On Thu, May 26, 2011 at 12:00 AM, Morten Kloster wrote:
[ snip ]
> I have, however, done some significant testing today: I downloaded
> the first 100 revisions of merge.c in libsvn_client (merge.c is the
> largest code file in the current HEAD revision, with almost 800
> revisions), and ran diff o
> Date: Wed, 25 May 2011 12:17:07 +0200
> From: s...@elego.de
> To: d...@daniel.shahaf.name
> CC: smith_winston_6...@hotmail.com; dev@subversion.apache.org
> Subject: Re: large number of large binary files in subversion
>
> On Wed, May 25, 2011 at 01:02:28P
On 18.05.2011 12:13, Daniel Shahaf wrote:
I found the docstring of the deserializer function types unclear as to
the use of @a POOL...
So, the following patch attempts to clarify the role of pools in that
interface:
[[[
Index: subversion/include/private/svn_cache.h
=
On Wed, May 25, 2011 at 1:44 AM, Johan Corveleyn wrote:
> On Tue, May 24, 2011 at 10:46 PM, Stefan Sperling wrote:
>> On Tue, May 24, 2011 at 10:22:58PM +0200, Morten Kloster wrote:
>>> Johan / Stefan - any progress on the reviewing part?
>>
>> I haven't had time to review this, sorry.
>> It got
Daniel Shahaf wrote on Thu, May 26, 2011 at 00:36:48 +0300:
> I've not read the code, but would an array of 'svn_revnum_t[2]' be
> a better representation?
>
> Specifically: an append-only array of [from_rev, to_rev] pairs, sorted
> by from_rev. That's less overhead (and we could take advantage o
I've not read the code, but would an array of 'svn_revnum_t[2]' be
a better representation?
Specifically: an append-only array of [from_rev, to_rev] pairs, sorted
by from_rev. That's less overhead (and we could take advantage of the
sorting to store less data), at the cost of O(log(n)) lookup.
G
On Wed, May 25, 2011 at 16:08, C. Michael Pilato wrote:
> On 05/25/2011 04:05 PM, C. Michael Pilato wrote:
>> On 05/25/2011 03:49 PM, Greg Stein wrote:
>>> On Wed, May 25, 2011 at 15:33, wrote:
...
+ /* A mapping of svn_revnum_t * dump stream revisions to their
+ correspondin
On 05/25/2011 04:05 PM, C. Michael Pilato wrote:
> On 05/25/2011 03:49 PM, Greg Stein wrote:
>> On Wed, May 25, 2011 at 15:33, wrote:
>>> ...
>>> + /* A mapping of svn_revnum_t * dump stream revisions to their
>>> + corresponding svn_revnum_t * target repository revisions. */
>>> + apr_hash
On 05/25/2011 03:49 PM, Greg Stein wrote:
> On Wed, May 25, 2011 at 15:33, wrote:
>> ...
>> + /* A mapping of svn_revnum_t * dump stream revisions to their
>> + corresponding svn_revnum_t * target repository revisions. */
>> + apr_hash_t *rev_map;
>
> How big can this grow? ie. what happen
On Wed, May 25, 2011 at 15:33, wrote:
>...
> + /* A mapping of svn_revnum_t * dump stream revisions to their
> + corresponding svn_revnum_t * target repository revisions. */
> + apr_hash_t *rev_map;
How big can this grow? ie. what happens when there are several million
revisions.
Cheers,
On 05/25/2011 10:42 AM, Bert Huijben wrote:
> To resolve this and similar problems we defined that we are not allowed
> to remove pristines with refcount zero until there are no more workqueue
> items, and nobody else has a working copy lock.
>
> (During update the only reference to a pristine is
Noorul Islam K M wrote:
> Noorul Islam K M writes:
> > Julian Foad writes:
[...]
> >> * Use SVN_ERR instead of svn_error_clear. There 'kind' variable is not
> >> guaranteed to be set to a valid value if you the function throws an
> >> error.
> >>
> >> * Name the variable the same way ('to_kind')
> -Original Message-
> From: Branko Čibej [mailto:br...@xbc.nu] On Behalf Of Branko Cibej
> Sent: dinsdag 24 mei 2011 18:22
> To: dev@subversion.apache.org
> Subject: Re: [PATCH] Server-transmitted "final SHA1 checksums" (Was: "large
> number of large binary files in subversion")
>
> On
Noorul Islam K M writes:
> Julian Foad writes:
>
>> Just taking a quick look.
>>
>> * Please always include the log message with the patch.
