Re: [sqlite] Possible inaccuracy in "Isolation In SQLite" paper

2017-07-20 Thread Olivier Mascia
> Le 20 juil. 2017 à 10:23, Olivier Mascia a écrit : > >> Le 20 juil. 2017 à 07:34, Rowan Worth a écrit : >> >> IMMEDIATE would take a RESERVED lock which is clearly not desired in this >> case -- the reader only wants a SHARED lock so as to minimise

Re: [sqlite] Possible inaccuracy in "Isolation In SQLite" paper

2017-07-20 Thread Gwendal Roué
> Le 20 juil. 2017 à 08:45, Clemens Ladisch a écrit : > > Rowan Worth wrote: >> On 18 July 2017 at 21:43, David Raymond wrote: >>> You can run "begin deferred transaction" then walk away for 3 months >>> without upsetting anything. If you need the

Re: [sqlite] Possible inaccuracy in "Isolation In SQLite" paper

2017-07-20 Thread Olivier Mascia
> Le 20 juil. 2017 à 07:34, Rowan Worth a écrit : > > IMMEDIATE would take a RESERVED lock which is clearly not desired in this > case -- the reader only wants a SHARED lock so as to minimise contention > with the writer. This discussion revolved around WAL mode. BEGIN

Re: [sqlite] Possible inaccuracy in "Isolation In SQLite" paper

2017-07-20 Thread Clemens Ladisch
Rowan Worth wrote: > On 18 July 2017 at 21:43, David Raymond wrote: >> You can run "begin deferred transaction" then walk away for 3 months >> without upsetting anything. If you need the precise timing then why not >> just use "begin immediate"? > > IMMEDIATE would take

Re: [sqlite] Possible inaccuracy in "Isolation In SQLite" paper

2017-07-19 Thread Rowan Worth
On 18 July 2017 at 21:43, David Raymond wrote: > You can run "begin deferred transaction" then walk away for 3 months > without upsetting anything. If you need the precise timing then why not > just use "begin immediate"? > IMMEDIATE would take a RESERVED lock which is

Re: [sqlite] Possible inaccuracy in "Isolation In SQLite" paper

2017-07-18 Thread Peter Da Silva
I notice that “read_uncommitted pragma” is spelled “PRAGMA read_uncommitted” in one place. The links all match. This small inconsistency is probably a mistake. ___ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org

Re: [sqlite] Possible inaccuracy in "Isolation In SQLite" paper

2017-07-18 Thread Gwendal Roué
rs [mailto:sqlite-users-boun...@mailinglists.sqlite.org] On > Behalf Of Gwendal Roué > Sent: Tuesday, July 18, 2017 9:10 AM > To: SQLite mailing list > Subject: [sqlite] Possible inaccuracy in "Isolation In SQLite" paper > > Hello all, > > The following sentence in

Re: [sqlite] Possible inaccuracy in "Isolation In SQLite" paper

2017-07-18 Thread David Raymond
s a RESERVED lock." -Original Message- From: sqlite-users [mailto:sqlite-users-boun...@mailinglists.sqlite.org] On Behalf Of Gwendal Roué Sent: Tuesday, July 18, 2017 9:10 AM To: SQLite mailing list Subject: [sqlite] Possible inaccuracy in "Isolation In SQLite"

Re: [sqlite] Possible inaccuracy in "Isolation In SQLite" paper

2017-07-18 Thread Gerry Snyder
"Deferred means that no locks are acquired on the database until the database is first accessed. Thus with a deferred transaction, the BEGIN statement itself does nothing to the filesystem. Locks are not acquired until the first read or write operation." On Jul 18, 2017 6:10 AM, "Gwendal Roué"

[sqlite] Possible inaccuracy in "Isolation In SQLite" paper

2017-07-18 Thread Gwendal Roué
Hello all, The following sentence in https://www.sqlite.org/isolation.html does not exactly describe the behavior of SQLite (since many versions): > In WAL mode, SQLite exhibits "snapshot isolation". When a read transaction > starts, that reader continues to see an unchanging "snapshot" of the