[Shotwell] Call for testing, round 2

[Roman Yepishev]

On Fri, 2009-12-18 at 08:22 -0500, Vadim Peretokin wrote:
> Welcome to sqlite :)

> All programs that make use of it I know of have this issue - had to
> ditch FF3 because it was taking ages to start
Yes, there was a problem in FF with fsync() when sqlite pragma
synchronous was set to FULL, i.e. every write to db flushed all
filesystem buffers.

Hm, firefox bug[1] seems to occur during database updates, not reads :/

> Liferea takes a good 2min to start too after grinding the disk.
Judging by the bug reports it does a lot of housekeeping on startup,
therefore it gets slow (I have never used Liferea). 
> Only Chrome manages to have an amazingly quick start, and I have no
> idea how do they accomplish it.

What makes me wonder is that there are 2k of reads from the file, each
16 bytes long, then there is a series of lock/unlock calls.

I feel that there is something being done more times than it should.

P.S. I have patched the sources to issue the pragma right after db
opening - no changes.

P.P.S. We might need to start a new thread for this issue :)

[1] https://bugzilla.mozilla.org/show_bug.cgi?id=430530#c4

-- 
Roman Yepishev <roman.yepishev at gmail.com>
-------------- next part --------------
A non-text attachment was scrubbed...
Name: signature.asc
Type: application/pgp-signature
Size: 197 bytes
Desc: This is a digitally signed message part
URL: <https://mail.gnome.org/archives/shotwell-list/attachments/20091218/2b238147/attachment.sig>