Mirror of the gdb mailing list
 help / color / mirror / Atom feed
* Changes to snapshot process ... ?
@ 2001-10-01 17:21 Andrew Cagney
  2001-10-02  1:14 ` Eli Zaretskii
  0 siblings, 1 reply; 10+ messages in thread
From: Andrew Cagney @ 2001-10-01 17:21 UTC (permalink / raw)
  To: gdb

Hello,

I'd like to propose a change to the way sanpshots are being created.  At 
present they are being drawn strictly from the 5.1 branch.  Instead, I'd 
like to suggest running both trunk and branch snapshots.  The process 
would alternate between creating one or the other.

Andrew


^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: Changes to snapshot process ... ?
  2001-10-01 17:21 Changes to snapshot process ... ? Andrew Cagney
@ 2001-10-02  1:14 ` Eli Zaretskii
  2001-10-02  8:58   ` Andrew Cagney
  0 siblings, 1 reply; 10+ messages in thread
From: Eli Zaretskii @ 2001-10-02  1:14 UTC (permalink / raw)
  To: ac131313; +Cc: gdb

> Date: Mon, 01 Oct 2001 20:21:08 -0400
> From: Andrew Cagney <ac131313@cygnus.com>
> 
> I'd like to propose a change to the way sanpshots are being created.  At 
> present they are being drawn strictly from the 5.1 branch.  Instead, I'd 
> like to suggest running both trunk and branch snapshots.  The process 
> would alternate between creating one or the other.

Any specific reason?  I rather like the way it works now.


^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: Changes to snapshot process ... ?
  2001-10-02  1:14 ` Eli Zaretskii
@ 2001-10-02  8:58   ` Andrew Cagney
  2001-10-02 10:07     ` Eli Zaretskii
  0 siblings, 1 reply; 10+ messages in thread
From: Andrew Cagney @ 2001-10-02  8:58 UTC (permalink / raw)
  To: Eli Zaretskii; +Cc: gdb

>> Date: Mon, 01 Oct 2001 20:21:08 -0400
>> From: Andrew Cagney <ac131313@cygnus.com>
>> 
>> I'd like to propose a change to the way sanpshots are being created.  At 
>> present they are being drawn strictly from the 5.1 branch.  Instead, I'd 
>> like to suggest running both trunk and branch snapshots.  The process 
>> would alternate between creating one or the other.
> 
> 
> Any specific reason?  I rather like the way it works now.

For following both the trunk and the branch:

So that both the current branch and the current trunk get put through 
the release process.  That includes things like:

	configury et.al. checks
	ARI checks (including DOS file name checks :-)
	documentation generation

as well as ending up with snaps and diffs.

Assuming I do get 5.1 (HP/UX, true64, s390) resolved I'd like to 
continue making the head of 5.1 available while also publishing current 
snaps.

For alternating:

That is pretty arbitrary.  The objective is to constrain the volume of 
raw data that shadow sites have to download every night.

An alternative would be to cut back on the snapshot frequency (weekly 
only) but maintain the nightly and weekly diffs.  (Although, to be 
honest, that is more tricky.)

Andrew



^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: Changes to snapshot process ... ?
  2001-10-02  8:58   ` Andrew Cagney
@ 2001-10-02 10:07     ` Eli Zaretskii
  2001-10-10 21:24       ` Andrew Cagney
  0 siblings, 1 reply; 10+ messages in thread
From: Eli Zaretskii @ 2001-10-02 10:07 UTC (permalink / raw)
  To: ac131313; +Cc: gdb

> Date: Tue, 02 Oct 2001 11:58:12 -0400
> From: Andrew Cagney <ac131313@cygnus.com>
> 
> For alternating:
> 
> That is pretty arbitrary.  The objective is to constrain the volume of 
> raw data that shadow sites have to download every night.
> 
> An alternative would be to cut back on the snapshot frequency (weekly 
> only) but maintain the nightly and weekly diffs.  (Although, to be 
> honest, that is more tricky.)

Suppose someone only wants the branch snapshots: can she easily find
out which snapshot is which?  If that's not easy, how about putting
each type of snapshot into its own directory, to avoid confusion?


