about search post article Documentation Mailing Lists Bug Tracking Development Installation Upgrading Download admin rdf |
from the squish-in-chinese dept. dept. I have set up a Squishdot site (excellent software Butch) in chinese. We have modified the DTML as well as Squishdot.py to add some extra functionality. However as the site is becoming more popular we have encountered the following problems: 1. The python process is consuming too much RAM >256Mb of RAM whereas the whole Data.fs is only 7Mb. I am using Zope-2b1 so the memory leak could be with Zope or with Squishdot, however we are running one more Zope-2b1 site (non-squishdot) and we never got such problems. 2. In two months we have accumulated 900 postings (are we the largest Squishdot site?) and it gets very difficult to maintain. Moreover a search can return too many results and we would like to use some batch viewing. The way the search function is currently implemented does not allow the use of the DTML batch processing constructs. (This is not a Squishdot problem per se, and I am already working to solve it, but if someone has an easy solution please let me know). I would have *really* liked it if the postings were normal DTML Documents with associated properties, seperated in folders according to subject or even better by month. That would have solved our third problem (also mentioned elsewhere on the Squishdot site) of migrating our postings to the newer version. In summary: I am going to make heavy modifications to Squishdot so I might as well ask all of you for advice. For ease of management of large Squishdot sites, do you think is a good idea to move the postings into folders by month? Butch, would this break too much of the internal structure? If not do you think is a good idea to use DTML Douments to represent postings so we can use Zope's regular import/export mechanisms? If not then I am willing to write an export/import function in XML. < | >
|
|
|||||||||||
|