On Wed, 2008-11-19 at 21:58 -0600, Thilina Gunarathne wrote: > Hi Asanka, > Hope you've seen the following thread before posting this.. > http://markmail.org/message/ayjxurjz2fow7hit > > > > > Quoting one of Oleg's mail, > "Presently the patch I submitted (WSCOMMONS-387) does not provide > support for the deferred parsing, but it should be relatively trivial > to add it." > I'm against committing the patch unless *at least* the above is > implemented. >
Thilina, There was no feedback on the patch so I felt it was pointless to put more work into it. Anyways, I updated the patch with support for deferred parsing of MIME streams. Folks, feel free to use this patch whichever way you please. > > IIRC the consensus of that thread are as follows, > 1. Create a new branch and complete the patch to perform what Axiom is > capable of doing at the moment. > 2. Put this behind an interface so that we can switch the > implementations. > > > IIRC Oleg has commit rights in WS. I have commit rights for Synapse, but I do not recall ever being granted WS commit rights. Oleg > Hence I assume he can go ahead and start developing in a branch on > Axiom (which at that time I thought will happen after the above mail > thread). > > > To be honest, I'm not fan of doing drastic changes to a working code > base without having extensive proof (in this case a branch that works) > that it'll work. Especially if it involves throwing away some existing > code. > > > thanks, > Thilina > > On Wed, Nov 19, 2008 at 12:52 PM, Davanum Srinivas <[EMAIL PROTECTED]> > wrote: > Let's do #2 and everyone will be happy... > > -- dims > > > On Wed, Nov 19, 2008 at 1:48 PM, Oleg Kalnichevski > <[EMAIL PROTECTED]> wrote: > > On Wed, 2008-11-19 at 10:32 -0500, Davanum Srinivas wrote: > >> Oleg, > >> > >> Yes, we need to do exactly what you mention first. > >> > >> "The whole trouble is that Axiom currently does not allow > for multiple > >> implementations of MIME processing as it does not have an > abstract API > >> for that to start with." > >> > >> -- dims > >> > > > > Davanum, > > > > Frankly, I could not care less about the underlying > implementation, as > > long as the Axiom API enabled me (1) parse and build MIME > messages > > without buffering various parts in memory (2) replace > certain bits If > > the default implementation did not quite meet the specific > application > > requirements. Currently neither 1 nor 2 is the case. Hence > my attempt at > > fixing both issues. I chose mime4j because I thought Axiom > project might > > want to depend on a generic library rather than maintaining > its own > > implementation. > > > > Oleg > > > > > > > >> On Wed, Nov 19, 2008 at 10:16 AM, Oleg Kalnichevski > <[EMAIL PROTECTED]> wrote: > >> > On Wed, 2008-11-19 at 08:39 -0500, Davanum Srinivas > wrote: > >> >> Let's please keep it as a parallel optional > implementation. just like > >> >> we have so many components with specific interfaces and > multiple > >> >> implementations. > >> > > >> > Davanum > >> > > >> > The whole trouble is that Axiom currently does not allow > for multiple > >> > implementations of MIME processing as it does not have an > abstract API > >> > for that to start with. > >> > > >> > > >> >> If/when we finally see the promised results, all we > >> >> need to do is switch the default to the other > implementation. > >> >> > >> > > >> > I am not sure I understand exactly what kind of results > you are > >> > referring to. The patch (1) provides an abstract API for > MIME processing > >> > and (2) eliminates in memory buffering of SOAP envelops. > The Mime4j > >> > based parser is currently ~15% slower than the existing > implementation, > >> > but I _personally_ see this a reasonable trade-off for a > smaller memory > >> > footprint and a better API. So, I do not think I made any > promises I am > >> > still to deliver upon. > >> > > >> > Anyways, this patch is just not worth the trouble. Scrap > it > >> > > >> > Oleg > >> > > >> >> thanks, > >> >> dims > >> >> > >> >> On Wed, Nov 19, 2008 at 8:15 AM, Davanum Srinivas > <[EMAIL PROTECTED]> wrote: > >> >> > I am asking for the MIME4J implementation to be an > option. Not the > >> >> > default. Default should be what is in right now. > >> >> > > >> >> > thanks, > >> >> > dims > >> >> > > >> >> > On Wed, Nov 19, 2008 at 5:41 AM, Paul Fremantle > <[EMAIL PROTECTED]> wrote: > >> >> >>> Asankha et al > >> >> >>> > >> >> >>> The patch was meant as a replacement for the > existing implementation. It > >> >> >>> is pointless to have two MIME frameworks in Axiom. > Just forget about the > >> >> >>> patch. > >> >> >>> > >> >> >>> Oleg > >> >> >> > >> >> >> -1! I think we should evaluate this patch just like > any other. I'm not > >> >> >> clear what Dims is asking I'm afraid, but I certainly > don't want to > >> >> >> forget about the patch. > >> >> >> > >> >> >> Paul > >> >> >> > >> >> >> > --------------------------------------------------------------------- > >> >> >> To unsubscribe, e-mail: > [EMAIL PROTECTED] > >> >> >> For additional commands, e-mail: > [EMAIL PROTECTED] > >> >> >> > >> >> >> > >> >> > > >> >> > > >> >> > > >> >> > -- > >> >> > Davanum Srinivas :: http://davanum.wordpress.com > >> >> > > >> >> > >> >> > >> >> > >> > > >> > > >> > >> > >> > > > > > > > > > -- > > Davanum Srinivas :: http://davanum.wordpress.com > > > > > -- > Thilina Gunarathne - http://thilinag.blogspot.com > --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
