sophie schrieb: > Hi Mathias, > Mathias Bauer wrote: > > [...] >> >> The refactoring is not influencing this problem directly. >> >> According to Oliver the problem is that we can't clone documents in >> memory and so we have to store the document and load it again and again >> for each and every mail. >> >> We already have talked about that and I'm not convinced that the >> in-memory-cloning is impossible. I agree with you that "OOoLater" is the >> wrong target. The problem is severe enough to deserve at least a "3.x" >> target. Some simple math tells me that mail merge performance is >> unacceptable for every non-trivial document. >> >> I heard from Oliver that he doubts that we can fix it in a reasonable >> time frame - but IMHO this has to be proven (not guessed) before we turn >> it down. So at least a close investigation should be done. As it looks >> we can't do that for 3.0 but perhaps not too far away. >> >> It would be nice to know whether the n load times for n mails are the >> only major performance problem here. Perhaps someone could take the mail >> merge document, measure the time for the mail merge, measure the time >> for loading the document and storing it (at best an average value taken >> from several load procedures) and compares how much loading contributes >> to the bad performance. > > Thanks a lot for your detailed answer, it will help :). > Is there a way we can help in measuring those different times between > loading and storing?
I think for exact measurements we would need to use some source code enabled tooling. If I'm nor wrong, for a first glance it could help to do the following: (1) Measure by stop watch how long it takes to load the mail merge document. Discard the first measurement and take an average time from 10 times loading the document. (2) Measure by stop watch how long it takes to store the mail merge document. Discard the first measurement and take an average time from 10 times storing the document. (3) Measure by stop watch how long mail merging using this document takes. If we have the numbers we can see whether this rough data is enough or if we must provide more exact data. If someone thinks that my suggestion is just a waste of time, please let us know - before someone tries is and finds out that it *is* a waste of time. :-) Ciao, Mathias -- Mathias Bauer (mba) - Project Lead OpenOffice.org Writer OpenOffice.org Engineering at Sun: http://blogs.sun.com/GullFOSS Please don't reply to "[EMAIL PROTECTED]". I use it for the OOo lists and only rarely read other mails sent to it. --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
