As has been pointed out, your function that is called can do the BEGIN and then a SELECT that hits the database to force the obtaining of the read lock. The fact that the BEGIN didn't get the lock is then not seen by the users of your API.
IF at some point a new BEGIN SHARED IMMEDIATE operation becomes available, you can change you function's internals, and the caller doesn't know the difference except that the call got perhaps a bit faster. On 7/31/19 6:36 AM, test user wrote: > Quote: What importance does it have for you that it already holds an > "end-mark? > Quote: Why would it matter that a writer did write and commit between the > "reader" BEGIN and its first read? > > Im writing a library and would like to have an API where the "read > transaction" has a clear beginning in time. > > BEGIN IMMEDIATE forces a "write transaction", but there is no counterpart > for a "read transaction". > > As an example, the client of this library could: > > - A. Obtain a "read transaction", *without running any SELECTs*. > - B. Complete 20 write transactions in another process. > - C. Begin reading from the read transaction (A) at the point before the > transactions had occurred. > > > At the moment, a "read transaction" is only started on the first SELECT. > > If a client tries to start a "read transaction" with BEGIN, and that > returns SQLITE_OK, its not clear that this has not actually begun any > transaction until the first SELECT query. > > This would enable an API like: > > const r = await db.startReadTx(); > const w = await db.startWriteTx(); > > // At this point in the runtime it clear when the transactions have begun, > and how they will impact other concurrent read/write transactions. > > > > On Tue, Jul 30, 2019 at 11:40 PM Olivier Mascia <o...@integral.be> wrote: > >>> Le 31 juil. 2019 à 00:22, Keith Medcalf <kmedc...@dessus.com> a écrit : >>> >>> I can see where a BEGIN IMMEDIATE SHARED would be useful in non-WAL mode >> though. I will grant that there may be cases where it might be useful in >> WAL mode, even though I cannot think of any. >> >> Fully agree. >> >> — >> Best Regards, Meilleures salutations, Met vriendelijke groeten, Mit besten >> Grüßen, >> Olivier Mascia >> >> _______________________________________________ >> sqlite-users mailing list >> sqlite-users@mailinglists.sqlite.org >> http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users >> > _______________________________________________ > sqlite-users mailing list > sqlite-users@mailinglists.sqlite.org > http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users -- Richard Damon _______________________________________________ sqlite-users mailing list sqlite-users@mailinglists.sqlite.org http://mailinglists.sqlite.org/cgi-bin/mailman/listinfo/sqlite-users
