up to look into the problem.  The first step will be to
understand what it is on the system that is causing it.  After we
understand the root cause then the solution will likely appear from
it.  And improving the monitoring so that we can be notified when it
appears as a problem.

Bob

Re: [savannah-help-public] Web interface doesn't work at 4am

[Bob Proulx]

Richard Stallman wrote:
> The housekeeping needs to be done, and 4am is a fine time to do it.

Agreed.  I was not suggesting anything different.

> I am only suggesting to lower its priority somehow.
> so that web interface requests don't get blocked.
> 
> Maybe there is a way to turn off the time-out.
> If they did not time out, they would work, just slowly.

To be clear it isn't normal for web sites to time out during the daily
housekeeping run of tasks.  That is abnormal.  The web server should
function acceptably well during the housekeeping task run.  It might
not be as responsive due to extra cpu load.  But it shouldn't be
timing out for normal user usage of it.  It should still operate 24x7.

Systems can normally do both web serving and housekeeping at the same
time.  Plus the Savannah systems are scaled up to be quite good high
performance systems with plenty of cpus and ram.  So I can only
believe that there is also a third thing happening at the same time.
It is that unknown third thing that I wish to know about.

Please give me a bit of time to look into the problem.  I am sure I
can figure out what is happening and improve it.

Bob

Re: [savannah-help-public] Web interface doesn't work at 4am

[Bob Proulx]

Richard Stallman wrote:
> The housekeeping needs to be done, and 4am is a fine time to do it.

Sorry.  I was mistaken.  The servers run natively in the UTC timezone.
4am Boston time would be 8am UTC.  The daily cron task is set to start
at 6:25am UTC daily.  That is 2:25am US/Eastern time.  They should be
well done and gone by 4am Boston time.

I have added additional monitoring.  So far I have not had a timeout
from the web interface in the last 24 hours.  I am polling it every
five minutes looking for delays and other problems.  I am looking from
off the local network so that the routers between are also being
end-to-end tested.  The load average on both systems look reasonable.
The process status looks okay.  I don't see any related errors in the
logs.  I was hoping I would be able to experience the problem myself
during that time period and then would be able to debug it.

> Maybe there is a way to turn off the time-out.
> If they did not time out, they would work, just slowly.

There isn't really a timeout on the Savannah side.  Your web browser
is likely timing out and causing the timeout you are seeing but that
is a timeout on your client and not on the Savannah server.

For a little description of the Savannah web interface the Apache2 web
server runs on frontend.savannah.gnu.org running an Apache PHP module
directly in Apache.  Savannah's PHP code accesses the MySQL database
on internal.savannah.gnu.org over tcp.  If the MySQL connection fails
or times out then an error display reporting the database connection
failure will be displayed.  If Savannah's Apache web server fails or
times out then your web browser will display a generic timeout
message.  Both machines are virtual systems and if the underlying host
is too busy then either of those systems may be reporting idle but
still acting slow.

The next time you see this problem can you note the exact time and
error that occurred?  Also please report any additional information
that you think might help identify this problem such as ability or not
to be able to ping it.  Perhaps it is a network connectivity issue?

I will continue to increase the monitoring capability so that we can
track this better.  I have just today added additional monitoring and
tracking and will need to look to tomorrow to see if anything in the
new history looks suspect.

Bob

Re: [savannah-help-public] cvs2svn

[Bob Proulx]

Edward Ned Harvey (savannah) wrote:
> Per the instructions at 
> http://savannah.gnu.org/maintenance/SvnImportExistingRepo
> I converted rdiff-backup from cvs to svn, and created a svn dump file.  The 
> dump file was uploaded via scp to:
>   dl.sv.nongnu.org:/srv/download/rdiff-backup/rdiff-backup-dump.bz2
> 
> Per the instructions at https://savannah.gnu.org/contact.php
> I tried to  submit a support request at 
> https://savannah.gnu.org/support/?func=addsupport&group=administration
> But I don't see any way to do that ...
> 
> So I am using the second method suggested.  Send email to
> savannah-help-pub...@gnu.org

Very good.  Thank you for following the procedures!

> Could you please restore the rdiff-backup svn dump, so I may begin
> using svn instead of cvs for this project?

I will set this up and let you know when it is done.

Bob

Re: [savannah-help-public] cvs2svn

[Bob Proulx]

Bob Proulx wrote:
> I will set this up and let you know when it is done.

Done and ready to go.

Bob

Re: [savannah-help-public] GSL: switching from bzr to git

[Bob Proulx]

Hi Patrick,

I can help you with your git on Savannah issue.  Sorry it took so long
to see your note and respond to it.

Patrick Alken wrote:
>   I'm the maintainer of the GNU Scientific Library (GSL) project. We
> were using git a few years ago and then switched to bzr. We recently
> took on a new developer who uses windows, and unfortunately has
> difficulties with committing code using the windows bzr client.
> Therefore we are considering switching back to git, however there is
> an old GSL git repository still on savannah (several years out of
> date).
> 
>   Would it be possible for the admins to remove the old GSL git
> repository, so we can start a new, fresh import of the current
> codebase?

Since it has been a few weeks I should ask first what the status is of
this problem?  Have you resolved it or does it still need some
attention?

I can rename the current master branch to oldmaster.  (I hate to
delete anything as a first step.  Renaming preserves the data.  Even
though I think you will tell me that it isn't good.  Better safe than
sorry.  Better to make a new connection before breaking the old
connection.)  That would open you a path to upload a new master
branch.  After a new master is uploaded then all can proceed and be
tested that way.  (After all is proven good then later the old master
can be removed as cleanup if desired.)

Bob

Re: [savannah-help-public] GSL: switching from bzr to git

[Bob Proulx]

Hi Patrick,

Patrick Alken wrote:
> Hello Bob, we would greatly appreciate if you could rename the
> current git master branch to something else, so we can re-import the
> current codebase into git.

Okay.  Done.  Upload away.  Let me know if you have problems.

Bob

[savannah-help-public] [sr #108409] Make the old GNU AUCTeX CVS repository read-only

[Bob Proulx]

Update of sr #108409 (project administration):

  Status:None => Done   
 Assigned to:None => rwp
 Open/Closed:Open => Closed 

___

Follow-up Comment #1:

Done.  

I don't do this every day so let me expand my thinking so that others can
check my work.  The group permission for the auctex cvs sources is the group
name auctex.  I believe that is what allows members of the auctex group to
commit changes.  I changed the group of all of those files to group root.  I
believe that will block all commits.  Please make an update if this turns out
to not be the case.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108409] Make the old GNU AUCTeX CVS repository read-only

[Bob Proulx]

Follow-up Comment #3, sr #108409 (project administration):

A commitinfo script is a good idea.  And it would have wider usage on other
projects that have moved.  Let's do it.

I modified auctex so that it once again group auctex and therefore could be
committed to.  Before doing that I created a
/usr/local/bin/commit_obsolete_notice script and configured it in the
commitinfo.

I created a rather generic message so that it could be used generically over
other projects that have also moved.  (I already have several in mind.  I
tested this on coreutils CVS.)

Please make another commit test to the CVS auctex and verify the new
behavior.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108411] Enable commit mails including diffs for GNU AUCTeX

[Bob Proulx]

Update of sr #108411 (project administration):

  Status:None => In Progress
 Assigned to:None => rwp

___

Follow-up Comment #1:

Hmm...  Doing *both* of those things seems not completely trivial.  Let me
explain.

Currently git projects all have a common hook script that reads the git config
file for the hooks.mailinglist value.  For auctex that value was
auctex-com...@gnu.org.  So the git commits have been going to that list.

I can trivially change the list name to auctex-diffs.  Then all of the commit
notifications would go there.  They would not go to auctex-commit anymore. 
Done.

I can trivially make the hooks.showrev change and then the custom showrev
would be in effect.  Done.

I have done the above.  This might be enough for you.  Let me know.

I can trivially add another email address to hooks.mailinglist and then the
email would go both places.  I did not do this.  I am simply mentioning the
capability.

In order to have *both* the current short commit notifications going to the
auctex-commit list *and* the full diffs going to the auctex-diffs list we will
need to split or duplicate the post-receive-email hook script so that it runs
twice.  Each run would have different values of hooks.showrev.  I would leave
one standard and then simply hack the other for the local showrev value.

I will stop here and wait for your feedback.  Let me know if the above is
sufficient.  I can pretty easily clone the hook script and make a local copy
that continues to send the current notifications to the auctex-commit list. 
If that is still desired.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108411] Enable commit mails including diffs for GNU AUCTeX

[Bob Proulx]

Follow-up Comment #3, sr #108411 (project administration):

Okay.  Let me tinker something together for you.

In CVS this is fully supported natively.  In git this isn't as nice since git
supports only one hook out of the box.  (Did I actually say that git wasn't as
nice as cvs?  No one is more shocked than myself!)  The git hook has stdin
piped into it and the script interprets it.  Therefore to run multiple hooks
you need to save off stdin into a temporary file and then hand it to each hook
script in turn.  Or the equivalent.  And to try to keep from breaking things I
will need to test in a victim location.  So please give me just a little bit
of tinker time to get it all tested and implemented.  Again this would be
useful for other projects too.