^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: Changes to snapshot process ... ?
  2001-10-02 10:07     ` Eli Zaretskii
@ 2001-10-10 21:24       ` Andrew Cagney
  2001-10-11  1:06         ` Eli Zaretskii
  0 siblings, 1 reply; 10+ messages in thread
From: Andrew Cagney @ 2001-10-10 21:24 UTC (permalink / raw)
  To: Eli Zaretskii; +Cc: gdb

>> Date: Tue, 02 Oct 2001 11:58:12 -0400
>> From: Andrew Cagney <ac131313@cygnus.com>
>> 
>> For alternating:
>> 
>> That is pretty arbitrary.  The objective is to constrain the volume of 
>> raw data that shadow sites have to download every night.
>> 
>> An alternative would be to cut back on the snapshot frequency (weekly 
>> only) but maintain the nightly and weekly diffs.  (Although, to be 
>> honest, that is more tricky.)
> 
> 
> Suppose someone only wants the branch snapshots: can she easily find
> out which snapshot is which?  If that's not easy, how about putting
> each type of snapshot into its own directory, to avoid confusion?

Yes, agreed.  You wouldn't have a suggestion for a directory name?  One 
would be gdb/snapshots/current/, as for the other.  One thought would be 
gdb/snapshots/stable/ however people may read more into that then they 
should.  I guess I should suggest gdb/snapshots/branch/ and see if 
someone comes up with something better :-)

Andrew



^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: Changes to snapshot process ... ?
  2001-10-10 21:24       ` Andrew Cagney
@ 2001-10-11  1:06         ` Eli Zaretskii
  2001-10-12 12:53           ` Andrew Cagney
  0 siblings, 1 reply; 10+ messages in thread
From: Eli Zaretskii @ 2001-10-11  1:06 UTC (permalink / raw)
  To: Andrew Cagney; +Cc: gdb

On Thu, 11 Oct 2001, Andrew Cagney wrote:
> > Suppose someone only wants the branch snapshots: can she easily find
> > out which snapshot is which?  If that's not easy, how about putting
> > each type of snapshot into its own directory, to avoid confusion?
> 
> Yes, agreed.  You wouldn't have a suggestion for a directory name?  One 
> would be gdb/snapshots/current/, as for the other.  One thought would be 
> gdb/snapshots/stable/ however people may read more into that then they 
> should.  I guess I should suggest gdb/snapshots/branch/ and see if 
> someone comes up with something better :-)

Yes, gdb/snapshots/branch/ sounds good.  Better yet, try something
like gdb/snapshots/gdb-5.1-release-branch/.


^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: Changes to snapshot process ... ?
  2001-10-11  1:06         ` Eli Zaretskii
@ 2001-10-12 12:53           ` Andrew Cagney
  2001-10-12 13:22             ` Christopher Faylor
  0 siblings, 1 reply; 10+ messages in thread
From: Andrew Cagney @ 2001-10-12 12:53 UTC (permalink / raw)
  To: Eli Zaretskii, gdb

> On Thu, 11 Oct 2001, Andrew Cagney wrote:
> 
>> > Suppose someone only wants the branch snapshots: can she easily find
>> > out which snapshot is which?  If that's not easy, how about putting
>> > each type of snapshot into its own directory, to avoid confusion?
> 
>> 
>> Yes, agreed.  You wouldn't have a suggestion for a directory name?  One 
>> would be gdb/snapshots/current/, as for the other.  One thought would be 
>> gdb/snapshots/stable/ however people may read more into that then they 
>> should.  I guess I should suggest gdb/snapshots/branch/ and see if 
>> someone comes up with something better  [:-)] 
> 
> 
> Yes, gdb/snapshots/branch/ sounds good.  Better yet, try something
> like gdb/snapshots/gdb-5.1-release-branch/.

I had some reservations over calling something 5.1 but I think I've 
overcome that fear. So gdb/snapshots/5.1-branch (I think release and gdb 
are redundant).

I'll tweek the scripts for this ready for tonight.

--

Next question is gdb/snapshots/gdb+dejagnu.tar.bz2.

Hmm, at present it is a hard link :-/  If it was a symlink would ftp 
shadow systems be smart enough to copy the link and not the file?

Eliminate this and GDB can run nightly snaps of both the 5.1-branch and 
current without a significant increase in shadow size.

Andrew



^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: Changes to snapshot process ... ?
  2001-10-12 12:53           ` Andrew Cagney
@ 2001-10-12 13:22             ` Christopher Faylor
  2001-10-13 10:25               ` Andrew Cagney
  0 siblings, 1 reply; 10+ messages in thread
From: Christopher Faylor @ 2001-10-12 13:22 UTC (permalink / raw)
  To: gdb

On Fri, Oct 12, 2001 at 03:53:22PM -0400, Andrew Cagney wrote:
>Next question is gdb/snapshots/gdb+dejagnu.tar.bz2.
>
>Hmm, at present it is a hard link :-/  If it was a symlink would ftp 
>shadow systems be smart enough to copy the link and not the file?

