Justin Erenkrantz wrote: > On Sat, May 25, 2002 at 10:07:54PM -0400, Cliff Woolley wrote: > >>Okay, I know there are still things that need to go in prior to 2.0.37, >>but I also know that HEAD is currently nice and stable, at least on Unix. >>So in order to start working toward a release without holding up any other >>work, I've tagged the tree as JCW_PRE_2037. We'll tag in changes to HEAD >>as necessary and aim to retag the result as 2.0.37 on, say, Wednesday or >>Thursday. > > > I think the mod_deflate issue needs to be resolved before 2.0.37. > You and Greg already addressed why that commit should be reverted.
I disagree. the commit doesn't fix the entire problem, but it is much better than what was there before. beforehand it would completly ignore what was in the content-encoding header and just replace it with gzip. now it takes a more conservative approach (although not 100% correct) and if it finds something already in there it just doesn't compress the file (which is ALWAYS a valid option for the module) I agree that it should make a special case for the 'identity' content-encoding and let that go through --Ian > > I have the ap_read_pid (sp?) function calling read_full in my > tree, but I'm also having a problem on Darwin starting up HEAD, so I > may just post the patch so that I'm not blocking this (I'm still > traveling, so my time/email is sporadic). > > It's hanging when it tries to look up the address in Listen (0.0.0.0 > in this case). I'm not enitrely certain this isn't an OS problem > (doing a clean rebuild now). If a rebuild fixes it, I'll test and > commit, otherwise, I'll post the ap_read_pid change and post a > stack trace of my hang for someone else to investigate. -- justin >