I looks to me like someone already tried to set up multiple hook scripts for
the auctex project.  The existing hook script was calling run-parts on a .d
subdirectory.  It worked because there was only one script in the .d
directory.  But it would not have worked if there were two.  Or rather the
second one would not have gotten any of the normal expected input.  Pretty
sure anyway. :-)

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108411] Enable commit mails including diffs for GNU AUCTeX

[Bob Proulx]

Update of sr #108411 (project administration):

  Status: In Progress => Done   

___

Follow-up Comment #4:

Done.  Tested on a vicitim repository.  Transfered to auctex.  I think it all
should be working.  I will monitor the -commit and -diffs lists and check to
verify that they are working.

For reference I basically did this to get multiple hook scripts.

while read oldrev newrev refname; do
  echo $oldrev $newrev $refname | /usr/local/bin/git-post-receive-hook
  echo $oldrev $newrev $refname | hooks/post-receive-mail-local
done


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108411] Enable commit mails including diffs for GNU AUCTeX

[Bob Proulx]

Update of sr #108411 (project administration):

 Open/Closed:Open => Closed 

___

Follow-up Comment #6:

Good deal!  Thanks for the feedback.  Closing out the ticket as done.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

Re: [savannah-help-public] (Bug on web.cvs.savannah.gnu.org) 500 Internal Server Error

[Bob Proulx]

joeko wrote:
> Is the website temporarily offline for maintenance work (or backup)?
> What's going on?

This has been a continuing problem.  The vcs server periodically
experiences a high load average greater than 20.  When it is in that
high load average state it is too slow responding to the frontend web
server and times out.  When the backend vcs times out then the
frontend web server emits the message you saw.

Presently there are various conjectures about the root cause of the
problem.  But it hasn't been solved.

Bob

[savannah-help-public] [sr #108415] Medias-admincfg

[Bob Proulx]

Update of sr #108415 (project administration):

  Status:None => Need Info  
 Assigned to:None => rwp

___

Follow-up Comment #1:

I have read your bug report in detail but I am unable to deduce
a problem.  Please say what problem you are having.  And please say how this
is related to Savannah, the project that you submitted the bug against.  It
appears to me that you may have been intending to contact the GNU arch project
about its documentation instead.  Note that the GNU arch project has been
decommissioned.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108417] sub-repository for Dragora

[Bob Proulx]

Update of sr #108417 (project administration):

  Status:None => Need Info  
 Assigned to:None => rwp

___

Follow-up Comment #1:

I don't see technical reasons why not.

Could you say a few words to describe your desire for this additional
respository?  The first thing that popped into my head was why not simply use
a branch?  Or perhaps if you have discussed this on a mailing list then a
pointer to the discussion?

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108417] sub-repository for Dragora

[Bob Proulx]

Update of sr #108417 (project administration):

  Status:   Need Info => Done   
 Open/Closed:Open => Closed 

___

Follow-up Comment #3:

Okay.  Done.  Same permissions as on dragora.

Let us know if you want any hook scripts enabled.  (Such as to mail off
notices of commits.)  Currently no hooks are enabled.  It is simply a bare
repository.




___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108420] Delete old user accounts.

[Bob Proulx]

Update of sr #108420 (project administration):

  Status:None => In Progress
 Assigned to:None => rwp

___

Follow-up Comment #1:

Okay.  Will do so.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108419] Delete user Account

[Bob Proulx]

Update of sr #108419 (project administration):

  Status:None => In Progress
 Assigned to:None => rwp

___

Follow-up Comment #1:

Okay.  Will do so.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108419] Delete user Account

[Bob Proulx]

Update of sr #108419 (project administration):

  Status: In Progress => Done   
 Open/Closed:Open => Closed 

___

Follow-up Comment #2:

Done.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108420] Delete old user accounts.

[Bob Proulx]

Update of sr #108420 (project administration):

  Status: In Progress => Done   
 Open/Closed:Open => Closed 

___

Follow-up Comment #2:

Done.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108415] Medias-admincfg

[Bob Proulx]

Update of sr #108415 (project administration):

 Open/Closed:Open => Closed 

___

Follow-up Comment #2:

Not having heard any response I assume this was a mistake and will close out
the ticket.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108421] Delete old user account.

[Bob Proulx]

Update of sr #108421 (project administration):

  Status:None => Done   
 Assigned to:None => rwp
 Open/Closed:Open => Closed 

___

Follow-up Comment #1:

Done.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

Re: [savannah-help-public] "Sysadmin volunteers to help with underlying infrastructure support"

[Bob Proulx]

Hi Ethan,

Ethan Henderson wrote:
> Hello, my name is Ethan Henderson, and I saw on the GNU help page that you
> guys could use a systems admin to "help with underlying infrastructure
> support," and would like to offer my help, if I can.

Great!  Many hands make light work.  Thanks for volunteering!

> I have little experience administrating systems, I do have basic experience
> with Ubuntu, I'm a fast learner, I do know a lot of PHP/SQL, and I have a
> desire to help.
> 
> I don't know what all the position requires, but maybe I can fill it, just
> let me know :)

There are many different facets of tasks for helping the Savannah
software forge and the GNU Project.  All could use some love and
attention.  A few of them in particular could use some help.

* Reviewing project submissions.  There is often a long backlog and
reviewers are always needed there.  See this page for a HOWTO on what
needs to be done there.  This is an area where anyone could be very
useful volunteering.
http://savannah.gnu.org/maintenance/HowToBecomeASavannahHacker/

* Savannah PHP development.  Since you mentioned PHP/SQL this might be
something that would interest you.  In particular there is some
concern that when the present Savane software forge was written that
the security attacks on the net now weren't known then.  There has
been some discussion recently about possible security vulnerabilities.

We could use someone to look at the current Savane / Savannah PHP code
with an eye toward security.  As well as adding features and fixing
other bugs.  Would you be interested in improving the Savannah PHP?

This could be done by different ways.  One is to review the source
code and fix bugs and improve it.  The Savannah source code is
available for git cloning.  Look at the Savannah defect tracker and
look for unresolved bugs for things that need to be done.  Submit
patches to fix them.
git clone git://git.sv.gnu.org/administration/savane

Another way is to try to use security analysis tools such as plugins
to Firefox and others such as "XSS Me" and "Tamper Data" and others
that are designed for security penetration testing.  And I am sure
there are other tools too.  This requires someone familiar with web
security issues.  And then of course any problems found will need to
be fixed.

* Subscribe to this mailing list.  (savannah-hackers-public.)  I
checked just now and didn't see your address on it.  That is why I
CC'd you specifically.  Normally we would just list-reply back to the
list.  Subscribe so that you can communicate with the rest of us and
us with you.  There are a confusing number of lists for this and that.
Initially the savannah-hackers-public is where you want to start.
https://savannah.gnu.org/maintenance/SavannahHackersCommunication/

Sound good?  Questions?  Ask on the mailing list.  Thanks!
Bob

[savannah-help-public] [sr #108426] Please import Subversion repository

[Bob Proulx]

Update of sr #108426 (project administration):

  Status:None => In Progress
 Assigned to:None => rwp

___

Follow-up Comment #1:

I will do this today and report when it is done.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108426] Please import Subversion repository

[Bob Proulx]

Update of sr #108426 (project administration):

  Status: In Progress => Done   
 Open/Closed:Open => Closed 

___

Follow-up Comment #2:

Done.  Let us know if there is anything else.  Such as adding notification
hooks or other.

Trivia: It took 1h50m of machine time to load 1161 revisions for 5.7 seconds
per revision on average.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108426] Please import Subversion repository

[Bob Proulx]

Follow-up Comment #5, sr #108426 (project administration):

Unfortunately this is currently not possible.  The web side of things only has
infrastructure to update from CVS.  Conceptually no one is against doing it
from other version control systems.  But practically speaking there hasn't
been anyone to actually make it happen.  And also to support the many various
options available today.  And by anyone I mean one of the FSF admins because
that is outside the Savannah umbrella and the Savannah Hackers do not have the
access to implement it.  It happens under the FSF umbrella.  Sorry.

If you wish to discuss the matter I suggest using the
savannah-hackers-pub...@gnu.org mailing list as it will involve the necessary
parties and easier than using the bug ticket system.  But eventually to have
something like that actually implemented we would need to involve the FSF
admins.



___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

Re: [savannah-help-public] "Sysadmin volunteers to help with underlying infrastructure support"

[Bob Proulx]

> * Subscribe to this mailing list.  (savannah-hackers-public.)
> 
> Ironically, "this list" is savannah-hackers@gnu.org, which is
> savannah-help-public, which is *not* savannah-hackers-public.
> (I know you know this, just pointing out for the record.)

Oops!  Somewhere long the way I got confused.  I only know it in
hindsight as you point out my error.  I had indeed confused them at
the moment I wrote that.

> As you say:
> There are a confusing number of lists for this and that.
> Indeed.

