Apache OpenOffice (AOO) Bugzilla – Issue 81962
pppopt02: file size estimation invalid in p-minimizer
Last modified: 2007-10-09 13:31:45 UTC
1. open bugdoc 2. p-minimizer 3. push the finish button (with default settings) current file size: 38.2 MB estimate new file size: 17.9 MB after minimizing: 0.4 MB => the estimation is that far away from the result that it makes no sense (invalid)
set keyword 'new_implementation' clu->sj: copy bugdoc file to your 'tausch' folder because of the size
sj: At the moment this issue can't be fixed. To determine an accurate estimation more information of each graphic is necessary, but this would require that each graphic has to be swapped in, this is very time consuming. We should discuss this problem within our next i-team meeting.
reopen issue -> wontfix is not the proper state clu->sj: if you do not have the time to fix this bug, we have to discuss the target milestone ;) p.s. one possibility could be to remove the estimation display till it works proper -> better no estimation than such a wrong one (i guess this would not take that much time)
sj->clu: It's not a problem of my time, and yes, I also didn't want to have this estimation. And it is also no proper state to lay down this issue to the developer without having a solution. So it seems to be something for mmp.
As long as this is a 1 in 50 case I suggest to live with it. Christian - how is the distribution of estimation/real new file size?
clu->mmp: i see this problem every time, if i use slides containing graphics and i guess it happens in a lot of more cases too - so i also guess not 1:50, but >50% (i do not know how many people use slides without graphics) -> maybe sj can predict in what cases exactly it goes wrong and then we can estimate the impact, but i think it is much more than an exception ;)
sj told me that wg will have a deeper look and try several 'real' docs to check the impact
I had a further look at the behaviour. It seems that we had bad luck that the first test doc we setup drove the estimation nuts. I checked with several other files and it is ok there. So it seems that it is 1 out of 50 files that gets a wrong estimation. Closing this issue. Thanks to all.
Closed.