>>
>
> I did include the log message in one of the previous mails. Since there
> was no change to the log message I did not include in the subsequent
> repl
Hi,
> Von: Stefan Sperling [mailto:s...@elego.de]
> > > using
> > > 'svnadmin hotcopy' to copy your repositories to it once a week or
so
> >
> > Yes, I planned to do that for a read-only backup repository as part
of
> > various backup schedules (daily, weekly, monthly, yearly).
>
> Unfortunately
Greg Stein writes:
> On Tue, May 24, 2011 at 14:37, Philip Martin
> wrote:
>> It's the expat releases tarballs that have changed.
>
> That's weird. Cuz we (the expat team) haven't done a release for about
> four years. Some downstream packager, I guess.
The relevant expat versions are 1.95.2 a
On Wed, May 25, 2011 at 01:02:28PM +0300, Daniel Shahaf wrote:
> Stefan Sperling wrote on Wed, May 25, 2011 at 11:45:21 +0200:
> > On Wed, May 25, 2011 at 10:03:16AM +1100, Winston Smith wrote:
> > > Yes, I planned to do that for a read-only backup repository as part of
> > > various backup schedul
On May 25, 2011 5:50 AM, "Stefan Sperling" wrote:
>
> On Wed, May 25, 2011 at 10:12:42AM +0100, Julian Foad wrote:
> > We can catch some WC DB inconsistencies as soon as they happen by adding
> > SQLite "trigger" functions that fire on adding or modifying rows, and
> > throw an error if some condi
Stefan Sperling wrote on Wed, May 25, 2011 at 11:45:21 +0200:
> On Wed, May 25, 2011 at 10:03:16AM +1100, Winston Smith wrote:
> > Yes, I planned to do that for a read-only backup repository as part of
> > various backup schedules (daily, weekly, monthly, yearly).
>
> Unfortunately there is no inc
Julian Foad writes:
> Just taking a quick look.
>
> * Please always include the log message with the patch.
>
I did include the log message in one of the previous mails. Since there
was no change to the log message I did not include in the subsequent
replies.
> * Use SVN_ERR instead of svn_erro
On Wed, May 25, 2011 at 10:12:42AM +0100, Julian Foad wrote:
> We can catch some WC DB inconsistencies as soon as they happen by adding
> SQLite "trigger" functions that fire on adding or modifying rows, and
> throw an error if some condition is not met. This is especially useful
> in the NODES ta
On Wed, May 25, 2011 at 10:03:16AM +1100, Winston Smith wrote:
>
> > > repo(s) is/are on a permanently mounted USB disk.
> >
> > The USB stick might give you less i/o throughput than an internal hard
> > disk, and it might fail early due to flash wearing out.
>
> I believe I mentioned a disk, not
On May 24, 2011 11:47 AM, "Stefan Sperling" wrote:
>
> On Tue, May 24, 2011 at 11:30:39AM -0400, C. Michael Pilato wrote:
> > I held off on committing my RA work because there were no consumers.
But
> > I've still considering making the change because a) there's no penalty
for
> > doing so, and b
On Tue, 2011-05-24 at 18:21 +0200, Branko Čibej wrote:
> On 24.05.2011 17:53, Arwin Arni wrote:
> >> I was hoping to slip in
> >> support for avoiding those text transfers altogether where possible.
> >> But I
> >> ran into the obvious problems with the editor interface. (Also, I
> >> became
> >>
We can catch some WC DB inconsistencies as soon as they happen by adding
SQLite "trigger" functions that fire on adding or modifying rows, and
throw an error if some condition is not met. This is especially useful
in the NODES table because of its complex rules.
This patch is presented as a conce
On Wed, May 25, 2011 at 10:31 AM, Daniel Shahaf wrote:
> Branko Čibej wrote on Wed, May 25, 2011 at 08:34:37 +0200:
>> On 25.05.2011 06:34, Ivan Zhakov wrote:
>> > On Wed, May 25, 2011 at 03:00, Bert Huijben wrote:
>> >>> 2011/5/24 Branko Čibej :
>> On 24.05.2011 11:02, Johan Corveleyn wrote
Branko Čibej wrote on Wed, May 25, 2011 at 08:34:37 +0200:
> On 25.05.2011 06:34, Ivan Zhakov wrote:
> > On Wed, May 25, 2011 at 03:00, Bert Huijben wrote:
> >>> 2011/5/24 Branko Čibej :
> On 24.05.2011 11:02, Johan Corveleyn wrote:
> [...]
> > I'm fully agree with Bert that moving truep
27 matches
Mail list logo