And I tripped on it again!  Oh well.

Bob

[savannah-help-public] [sr #108426] Please import Subversion repository

[Bob Proulx]

Follow-up Comment #7, sr #108426 (project administration):

OT drift: If you were using git instead of svn then I would suggest using the
git-cvs adaptor.  Then everything on the client side appears in git while
everything on the server side stays cvs.  There is the equiv git-svn adaptor
too.

However although I am a git person I still use native cvs for the web pages. 
They just are not complicated enough to need anything more.

As far as I know there isn't an svn-cvs wrapper in the same way.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108426] Please import Subversion repository

[Bob Proulx]

Follow-up Comment #9, sr #108426 (project administration):

The web pages may be checked out read-write with:

cvs -d :ext:yourusernameh...@cvs.savannah.gnu.org:/webcvs/classpathx co
classpathx

Or anonymous read-only with:

cvs -z3 -d:pserver:anonym...@cvs.savannah.gnu.org:/web/classpath co
classpathx

When you make commits the web pages are updated.  Use the .symlinks file to
manage web symlinks.  (Since cvs does not otherwise handle non-text special
files.)

I believe the control to see the web cvs documentation is under the project
administration select features page. Activate cvs and I think then the web cvs
menus will be active.  I think.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

Re: [savannah-help-public] Webpage with information on Savannah status?

[Bob Proulx]

Joris van der Hoeven wrote:
> These last days, I have been experiencing considerable slowdowns
> when using SVN on Savannah.  Is there some reason for that?

Could you keep track of the time of day when you have experienced
problems?  Also if you were dealing only with subversion or if you
were also dealing with a web page interface to subversion.  (The web
page interface uses a frontend machine that interfaces to the backend
vcs machine.)

The subversion backend VM instance is named "vcs" and has a load
problem that happens in the middle of the night local -0500 timezone
time.  It sometimes hits a load of 30 or so.  During that time it just
doesn't respond very fast.  The middle of the night part has gotten in
the way of anyone having been able to debug it.  Several of us have
looked at the problem but mostly we are not awake when the problem is
occurring.  It has been discussed on the mailing list.

Our best thoughts now is that it is from backup.  We have asked the
FSF admins to implement a bandwidth limit on the backup process but so
far AFAIK this has not been implemented yet.

> It might be nice to add a link in the margin of the website to
> some page with information on possible network problems,
> e.g. under the Site Help header.  Maybe such a page already exists,
> but I could not easily find it.

Just recently a new status notification page was announced.  I will
note that there isn't much activity there.  It was announced two
months ago and there have been no postings there yet.

  https://www.fsf.org/blogs/sysadmin/new-fsfstatus-notification-account

Bob

[savannah-help-public] [sr #108428] Please import dump into existing Subversion repository

[Bob Proulx]

Update of sr #108428 (project administration):

  Status:None => Need Info  
 Assigned to:None => rwp

___

Follow-up Comment #1:

Unfortunately that dump file isn't set up to merge on top of the existing
previous import.

 svnadmin: File already exists: filesystem '/srv/svn/classpathx/db',
transaction '1163-wb', path 'branches'

It is going to need some adjustment in order to load this history into the
existing subversion repository.  I would be happy to help but am swamped and
won't be able to research this.

Could you try the task yourself?  Import the previous classpathx dump.  Then
try importing your new dump.  That way you will see the errors yourself.  Then
create the new dump again but doing whatever it takes to make it work.  I will
be happy to help as I can along the way.

See also
http://svnbook.red-bean.com/en/1.7/svn.reposadmin.maint.html#svn.reposadmin.maint.migrate.svnadmin
for good information on doing this.  Perhaps the --parent-dir option?  Or
maybe filtered with svndumpfilter?


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

Re: [savannah-help-public] GSL: switching from bzr to git

[Bob Proulx]

Hi Patrick,

Patrick Alken wrote:
>   Thanks for your help earlier. We have imported a new git
> repository, under the branch 'master' - it looks like you changed
> the old repository to 'oldmaster'. Unfortunately when we do a clone
> of the repository, the default branch appears to be 'oldmaster'
> instead of 'master'. I don't know any way to change this remotely -
> is this something you can do for us?

Drat.  I didn't think about testing that when I did it.

I don't know if it is something that can be done remotely or not.  But
locally it is simply the name in the gsl.git/HEAD file.

  $ cat HEAD 
  ref: refs/heads/oldmaster

I simply changed that to point to refs/heads/master to fix it.  I just
now did a clone to test it and the branch is now master.

  $ git branch -l
  * master

Sorry about that.  Fixed now.

> The command appears to be:
> git symbolic-ref HEAD refs/heads/master

Hmm...  A new command!  It edits the file and does the same thing as
the above.  That would do it!  I will add that command to the toolbox.

> however this may need to be done directly on the savannah repository.

Yes.  Sorry again for missing that problem with the above.  I should
have double checked with a checkout.

Bob

[savannah-help-public] [sr #108428] Please import dump into existing Subversion repository

[Bob Proulx]

Update of sr #108428 (project administration):

  Status:   Need Info => None   
 Open/Closed:Open => Closed 

___

Follow-up Comment #3:

Okay.  I am sure that the history could be imported correctly.  It just needs
to be either dumped or filtered in the way that supports it.  (And I didn't
have the time to do the research myself.)  I would still be happy to do the
import for you if you wanted.  :-)

But okay on your last status.  I will close the ticket.  Feel free to ask
again if you decide otherwise.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

Re: [savannah-help-public] [Savannah-users] Certi:error(cannot connect to RTIA)

[Bob Proulx]

希望 wrote:
> Hi,I'm a certi library user of China.While using certi in my projects I met a 
> strange question.

You have reached the Savannah mailing list.  Savannah is the software
forge for people committed to free software.  The web site, the
version control hosting, that type of thing.

But you are asking about the certi project.  We don't know too much
about specific projects here.  It would be better to look at the
project and see what it says about where to get help.

The project web site:

  https://savannah.nongnu.org/projects/certi

Contains information about the project mailing lists.

  https://lists.nongnu.org/mailman/listinfo/certi-devel

Also, please do not send images.  That is a very large image at 3.5M!
The best thing is to cut and paste the *text* of the error verbatim.
The text of the error would be very small.  The text is searchable.
Images are not.  Text is small.  Images are not.  Sending the plain
text of the error and how to recreate it is best.

Bob


> I have two different projects(one is a 3D application called
> A,another is a Gis application called B),A and B were installed on
> Local PC,they shared data each other through CERTI.rtig.exe is
> running on another remote PC.
> Now A app can connect to rtia.exe (create hidden process
> successfully) but B app failed everytime.
> try{
> RTI:RTIambassador *Gp_rtiamb=new RTI:RTIambassador();}
> catch(RTI:RTIinternalError&e)
> {
>  e._reason;//
> }
> I found the failed reason is when new RTI:RTIambassador.A app can
> pass here but B app failed everytime.sometimes B app can pop up a
> console and output a line(rtia.exe :invalid numeric value:1,056).It
> means the tcp port used an invalid number?
> My two applications use the same certi configuration.the only
> difference is A has WinMain entry function but B has main entry.Is
> it the reason?
>
>
> YangJing @China
> 20131107

Re: [savannah-help-public] encryption key not recognized

[Bob Proulx]

Hello Frank,

Frank Ross (fross) wrote:
> Sorry, but having a lot of trouble getting keys to be recognized so
> that I can establish a connection for CVS.

This is the right place to ask for help with such things.

> I created the key using ssh-keygen, and added the key to the system.
> ...
> linux-ea8y:/usr/src # cvs -z3 -d:ext:ros...@cvs.savannah.gnu.org:/web/grub co 
> grub
> Permission denied (publickey).

Looking in the logs I see:

  Nov 11 17:33:35 vcs sshd[29044]: Invalid user rossff from [redacted]
  Nov 11 17:33:35 vcs sshd[29044]: input_userauth_request: invalid user rossff 
[preauth]
  Nov 11 17:33:35 vcs sshd[29044]: Connection closed by [redacted] [preauth]

The problem is that your account isn't considered valid.  But I don't
immediately see why not.  I see an entry for rossff in the user
database.  I tested this with myself and my user is valid.  But your
account isn't being seen as valid.  And so at the moment I am a little
bit at a loss.

  getent passwd rossff
  ...nothing...

Normally for a valid user that will return the account information.

I am still learning this part of the system.  If one of the other
Savannah Hackers doesn't jump in then I will dig into it further and
figure it out.  But off the top of my head I don't know and think it
should be working.

> Note, the sample key and my key differ at the end of the key where
> the user is given, my key file just has a space vs. the "==" given
> in the sample.  But, I tried modifying with no change in result.

The key is a binary blob encoded as a base64 string.  The equals signs
are part of the trailing padding.  Either one, two or none will appear
depending upon the length of the input that gets encoded.  Whatever is
generated by ssh-keygen will be the correct string.

> I also tried changing the key to root@[ip_addr] since I do not have
> a domain associated with this machine, it normally gives
> r...@linux-ea8y.site - perhaps the problem?