Yes.  I have a symbolic link in the cygwin directory that seems to be
faithfully mirrored and there is a 'egcs -> gcc' link elsewhere that
has been around for a long time.  It also is mirrored correctly on
the sites that I looked at.

cgf


^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: Changes to snapshot process ... ?
  2001-10-12 13:22             ` Christopher Faylor
@ 2001-10-13 10:25               ` Andrew Cagney
  2001-10-13 10:53                 ` Christopher Faylor
  0 siblings, 1 reply; 10+ messages in thread
From: Andrew Cagney @ 2001-10-13 10:25 UTC (permalink / raw)
  To: Christopher Faylor; +Cc: gdb

> On Fri, Oct 12, 2001 at 03:53:22PM -0400, Andrew Cagney wrote:
> 
>>Next question is gdb/snapshots/gdb+dejagnu.tar.bz2.
>>
>>Hmm, at present it is a hard link :-/  If it was a symlink would ftp 
>>shadow systems be smart enough to copy the link and not the file?
> 
> 
> Yes.  I have a symbolic link in the cygwin directory that seems to be
> faithfully mirrored and there is a 'egcs -> gcc' link elsewhere that
> has been around for a long time.  It also is mirrored correctly on
> the sites that I looked at.

Nice, if you look you'll notice that I used softlinks recklessly.  The 
snapshots will slowly flush them.

I have one tweek left.  For the 5.1-branch, change the version to 
``5.0.91_YYYYMMDD'' so that it is different to the trunk (YYYYMMDD).

thanks,
Andrew


PS: http://sources.redhat.com/gdb/5.1/{onlinedocs,ari} are for the 
branch while http://sources.redhat.com/gdb/{ari,onlinedocs} are for 
current.  I might one day move them to /gdb/current/... but not today.



^ permalink raw reply	[flat|nested] 10+ messages in thread

* Re: Changes to snapshot process ... ?
  2001-10-13 10:25               ` Andrew Cagney
@ 2001-10-13 10:53                 ` Christopher Faylor
  0 siblings, 0 replies; 10+ messages in thread
From: Christopher Faylor @ 2001-10-13 10:53 UTC (permalink / raw)
  To: gdb

On Sat, Oct 13, 2001 at 01:25:19PM -0400, Andrew Cagney wrote:
>>On Fri, Oct 12, 2001 at 03:53:22PM -0400, Andrew Cagney wrote:
>>
>>>Next question is gdb/snapshots/gdb+dejagnu.tar.bz2.
>>>
>>>Hmm, at present it is a hard link :-/  If it was a symlink would ftp 
>>>shadow systems be smart enough to copy the link and not the file?
>>
>>
>>Yes.  I have a symbolic link in the cygwin directory that seems to be
>>faithfully mirrored and there is a 'egcs -> gcc' link elsewhere that
>>has been around for a long time.  It also is mirrored correctly on
>>the sites that I looked at.
>
>Nice, if you look you'll notice that I used softlinks recklessly.  The 
>snapshots will slowly flush them.

One irritating thing that I noticed on some mirrors is that they don't
seem to notice deletions.  So, many sources.redhat.com directories are
slowly being filled with past snapshots that have since been purged from
our server.

cgf


^ permalink raw reply	[flat|nested] 10+ messages in thread

end of thread, other threads:[~2001-10-13 10:53 UTC | newest]

Thread overview: 10+ messages (download: mbox.gz / follow: Atom feed)
-- links below jump to the message on this page --
2001-10-01 17:21 Changes to snapshot process ... ? Andrew Cagney
2001-10-02  1:14 ` Eli Zaretskii
2001-10-02  8:58   ` Andrew Cagney
2001-10-02 10:07     ` Eli Zaretskii
2001-10-10 21:24       ` Andrew Cagney
2001-10-11  1:06         ` Eli Zaretskii
2001-10-12 12:53           ` Andrew Cagney
2001-10-12 13:22             ` Christopher Faylor
2001-10-13 10:25               ` Andrew Cagney
2001-10-13 10:53                 ` Christopher Faylor

This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox