4206909 - adding Z_SYNC_FLUSH support to deflaters

Xueming Shen Xueming.Shen at Sun.COM
Tue Sep 22 20:03:00 UTC 2009

Thanks Alan!

The webrev has been updated accordingly to address your comments (use 
syncFlush instead of
doSyncFlush and those suggestions for Deflater.java).



PS. There was a "take2" for DOS that I think might be more consistent 
with the existing APIs, but it might be the
time to focus on one approach. If you're interested the "take2" is at


Alan Bateman wrote:
> Welcome back (for some reason I thought you were gone for two weeks).
> It would be best to send the proposal to the mailing list so that 
> others can comment.
> Personally, I don't like exposing the flush mode to subclasses but you 
> are right that it is more consistent with the original design.  If 
> this approach is chosen then I would suggest that the flush 
> implementation make a copy of the flush mode before testing and using 
> it. Also, the new constructors will need to say that they initialize 
> the flush mode based on the parameter (since it can change on the 
> fly). BTW: What is the reason for renaming the parameter to flushDef. 
> I prefer "syncFlush" over "doSyncFlush".
> I'm happy with Deflater. In the description of SYNC_FLUSH I would 
> suggest changing the "," into a ";" or else make it into two 
> sentences. In the new deflate method the description of NO_FLUSH needs 
> a comma after "accumulate" for it to flow well. Similar issue in the 
> SYNC_FLUSH description where you need a comma after "is flushed". 
> There is also a typo [ comparssed :-) ]
> -Alan.

More information about the core-libs-dev mailing list