Not sure what you are referring to here.  If it is the last field in
the ssh key string then that can be ignored.  It is the comment
field.  By default ssh-keygen will put something that it tries to be
useful in there.  But the field isn't used for anything other than a
comment.  You are free to change it to something that makes more sense
for you.  By convention it is an email address.  But for example I
have several and therefore I use different hostnames in my address so
that I can identify which key is which key.

> linux-ea8y:/usr/src # ssh -v ros...@cvs.savannah.gnu.org

Note that only vcs commands are allowed.  You won't be able to
actually log in that way.  You will only be able to run cvs commands.

Note that on your end you are running as root.  That definitely isn't
recommended.  Normally we check out as ourselves and not as root.

Note that before you can commit you will need to be added to a group.
I looked at your profile on Savannah and you are not in any groups.
You don't have any permission for any projects.

Project/Group Information
This user is not a member of any Project/Group 

But that isn't the current problem.  (At least I don't think it is.)
First things first.

Bob

Re: [savannah-help-public] encryption key not recognized

[Bob Proulx]

Hi Frank,

Frank Ross (fross) wrote:
> Not worried about commits of code, in fact, I was just wanting to
> get the latest trunk of grub, which has a fix that I need.

Oh!  In that case you have the wrong repository.  The CVS repository
you were trying for is for the Savannah web pages.  That is why I
thought you must have been planning to make project web page
improvements.  And you don't need an account to read the archives.
You only need an account to push commits.

And the grub source code is in git not cvs.  Only the web pages are in cvs.

> I think I (at least) figured out that I could do an anonymous git
> and have the latest master of the grub source, which is my primary
> need at the moment.

Yes.  Exactly!  Like this in case anyone else is reading along.

  git clone git://git.savannah.gnu.org/grub.git

Bob

Re: [savannah-help-public] Using mod_proxy to host the project homepage elsewhere

[Bob Proulx]

Savannah Hackers,

Paulo H. "Taka" Torrens wrote:
> Is there any way you could activate it? Otherwise, is there any
> other way I could achieve this (without using ugly iframes)?

I am not sure how to answer Paulo.  Any ideas?

Here are some thoughts that appeared in my head...

He seems to be trying to use a .htaccess file.  I will probably
trigger a long thread (if you do then start a new thread please) but
those have a bad reputation for slowing server performance.  Not to
mention other concerns.  They are not enabled by default and I
wouldn't suggest to the FSF to enable them.  I wouldn't enable them.

AFAICS the redirect he tried to implement would take the user away
from the gnu.org site completely anyway.  So effectively he wouldn't
be using that domain name in that case anyway.  After the redirect
anyone who bookmarked would bookmark the new location.

I don't like the idea of redirecting off to a www.google.com page as
Paulo suggested in his email.  But I know that there are projects that
simply redirect off to other sites.  But I have no idea how those are
set up.  (Nor can I think of one off the top of my head.  Maybe I am
mistaken.)  If so I think those would simply be top level redirects
and very light weight on the FSF web page side of things.  Probably
simply an RT request to get them installed.

AFAIK the web pages are controlled by the FSF not the Savannah Hackers
team.  So appealing here isn't something we can affect any changes to
anyway.  There is only the upload to cvs for project teams to update
the web pages.  I don't see them setting up generic PHP access anytime
soon.  And Rails/Django would take even more effort.

Rails and Django are relatively heavy and really should have their own
contained daemons to serve their content.  It would be pretty heavy if
every project wanted their own Rails/Django framework running.  Not to
mention the problems with trying to keep those updated for security
vulnerabilities.

Bob

Re: [savannah-help-public] Using mod_proxy to host the project homepage elsewhere

[Bob Proulx]

Hi Karl,

Karl Berry wrote:
> Beyond that, in principle, installing mod_proxy with its potential
> hazards just to make it easy for people to avoid using us seems, um,
> wrong.  I personally am not inclined to change the world just to accede
> to one person's request but hey, it's not up to me.

I am sure that others would like it.  So it is more than just one
person's request.  But as a practical matter I don't think it scales.
There just aren't the resources all in one place for thousands of
unique project sites.  And for the resources needed to maintain them
for security upgrades.  I just don't think it is practical.  Nice in
concept.  Just unable to actually do it.

> AFAIK the web pages are controlled by the FSF not the Savannah Hackers
> 
> The web pages for www.gnu.org are controlled by the FSF.
> But savannah serves www.nongnu.org (from frontend).

I don't think so.  Isn't www.nongnu.org served from the FSF and is
just a top level http redirect back to savannah.nongnu.org?  The
savannah.nongnu.org site is served on frontend.  That is after the
redirect.  But www.nongnu.org is as far as I know served on the FSF
servers.

I guess I should ask about DNS.  I suppose it would be possible for
the DNS name to point to a completely different server.  Again that is
administered by the FSF admins.  I don't know if there is precedent
for it.  And philosophically and politically would the FSF want to be
a DNS registrar for nongnu projects?  Don't know that either.  Maybe
they would since they host them on nongnu.org.  WDYT?

Bob

Re: [savannah-help-public] Using mod_proxy to host the project homepage elsewhere

[Bob Proulx]

> The thing is, when someone uses www.nongnu.org, they get
> their access after agreeing to our policies (when they registered
> their savannah project).  E.g., they agreed not to promote proprietary
> software on their web pages.
> 
> Spreading that a step further to arbitrary hosting anywhere ... no.
> This is also a (philosophical) argument against mod_proxy, or
> mod_rewrite stuff, etc.  
> If people host web pages elsewhere, we don't want there to be any
> necessary connection to gnu or savannah.

Okay.  I think I have enough to answer this request.  I am going to
basically say:

Hello Paulo,

> Is there any way you could activate it? Otherwise, is there any
> other way I could achieve this (without using ugly iframes)?

In summary I think the answer is no.  Sorry.  I don't know of any way
to accomplish what you are asking.  But there are several points that
your question raises.  Rather than simply saying, sigh, not possible,
let me address some of them.

* Being able to host more fancy web pages is nice in concept but that
we just don't have the resources to actually do it.  Sorry.  PHP might
be lightweight enough but Rails/Django are both rather heavy weight
servers and both require significant resources to keep updated against
security vulnerabilities.  Can you imagine if just 10% of the projects
hosted wanted to have a unique web service hosted?  That would be
completely overwhelming!

* The web pages are actually hosted on the FSF servers.  Those are
organizationally distinct from the Savannah site and team.  We the
Savannah Hackers don't have administrative access to them.  For
anything related to the web pages all administrative control is
through the FSF admins.

* Your mod_proxy wouldn't actually do what you wanted it to do.  You
would redirect away from black.nongnu.org and over to www.google.com
and then the user would be there.  Any bookmarks would see the new
site.  Effectively you wouldn't be using black.nongnu.org at all at
that point.  (And I don't think anyone would like the idea of
redirecting over to www.google.com.)

* It looks like you are trying to use a .htaccess file.  Those have a
bad reputation for slowing down servers.  If enabled they require the
server to make a large number of stat(2) calls for every page served.
They are never enabled by default.  Again this is up to the FSF admins
but personally I would never enable them.  You might want to read:
https://wiki.apache.org/httpd/Htaccess

And so I don't really have an answer that you will like.  I might
suggest that for some of the things you are asking you might try
setting them up for yourself.  Then you will see the problems of
trying to do it on a thousand projects.  Whatever is done must be able
to scale to a large number of projects and must be manageable for a
long time with minimal people resources.  It's a problem!

Bob

Re: [savannah-help-public] Using mod_proxy to host the project homepage elsewhere

[Bob Proulx]

Hello Paulo,

Paulo H. "Taka" Torrens wrote:
> I understand your security issues, but I'd really like to use some
> complex features, so I'd need PHP (or Rails, or Jango, or...). A
> Wiki, a script to build compiler packages (for the user to download
> exactly what he needs), a pretty-printer to render the source code
> (including documentation comments and stuff), and so on...
> ...
> Is there any way you could activate it? Otherwise, is there any
> other way I could achieve this (without using ugly iframes)?

In summary I think the answer is no.  Sorry.  I don't know of any way
to accomplish what you are asking.  But there are several points that
your question raises.  Rather than simply saying, sigh, not possible,
let me address some of them.

* Being able to host more fancy web pages is nice in concept but that
we just don't have the resources to actually do it.  PHP might be
lightweight enough but Rails/Django are both rather heavy weight
servers and both require significant resources to keep updated against
security vulnerabilities.  Can you imagine if just 10% of the projects
hosted wanted to have a unique web service hosted?  That would be
completely overwhelming!

* The web pages are actually hosted on the FSF servers.  Those are
organizationally distinct from the Savannah site and team.  We the
Savannah Hackers don't have administrative access to them.  For
anything related to the web pages all administrative control is
through the FSF admins.

* Your mod_proxy wouldn't actually do what you wanted it to do.  You
would redirect away from black.nongnu.org and over to www.google.com
and then the user would be there.  Any bookmarks would see the new
site.  Effectively you wouldn't be using black.nongnu.org at all at
that point.  (And I don't think anyone would like the idea of
redirecting over to www.google.com.)

