Обсуждение: Support for invalid foreign keys
Hi, This is a new (as in 9.1) feature we have to support in pgAdmin. From the git log on https://github.com/gleu/pgadmin3/tree/ticket304: Add support for invalid foreign key constraints This patch adds a checkbox in the dlgForeignKey dialog, so that the user can create an "not validated" foreign key. Each foreign key of this kind is displayed with the usual icon with a red cross on it. A contextual menu allows the user to force the validation of the constraint. The user can also do this via the dlgForeignKey dialog. Patch attached. Comments welcomed :) -- Guillaume http://www.postgresql.fr http://dalibo.com
Вложения
On Sun, Mar 6, 2011 at 10:28 PM, Guillaume Lelarge <guillaume@lelarge.info> wrote: > Hi, > > This is a new (as in 9.1) feature we have to support in pgAdmin. > > From the git log on https://github.com/gleu/pgadmin3/tree/ticket304: > > Add support for invalid foreign key constraints > > This patch adds a checkbox in the dlgForeignKey dialog, so that the > user can create an "not validated" foreign key. Each foreign key of > this kind is displayed with the usual icon with a red cross on it. A > contextual menu allows the user to force the validation of the > constraint. The user can also do this via the dlgForeignKey dialog. So it basically disables the constraint, leaving it as a sort of documentation of intent? -- Dave Page Blog: http://pgsnake.blogspot.com Twitter: @pgsnake EnterpriseDB UK: http://www.enterprisedb.com The Enterprise PostgreSQL Company
On 7 March 2011 10:18, Dave Page <dpage@pgadmin.org> wrote: > On Sun, Mar 6, 2011 at 10:28 PM, Guillaume Lelarge > <guillaume@lelarge.info> wrote: >> Hi, >> >> This is a new (as in 9.1) feature we have to support in pgAdmin. >> >> From the git log on https://github.com/gleu/pgadmin3/tree/ticket304: >> >> Add support for invalid foreign key constraints >> >> This patch adds a checkbox in the dlgForeignKey dialog, so that the >> user can create an "not validated" foreign key. Each foreign key of >> this kind is displayed with the usual icon with a red cross on it. A >> contextual menu allows the user to force the validation of the >> constraint. The user can also do this via the dlgForeignKey dialog. > > So it basically disables the constraint, leaving it as a sort of > documentation of intent? Yes, I believe the purpose is to allow the creation of a foreign key with far less disruptive locking meaning foreign keys can be created in parallel. Guillaume, you appear to have commented out an if container, while also adding a condition to it. Is this supposed to be commented out at all?.. + //if (sql.IsNull() || forceRefresh) -- Thom Brown Twitter: @darkixion IRC (freenode): dark_ixion Registered Linux user: #516935
Le 07/03/2011 06:49, Thom Brown a écrit : > On 7 March 2011 10:18, Dave Page <dpage@pgadmin.org> wrote: >> On Sun, Mar 6, 2011 at 10:28 PM, Guillaume Lelarge >> <guillaume@lelarge.info> wrote: >>> Hi, >>> >>> This is a new (as in 9.1) feature we have to support in pgAdmin. >>> >>> From the git log on https://github.com/gleu/pgadmin3/tree/ticket304: >>> >>> Add support for invalid foreign key constraints >>> >>> This patch adds a checkbox in the dlgForeignKey dialog, so that the >>> user can create an "not validated" foreign key. Each foreign key of >>> this kind is displayed with the usual icon with a red cross on it. A >>> contextual menu allows the user to force the validation of the >>> constraint. The user can also do this via the dlgForeignKey dialog. >> >> So it basically disables the constraint, leaving it as a sort of >> documentation of intent? > > Yes, I believe the purpose is to allow the creation of a foreign key > with far less disruptive locking meaning foreign keys can be created > in parallel. > Yeah, you first create the constraint without being disruptive (on big tables), and you validate the constraint at a time the system is less used (for example during the night when everyone's asleep :) ). > Guillaume, you appear to have commented out an if container, while > also adding a condition to it. Is this supposed to be commented out > at all?.. > > + //if (sql.IsNull() || forceRefresh) > Well, it was supposed to be deleted. Actually, I was supposed to ask if anyone knows why there is the "if (sql.IsNull())" check at all (the "|| forceRefresh" is a test I should have deleted). I have no idea why there is this sql.IsNull() test. It doesn't appear in other same sources. Dave, any idea? -- Guillaume http://www.postgresql.fr http://dalibo.com
On Mon, Mar 7, 2011 at 6:21 AM, Guillaume Lelarge <guillaume@lelarge.info> wrote: > Well, it was supposed to be deleted. Actually, I was supposed to ask if > anyone knows why there is the "if (sql.IsNull())" check at all (the "|| > forceRefresh" is a test I should have deleted). I have no idea why there > is this sql.IsNull() test. It doesn't appear in other same sources. > Dave, any idea? It's an optimisation to prevent us regenerating the SQL every time we need it. For most objects that's cheap - for some (eg. tables) - far from it. When refreshing, frmMain->Refresh calls the SetDirty() function on the object, which causes sql to be cleared ready for regeneration. -- Dave Page Blog: http://pgsnake.blogspot.com Twitter: @pgsnake EnterpriseDB UK: http://www.enterprisedb.com The Enterprise PostgreSQL Company
Le 07/03/2011 07:49, Dave Page a écrit : > On Mon, Mar 7, 2011 at 6:21 AM, Guillaume Lelarge > <guillaume@lelarge.info> wrote: >> Well, it was supposed to be deleted. Actually, I was supposed to ask if >> anyone knows why there is the "if (sql.IsNull())" check at all (the "|| >> forceRefresh" is a test I should have deleted). I have no idea why there >> is this sql.IsNull() test. It doesn't appear in other same sources. >> Dave, any idea? > > It's an optimisation to prevent us regenerating the SQL every time we > need it. For most objects that's cheap - for some (eg. tables) - far > from it. > > When refreshing, frmMain->Refresh calls the SetDirty() function on the > object, which causes sql to be cleared ready for regeneration. > OK, commited with a few fixes according to this discussion. Thanks. -- Guillaume http://www.postgresql.fr http://dalibo.com