* It looks like you are trying to use a .htaccess file.  Those have a
bad reputation for slowing down servers.  If enabled they require the
server to make a large number of stat(2) calls for every page served.
They are never enabled by default.  Again this is up to the FSF admins
but personally I would never enable them.  You might want to read:
https://wiki.apache.org/httpd/Htaccess

And so I don't really have an answer that you will like.  I might
suggest that for some of the things you are asking you might try
setting them up for yourself.  Then you will see the problems of
trying to do it on a thousand projects.  Whatever is done must be able
to scale to a large number of projects and must be manageable for a
long time with minimal people resources.  It's a problem!

Bob

[savannah-help-public] [sr #108445] Savannah download statistics issues

[Bob Proulx]

Follow-up Comment #1, sr #108445 (project administration):

I don't know anything about awstats but I do know that it is spewing error
messages in the nightly run.  Here is the error.  Maybe showing this will
spark recognition of the problem from someone who has seen it before.  I
looked in the log file but the data looked okay to me.  If not then I will dig
into it.

Create/Update database for config "/etc/awstats/awstats.conf" by AWStats
version 6.95 (build 1.943)
>From data in log file "/var/log/apache2/access.log"...
Phase 1 : First bypass old records, searching new record...
Direct access after last parsed record (after line 285021)
AWStats did not find any valid log lines that match your LogFormat parameter,
in the 50th first non commented lines read of your log.
Your log file /var/log/apache2/access.log must have a bad format or LogFormat
parameter setup does not match this format.
Your AWStats LogFormat parameter is:
1
This means each line in your web server log file need to have "combined log
format" like this:
111.22.33.44 - - [10/Jan/2001:02:14:14 +0200] "GET / HTTP/1.1" 200 1234
"http://www.fromserver.com/from.htm"; "Mozilla/4.0 (compatible; MSIE 5.01;
Windows NT 5.0)"
And this is an example of records AWStats found in your log file (the record
number 50 in your log):
173.177.131.229 - - [18/Nov/2013:23:58:47 +] "-" 408 0 "-" "-"
Setup ('/etc/awstats/awstats.conf' file, web server or permissions) may be
wrong.
+http://www.google.com/bot.html)"
Setup ('/etc/awstats/awstats.conf' file, web server or permissions) may be
wrong.
Check config file, permissions and AWStats documentation (in 'docs'
directory).


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108446] git unpack-objects abnormal exit

[Bob Proulx]

Update of sr #108446 (project administration):

  Status:None => In Progress
 Assigned to:None => rwp

___

Follow-up Comment #1:

Hmm...  I don't immediately see the problem.  I was able to clone from it
okay.  I may need to add myself to the project and do a test push in order to
recreate it and debug it.  In the meantime I ran git-fsck against it.  It
didn't report any problems.  And then git-gc which also ran normally.  Please
try it again.  Maybe we will be lucky and this will have fixed it.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108445] Savannah download statistics issues

[Bob Proulx]

Follow-up Comment #2, sr #108445 (project administration):

Ivaylo, You said you noticed that awstats was offline.  Survey time.  Do you
regularly use awstats?  What information does it provide that you find useful?
 Because one of the proposals to fix it is simply to shut it down as being
unnecessary.  Would that cause any hardship?  Thanks.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108445] Savannah download statistics issues

[Bob Proulx]

Follow-up Comment #4, sr #108445 (project administration):

The stats on Savannah will not show the downloads from users using mirrors. 
That would seem to be a big disadvantage.

I won't be able to get to it immediately but I will look into the problem in
more detail and see about getting the awstats back online again.  It doesn't
seem like it should be too hard.  I will leave the assigned-to open so that if
someone wanted to jump in they should feel free to do so.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108451] GNU Emacs web repository embarrassingly slow

[Bob Proulx]

Update of sr #108451 (project administration):

  Status:None => Confirmed  
Operating System:None => GNU/Linux  

___

Follow-up Comment #1:

There has definitely been a performance problem for a long time. It has been
noticed.  It has been discussed.  (Try the savannah-help-pub...@gnu.org
mailing list, aka savannah-hackers@gnu.org list.)  But no one understands it
well enough to propose a solution.  Why are things slow?

The problem is suspected to be on the vcs VM.  The web site frontend
interfaces to the vcs backend.  (Could be on the frontend.)  The vcs VM is
configured for 7 cpus and 6G of ram.  It is mostly cpu idle and lots of ram. 
But at times it runs a very high load average up to 20, but still with an idle
cpu.  Theories have come and gone.  I could tell you my current theory is that
the xen vm is I/O slow.  Whenever it is doing a lot of I/O is when it seems to
really bog down.  But other theories have come and gone already so this one
might too.

I don't know anything about how BZR works so that is harder for me to debug. 
But CVS and SVN are similarly slow.  Dora has reported problems with those
months ago.  There is a cycle when it is worse than other times.  The worst
times seem to be around zero-dark-thirty for my timezone making it painful for
me to observe the problem first hand.  Unfortunately that is midmorning for
Dora and the folks over there.  Some folks have been experiencing it much
worse than others.

BTW I tried your sample URL and it came up within a couple of seconds.  No
problem at this moment.  (shrug.)  But I don't doubt that it has been bad
because we all have seen it be terrible before.  It must be an interaction of
things.

I have been trying to get dom0 access so that I can fix the grub boot process
to be reliable.  (It isn't set to upgrade well.)  And until then I am fearful
of reboots because it sometimes fails to come up and then it is a mad scramble
to find an FSF admin to rescue the system.  This request is in the pipeline
and at some point it will pop through.  Then it will be reboots and upgrades
all around.  From my viewpoint that is the next thing to do to it.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

Re: [savannah-help-public] Fwd: Issue : Stack unwinding and Frame pointer on Optimized Code in Linux x86_64.

[Bob Proulx]

Hello Pallav,

Pallav Singh wrote:
> Please provide pointers regarding the same. If you are not right person
> working on such issues, then please provide reference so that I can reach
> out to them regarding the same.

You have reached the Savannah mailing list.  Savannah is the software
forge for people committed to free software.  The web site, the
version control hosting, that type of thing.

> *Problem Statement :* We have stack unwinding code that relies on frame
> pointers. It not working in Optimized Code in Linux x86_64.

I think you are wanting the gcc mailing list.  Probably the gcc-help
mailing list.  Here is a pointer to their mailing lists.

  http://gcc.gnu.org/lists.html

Bob

[savannah-help-public] [sr #108457] where to find cookie in the email sent to confirm subscription

[Bob Proulx]

Update of sr #108457 (project administration):

  Status:None => Invalid
 Assigned to:None => rwp

___

Follow-up Comment #1:

The quagga-dev mailing list isn't hosted at gnu.org and the Savannah Hackers
do not administer it.  You will need to talk to the quagga admins.

Having said that the mailing list looks to be a normal Mailman mailing list
and should work the same.  It should be sending you the confirmation token by
email.  The simplest thing is to reply to the email and send that token back
by email.  It is a mailing list and if you can't receive and send email then
that is a good test that there is a problem.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108457] where to find cookie in the email sent to confirm subscription

[Bob Proulx]

Follow-up Comment #2, sr #108457 (project administration):

I could have mentioned that the owner address for mailing lists is usually the
address of the mailing list with -OWNER appended to it.  In this case
quagga-dev-ow...@lists.quagga.net is the address to write to for help from the
quagga-dev admins.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108457] where to find cookie in the email sent to confirm subscription

[Bob Proulx]

Update of sr #108457 (project administration):

 Open/Closed:Open => Closed 

___

Follow-up Comment #3:

Lastly since this was submitted by an anonymous submitter there isn't any way
to communicate this back to them.  Closing.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

Re: [savannah-help-public] please adjust the groff-commit mailing list

[Bob Proulx]

Hello Werner,

Werner LEMBERG wrote:
> right now, commits to the groff CVS repository are sent to the
> groff-commit list.  However, we've recently moved to git, so I ask you
> to change that, making git sending commits to the groff-commit list.
> 
> I've looked around, but it seems that Savannah doesn't provide an
> interface for this feature...

There isn't.  You have to ask for it on this mailing list (or put in a
support task in the tracker) and one of the Savannah Hackers will do
it for you.  It is easy to do by hand and it hasn't ever been
automated.

One question must be asked.  Some people want only the short commit
notifications and others want the notification with the full commit
diff in the email too.  Which would you prefer?

Bob

Re: [savannah-help-public] please adjust the groff-commit mailing list

[Bob Proulx]

Bob Proulx wrote:
> One question must be asked.  Some people want only the short commit
> notifications and others want the notification with the full commit
> diff in the email too.  Which would you prefer?

I have speculatively turned on the full commitdiff.  Let me know if
you need anything different.

Bob

Re: [savannah-help-public] please adjust the groff-commit mailing list

[Bob Proulx]

Werner LEMBERG wrote:
> > Some people want only the short commit notifications and others want
> > the notification with the full commit diff in the email too.  Which
> > would you prefer?
> 
> The latter, please.  Commits are not so frequent that this would
> overflow the list.

Done!  Of course I made the change blind by just copying in the
config.  You should see something with the next commit.  Please let me
know if it doesn't work right.  Or if you have anything else you want.

Enjoy!
Bob

Re: [savannah-help-public] please adjust the groff-commit mailing list

[Bob Proulx]

Bob Proulx wrote:
> Werner LEMBERG wrote:
> > > Some people want only the short commit notifications and others want
> > > the notification with the full commit diff in the email too.  Which
> > > would you prefer?
> > 
> > The latter, please.  Commits are not so frequent that this would
> > overflow the list.

I should have said it isn't the number of messages but the size of an
individual message.  If you have an extremely large commit then it
will generate an extremely large email.  Sometimes so large that MTAs
fail to deliver them.  Size limits.  Transport timeouts.  Various.
Personally I still like to see the full commitdiff with the
understanding of the limitations.  It encourages more smaller commits
over the one large super huge commit.

> Done!  Of course I made the change blind by just copying in the
> config.  You should see something with the next commit.  Please let me
> know if it doesn't work right.  Or if you have anything else you want.

I should have also asked if you want me to disable commits to the
previous CVS repository?  That way any commits there will fail back
with a message letting someone know that they are pushing to the wrong
location?  It is a CVSROOT/commitinfo entry:

  ALL /usr/local/bin/commit_obsolete_notice %r/%p %s

To print out:

  echo "This repository is obsolete and marked read-only to prevent commits."
  echo "The sources have been moved to another repository."
  echo "Please see the package web pages and documentation for more 
information."
  exit 1

Sorry for the too many exchanges.  I didn't think of everything all at
once.  I should update a wiki page with a checklist for these details.

Bob

Re: [savannah-help-public] Slowness of Savannah in the morning

[Bob Proulx]

Karl Berry wrote:
> I added an iotop invocation to vcs:/etc/cron.d/sv, at the same time as
> top, to see if it gets us anything useful.  Also nice-d all the relevant
> cron jobs.

Thanks.  I think your theory about being I/O bound is probably
correct.  But in that case I am not sure what to do about it.

> There was a load average of 10+ on both Saturday and Sunday from 0700
> though 0800 (more or less), so clearly the backup job change was not a
> panacea (as expected).  /var/tmp/toplog shows the usual array of git,
> bzr, cvs, etc., processes, mostly niced, mostly in disk wait.  The
> backup job doesn't even show up any more.

Perhaps it is time to ask for another VM to split the workload?
Because it would be nice to be able to do A-B comparisons between good
and bad until able to zero in on the problem.

If we started moving tasks from the current VM onto a different one
then eventually we would move the "heavy" one, the problem would
follow it, and then point to that task as being the heavy problem one
that needs more power.  I think all of the tasks are tied to their
name.  That is git.sv.gnu.org and bzr.sv.gnu.org and so forth.  It
should be trivial to move individual services from one place to
another.

However if the host system is I/O bound and all of the VMs are on the
same host system then I doubt this would change anything.  But it
might prove that the problem isn't in the VM but in the host system.
I am somewhat grasping at things because there just doesn't seem to be
any obvious pointer to the problem.

> Sigh.

Agreed.

Bob

Re: [savannah-help-public] Slowness of Savannah in the morning

[Bob Proulx]

Karl Berry wrote:
> I added an iotop invocation to vcs:/etc/cron.d/sv, at the same time as
> top, to see if it gets us anything useful.  Also nice-d all the relevant
> cron jobs.

That was a good idea.  But it doesn't seem to be working without a tty
device and generates a python backtrace.  (I really hate python's
proclivity for puking guts all over the place instead of a proper
error message.)

I fiddled with it and changed it to this set of options which seems to
provide some useful details.  Nothing unusual showing up yet though.

  iotop -b -o -n 5

I will also note that vcs is acting really slow right now to my
interactivity.  It is running a load of 11.  But the 7 cpus look
mostly idle.  Every so often one cpus shows some activity, every so
often two show activity, and then everything goes back to idle.

Being on the system now it just seems very slugish.  Doing simple
things like running "cat /var/tmp/iotoplog" for a 1k file takes 20
seconds to complete.

Bob

[savannah-help-public] generate_log_accum.pl ?? What does it do?

[Bob Proulx]

Savannah Hackers,

I see generate_log_accum.pl using up a lot of I/O and cpu time.  It is
set to run by cron every 20 minutes.  The script header says only:

  Generates all projects CVSROOT/commitinfo and CVSROOT/loginfo.

Of course I know about commitinfo and loginfo.  A better descriptive
above might be "edit" intead of "generate" but that still doesn't say
what it does when it edits those files.  I can't see any editing being
done on those files.  They seem pretty static to me.  Except all of
the files for every project are getting updated every 20 minutes.
That is 2299 projects and files updated every time it runs.

I browsed the script very briefly but the intent was not obvious.
Does anyone already know what this script is doing?

I did a quick analysis and there are 2007 identical commitinfo files
empty except for  tags, 155 identical except for  tags
plus one blank line, 14 identical that are the default CVS upstream
version of the file, and 122 uniquely different commitinfo files.

Thanks,
Bob

[savannah-help-public] How to coordinate a VM reboot?

[Bob Proulx]

Savannah Hackers,

Everyone wants to reboot vcs.sv since it has been so long.  It has
been up 514 days.  I want to fsck the file system since it has been
three years since the last check!  Perhaps the reboot will help with
the performance problems.

The file system is 120G.  How long will an fsck of the 120G file
system take the VM to complete?  I don't know but could take more than
30 minutes.  Could take 60 minutes.  Could take 10 minutes.  Unknown
until it is done.  I will guess 20 minutes, possibly 40 minutes if the
system is slow.  Or maybe stated as 30 minutes +/- 10 minutes.

I now have console access and can monitor fsck progress which is a
very good thing if needed.  Can also monitor and debug the grub
problems previously seen too.  I was too scared to try this in the
blind without access since I already had one VM fail to boot and need
to be rescued.

If vcs.sv is offline for possibly 60 minutes that is a very long
time.  How should this be communicated to the users?  Ideas?  Should
we just steel ourselves and simply reboot it and fsck and then
apologize for any problem reports due to it?  Ideas?

Thanks,
Bob

Re: [savannah-help-public] Slowness of Savannah in the morning

[Bob Proulx]

Karl Berry wrote:
>   iotop -b -o -n 5
> 
> Thanks.  I think -n 1 is better; -n 5 is too much information.

That is fine.  I looked at -n1 and it looked like it may have been
missing processes that were listed in one of the subsequent passes.
That is why I added another few iterations to save.  But to be honest
I was simply trying to keep moving and wasn't looking too closely.

> But the 7 cpus look mostly idle.  
> 
> I think it is almost certain that the underlying problem is disk I/O.
> That's been clear to me for quite a while, since the top processes are
> almost always in disk wait, as I said.

I now see that the dom0 is a 8-cpu host but is hosting 24 VMs.  It
isn't overbooked on ram as far as I can tell.  But if several of those
VMs all became active at around the same time then that would saturate
the underlying host hardware.  I/O saturation would line up with your
correlation of disk I/O being problematic.  But we really need to
look at the entire set of VMs across the underlying dom0 host.  Having
visibility into some of the individual VMs has definitely made
diagnosing this more difficult.

> What's not clear is how to resolve it, given that updating hardware
> is a massive, expensive, and unlikely-to-happen project.

Agreed.  If all things were possible I would suggest adding more
hardware.  Now that I know it is one 8-cpu host I think we are
probably overloading the available hardware resource.  We should move
services off of the current vcs.sv VM and onto different additional
hardware.

I will just stop there for the moment because I agree that does
escalate into a serious project.  Because I wouldn't want to add one
single machine.  I would want to add at least two such that there is
redundant hardware in the case of a failure.  It would need to be
thought about in order to support this much infrastructure.

Bob

Re: [savannah-help-public] please adjust the groff-commit mailing list

[Bob Proulx]

Werner LEMBERG wrote:
> > Shall I really subscribe `invalid.nore...@gnu.org' to the
> > bug-groff mailing list?
> > 
> > You don't need to (and shouldn't) subscribe that address.  You just
> > need to add it to accept_these_nonmembers on
> > https://lists.gnu.org/mailman/admin/bug-groff/privacy/sender.
> 
> I've added `invalid.nore...@gnu.org' accordingly, thanks.

Just an an oh, by the by, comment...

If you hadn't done anything then eventually someone would make a
commit.  That would generate an email.  The first email would
initially be held in the Mailman hold queue.  One of the listhelpers
(likely myself or Karl) would see it and we would both approve the
message and as standard operating procedure add the message's address
to the above accept_these_nonmembers list.  The Mailman "summary" web
page has a checkbox for doing so making it easy.  I would normally
check that box and have the web interface do this for me when I
approved the message.  (After having looked at the contents of the
message in the "details" page.)  At that point subsequent messages
would pass through without any delay.  It is only the initial message
that would be held.

Just to show that for the most part for most projects things will get
taken care of without needing a lot of tedious administrivia from each
project maintainer.  Very happy that you have such a good hands-on
status for your projects.  Most do not and so we mother over things as
a general umbrella.  When it is cold outside we try to make sure
people have their warm sweaters on, and that type of thing.  :-)

I am sure from your perspective there may seem like this is a lot of
administrivia work that you needed to do.  But you really didn't need
to do it.  From our perspective you did a lot of work that would have
happened through the normal course of events the same as for the other
thousand plus mailing lists.  But I don't want to discourage you from
doing any of these things.  First hand experience and understanding
how things are working are what enable us to do other more interesting
things.  I just don't want you to think the drudgery is required of you.

Just thought I would try to give a view to what happens as a normal
routine.  :-)

Bob

Re: [savannah-help-public] please adjust the groff-commit mailing list

[Bob Proulx]

Werner LEMBERG wrote:
> > You asked about making the CVS frozen message include an explicit
> > url.  Anything can be done, but time is fleeting.  The issue is that
> > it's extra work to figure out what that url is.  savannah.gnu.org
> > vs. nongnu.org ...
> 
> :-) Ideally, the message should be editable with the web interface,
> but Bob explained that the Savannah infrastructure system itself is
> orphaned (sort of), so please forget my request for now.

I had put it on the back burner because I wanted to research the
available format options.

> BTW, can the message be slightly refined to directly
> give the project's URL?

The CVS commitinfo has by default "%r/%p %s" but other options are
possible too.

  ALL /usr/local/bin/commit_obsolete_notice %r/%p %s

What are %r/%p %s?  What are the other possible set of format options?
I forget!  I need to go look at the CVS docs for it.  Is the
information needed to for a correct URL in those options?  These seem
to be the useful options of the available ones.

 p
  The name of the directory being operated on within the repository.
 r
  The name of the repository (the path portion of '$CVSROOT').
 s
  a list of the names of files to be committed

So we can get the project name out of it.  But then as Karl mentioned
would need to map that into gnu versus nongnu.

Like Karl I think that anyone who was a project committer but then got
a message that the repository had moved on would be able to figure
things out from there.  How much trouble could it be for someone who
was a project committer to be able to find the project's home page?
Would a project committer really be unaware of a repository change?

I think most likely they would be fully aware but just accidentally
use an old working copy that they had forgotten to convert.  In which
case the simple message should be enough to tell them what happened
and clue them in to moving to the new repository.

Bob

Re: [savannah-help-public] How to coordinate a VM reboot?

[Bob Proulx]

Hi Karl,

Karl Berry wrote:
> If vcs.sv is offline for possibly 60 minutes that is a very long
> time.  How should this be communicated to the users?
> 
> Please choose a time convenient for you a couple days from now and we
> can post a news item.  Also email savannah-annou...@gnu.org and
> savannah-us...@gnu.org.  Christmas day, boxing day?  I think it is good
> to tell people in advance.

Your suggestion of Boxing Day, the day after Christmas, seems good to
me.  I will put together an announcement and send it out as you
suggest as soon as I can type one in after this message.

I will also ask the FSF admins to post something about it to their
status feed on https://pumprock.net/fsfstatus too.  It is an FSF admin
feed but I think they would add something there for us.

> Wdyt about rebooting download at the same time, to "parallelize" the
> potentially long fsck's?  Pros and cons ...

Yes!  If things go smoothly then let's roll through the cluster and
reboot and fsck the root file systems on all of them.  Good idea.

> download$ df -BG
> Filesystem 1073741824-blocks  Used Available Capacity Mounted on
> /dev/xvda2  197G  150G   48G  77% /

Yes.  That is another big one.  Since it is on the same dom0 I don't
think it would be a good idea to do it at the exact same time.  I am
not sure what the number of mechanisms are on the disk farm but likely
each will individually saturate the bus.  In which case doing more at
one time in parallel will take the total as twice as long.  But if one
goes smoothly and with reasonable time then let's roll on to the next
and do it sequentially after it.

> download$ uptime
>  22:42:19 up 616 days, 18:17,  2 users,  load average: 0.00, 0.00, 0.00

That is a long time.  And of course it means that the kernel has not
been upgraded in that time either.  It will be very good to reset the
clock on everything.

Bob

[savannah-help-public] [sr #108460] ease.js: git push --force

[Bob Proulx]

Update of sr #108460 (project administration):

  Status:None => Need Info  
 Assigned to:None => rwp

___

Follow-up Comment #2:

As you know once you publish source code it is then published forever.  The
general rule for public master branches is never to rewind public master
branches.  We are especially cautious about doing so on Savannah because
people have tried to suppress previously released source code in this way. 
This is specifically disabled due to this reason.  Therefore I never simply
remove source code lightly.  I always investigate in detail.

I cloned both projects and compared them.  I see three commits to master on
Savannah that are identical to what are on gitorious but with different commit
dates and therefore different commit hashes forcing a fork in the version
history.  I can see that isn't a real difference and therefore I have gone
ahead and rewound the head on Savannah to the last common commit between the
two repositories.  You should be able to push your new version history to
Savannah on top of this.

This is clearly bad for anyone that has already pulled a copy from Savannah
previously.  Rewinding public branches is a bad thing.

I noticed that the project description for the web page was the default
template.  I updated it.

Please let me know that you have uploaded your version history successfully
and I will close out this ticket.  I branched master over to master.mistake
and the same for website.  I will clean up by removing master.mistake when I
know that all is good.  Let me know if there is anything else that needs to be
done.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108460] ease.js: git push --force

[Bob Proulx]

Update of sr #108460 (project administration):

  Status:   Need Info => Done   
 Open/Closed:Open => Closed 

___

Follow-up Comment #4:

Yay!  Good to hear that everything is good now.  I removed the master.mistake
and website.mistake safety copies I made so that things should be all normal
now.

I may have come on a little strong.  Let me soften that up quite a bit.  I can
see you understand and have the right attitude.  That was all I needed to
hear.  Please don't get shy about doing things that need to be done pushing
commits and that kind of things for fear of messing something up.  Stuff
happens.  Everyone understands that and it isn't a big deal.  I have been on
the other end of things as recently as today when I seriously broke something
and needed a rescue!  We are all in this together.

So please keep up the good work and keep pushing updates and doing whatever
needs to be done.  Life is good!  If you need anything please feel free to
contact us either with a support request or with an email to
savannah-help-pub...@gnu.org either way.  Things only work when we all work
together.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

Re: [savannah-help-public] How to coordinate a VM reboot?

[Bob Proulx]

The VMs notably vcs and download were rebooted.  All is back up and
online afterward.

Things did not go completely smoothly.  Here is the post mortem of
today's operation.

No one was available who could update the FSF status location
https://pumprock.net/fsfstatus for us.  Hopefully that will get
resolved before next time as having an out of band status is useful.

The boot of vcs took twenty minutes to clear /tmp.  I had previously
captured a du listing of /tmp and discovered 1,377,775,616 bytes in
113,283 files there.  Mostly this resided in /tmp/loggerhead-cache-*
and some more in /tmp/cvs-srv12738 directories.  It appears that the
loggerhead cache is not getting pruned.  This should be looked into in
more detail.  Everything is reset fresh now but it will grow again.

To be honest I had forgotten about the /tmp disk space and didn't
figure it into the timeline.  At that point I just needed to wait out
the /tmp clearing which was taking a while since there were so many
files.

I hit a snag with the Xen captured management interface.  While
booting frontend I was looking at the grub boot screen and examining
the configuration.  Then a python backtrace appeared.  There isn't any
python in grub itself and this backtrace was coming from xen.  At that
point Xen had lost track of the frontend and could not control it.
Xen reported errors that it could not open the config file
frontend.savannah.gnu.org and then gave me an 'xm' usage dump.  I then
had no control over the VM and could not boot it.  Argh!  Note that I
only have access to xen-shell and not shell access to the dom0.

In theory there is no difference between theory and practice.  In
practice there is a difference.  In theory the xen-shell is a captured
interface and should prevent this type of breakage.  In practice I
broke it without trying.  It was a very stressful time!

Many thanks to "nully" our newest FSF admin who was in the hot seat on
site today.  She was able to rescue the configuration file and get xen
back online again.  With that was able to boot the system again.  This
excursion took us almost an hour to figure out before we got things
going again.  Therefore I decided to stop at that point, declare
victory, and withdraw to safety.  No fsck's were done.  However the
very long running vcs and download systems were rebooted which was the
major goal of the day.  That was accomplished successfully.

Savannah is back in normal service.

Bob

[savannah-help-public] [sr #108463] How is write access to SVN/git controlled

[Bob Proulx]

Follow-up Comment #1, sr #108463 (project administration):

I have been wondering about this myself.  Which may seem strange that I would
reply in that case.  But since I don't think anyone else will reply for
perhaps a long time I thought I might as well say what I knew regardless.

As far as I know (could be wrong) any project member may commit to the source
code of any project for which they are a member.

The implementation uses a local PAM plugin that implements user accounts
through mysql.  Using mysql the database is manipulated through the PHP code
of the Savane code base that runs Savannah.  On the vcs.sv.gnu.org machine the
/etc/nsswitch.conf file lists "passwd: compat mysql".  The "mysql" addition
allows accounts to exist in the shared msyql database tables in addition to
the machine local /etc/passwd location.  Projects name Unix account groups. 
The effect is that users will have additional groups for every project that
they are a member.  Looking at 'id' I see that you are in the group nano for
example.

The directories of /srv/git and /srv/svn and others that store the vcs
backends are group writable by the project group.  Therefore commits are
allowed if they have write permission to the directory.  Version commits that
write to the file system will be allowed if they are a member of the group. 
If the user is not a member of the group then write access is denied.

I see nothing that differentiates between users who are a member of a group
and users that are a member of the group and also have project admin status. 
It appears to be purely group membership.

I report the above based upon reverse engineering what I see on the system.  I
could easily be wrong.

Just FYI but for these types of discussion questions (instead of tasks to be
tracked) I prefer using the mailing lists over the web forum.  On the mailing
list I get to use a real editor instead of a browser form.  Your preferences
may be quite different form mine however and I know many people prefer the web
browser forms.  The savannah-help-pub...@gnu.org mailing list (also known as
savannah-hackers@gnu.org) would be a perfect place for the discussion. 
https://lists.gnu.org/mailman/listinfo/savannah-hackers

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

Re: [savannah-help-public] GSL: switching from bzr to git

[Bob Proulx]

Hi Patrick,

Patrick Alken wrote:
>   I was hoping I could trouble you to again make HEAD point to a
> different branch by default for GSL. We would like our 'develop-1'
> branch to be the default HEAD. I think the command will be:

The problem is that you would be rewinding a publicly published master
branch.  That is always a bad thing to do.  Anyone who has cloned that
branch will be broken.  Because it is the master branch it will be the
default one cloned.  The master branch is one that is expected to have
an infinite lifetime.  They will need to take manual steps to fix the
broken clone.  Having run continuous integration builds and being an
advocate for CI builds I am sensitive that it will also break them too
for the same reason.

Sometimes mistakes are made and so we do things to correct mistakes.
That is just life.  I am happy to help fix something that is broken.
Previously when you had a broken master branch I was happy to correct
it.  But you are proposing to break it every time you make a release
as a standard release strategy.

I think you should use a different branch release strategy.  Many
others have been this way many times before.  Instead of trying to
invent a new process and tripping over the snags along the way I would
try to learn from the mistakes of others.  There are endless
discussions of release branch models.  I recommend that you choose one
of the many strategies already available.  That way the path is
already smoothed out.

Here is an overview of git specific workflows.  It does a good job of
documention possible branching work flows using git.  (Some branches
of which are routinely expected to be rewound.)

  https://www.kernel.org/pub/software/scm/git/docs/gitworkflows.html

I did a quick web search and found this reference.  I can't say as I
worked through the article in enough detail to know if I agree with it
completely but on first pass it sounded like one of the reasonable and
nicely written ones.

  http://nvie.com/posts/a-successful-git-branching-model/

I would follow one of the above branch and release strategies as a
known best practice.  It will avoid a lot of pain for others.

> Also I'd like to ask for this as a feature request in the web interface
> - the GSL team prefers the default branch to be associated with a major
> version number (ie: v1, v2, etc) so for each major release we'll
> probably want to update the default HEAD. I know github has a pull down
> menu on their web interface which allows users to update the remote HEAD
> settings (like the command above). Perhaps savannah could do something
> similar?

I personally would not want to encourage this behavior.

Bob

Re: [savannah-help-public] Delete Bazaar branches

[Bob Proulx]

Sam Geeraerts wrote:
> We mistakenly created 2 Bazaar branches, i.e.:
> 
> bzr://bzr.savannah.gnu.org/gnewsense/packages-parkes/debderiver
> bzr://bzr.savannah.gnu.org/gnewsense/packages-parkes/freetype
> 
> These are duplicates of
> bzr://bzr.savannah.gnu.org/gnewsense/debderiver. Could someone please
> delete them for us?

If no one else jumps in first I will look into doing this a little
later today.  Not being a bzr user I will first need to learn how to
use it.  ;-)

Bob

Re: [savannah-help-public] Delete Bazaar branches

[Bob Proulx]

Sam Geeraerts wrote:
> As far as I know, deleting the debderiver and freetype diretories
> should do the trick. Be sure to pick the right debderiver directory. :)

Hmm...  Okay.  I can certainly do that!  :-)  (Safely.  I will move
them to a trashcan first.)

  mv gnewsense/packages-parkes/debderiver gnewsense/packages-parkes/freetype 
/var/tmp/gnewsense/

Take a look at things now.  Good?  Bad?

Bob

Re: [savannah-help-public] Delete Bazaar branches

[Bob Proulx]

Sam Geeraerts wrote:
> > Take a look at things now.  Good?  Bad?
> 
> Looks good from here. Thanks.

Great!  Let us know if you need anything else.

Bob

Re: [savannah-help-public] Delete Bazaar branches

[Bob Proulx]

Sam Geeraerts wrote:
> Op Thu, 2 Jan 2014 11:23:05 -0700
> schreef Bob Proulx :
> 
> > Great!  Let us know if you need anything else.
> 
> Sure thing. FYI, we're reusing the freetype location. You can
> leave it alone now. :)

Good deal.  I am done until or unless you submit a new request. :-)

Bob

[savannah-help-public] [sr #108446] git unpack-objects abnormal exit

[Bob Proulx]

Update of sr #108446 (project administration):

  Status: In Progress => Done   
 Open/Closed:Open => Closed 

___

Follow-up Comment #5:

This report was left for way too long. Sorry about that.  I assume after the
last exchange that the problem has been long resolved and so will close the
report.  If not then please let me know.

Just to tidy up some previous information the version of git on the server is
from Debian Squeeze which was Stable.  It is past due to be upgraded to Wheezy
but will be upgraded soon.  The Squeeze version of git is 1.7.2.5 so going
newer isn't the right direction.  The Wheezy version of git is 1.7.10.4.

Also if there has been long term server performance issues during some times
of the day.  When reporting service problems please note the time of day when
the problem is happening.  Some times of day things are great.  Some times of
day the machine is overloaded with a high load average and maximum I/O.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108467] Please change GNU ELPA post receive hook to new multimail email hook

[Bob Proulx]

Update of sr #108467 (project administration):

  Status:None => In Progress
 Assigned to:None => rwp


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108467] Please change GNU ELPA post receive hook to new multimail email hook

[Bob Proulx]

Follow-up Comment #1, sr #108467 (project administration):

Done.  Sorry for the delay.  I needed to do some testing of the new hook
script before I set it up live.  Let me know what you think of the new format.
 I will leave the ticket open until then.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

Re: [savannah-help-public] [sr #108451] GNU Emacs web repository embarrassingly slow

[Bob Proulx]

Paul Eggert wrote:
> Dmitry Gutov wrote:
> >Follow-up Comment #3, sr #108451 (project administration):
> >>They both responded pretty fast ( < 5 sec).
> >
> >Same here.
> 
> It's < 5 sec for me now too.  Thanks to whoever improved
> the performance.

It is just a time of day thing.  There haven't been any significant
changes to address the problem.  I believe if this were tested at
another time of day the problem would be visible then.

Bob

[savannah-help-public] [sr #108467] Please change GNU ELPA post receive hook to new multimail email hook

[Bob Proulx]

Follow-up Comment #4, sr #108467 (project administration):

Drat!  I will look at it.  (It was fine for me in a test on another system.)

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108467] Please change GNU ELPA post receive hook to new multimail email hook

[Bob Proulx]

Follow-up Comment #5, sr #108467 (project administration):

I believe this should be working now.  I re-triggered this commit through the
post-receive hook.  Please let me know about any details that should be
changed.  There are several configuration options available for tweaking.

Also I set the from address to be the emacs-devel mailing list.  Please let me
know if that should be a different mailing list.


___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108467] Please change GNU ELPA post receive hook to new multimail email hook

[Bob Proulx]

Follow-up Comment #6, sr #108467 (project administration):

Looking at the result I think setting the from address is not as good as
letting it default to the committer email as it had before.  I removed that
setting.  The next commit will default to the submitter as it had before.

Initially I was thinking that setting it would allow the email through the
mailing list through the whitelisted email without further approval.  But that
really isn't much of a problem so shouldn't be solved.  And it hides the
committer in the email summary of the commits.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/

[savannah-help-public] [sr #108471] installing octave causes numpy to fail when running python

[Bob Proulx]

Follow-up Comment #1, sr #108471 (project administration):

You appear to be having a problem with octave.  But you have submitted this
report against savannah administration.  I am reassigning this ticket to the
octave project.

___

Reply to this item at:

  

___
  Message sent via/by Savannah
  http://savannah.gnu.org/