Re: [Evolution] Evo Mail component crashing on this message: [Fwd: RE: Question about TSP (rfc 3161)]

Btw, this is now fixed in CVS in several ways.

1. if iconv doesn't recognize the charset, just write out the raw text.

2. wrote a parser for windows-[cp]#### charsets and convert them to the
iconv-friendly format of cp####.

I love iconv suckage...


On Thu, 2001-09-13 at 12:56, Rick Ziegler wrote:
The forwarded message is consistently crashing the latest RH 6.2
RedCarpet snapshot.

Richard Ziegler
Release Engineer / ClearCase Administrator
(617) 503-0442
CertCo Inc.   

From: Cristian Marinescu <cristian marinescu omicron at>
To: libel paris md
Cc: ietf-pkix imc org
Subject: RE: Question about TSP (rfc 3161)
Date: 13 Sep 2001 16:27:30 +0200

Hash: SHA1


I agree, there is in practice (at least for the moment,
when everyone is trying to put a TSA together, more or less
draft compliant), no reason for having such flags.

But, it is also not stupid.
Let's imagine that some day the TSP will be really used
by everyone.  :)  Hard to imagine, but let's try.
This will rise the problem of DOS attack, and some
people, (as I have done myself) will implement the TSA
and limit the number of parallel requests (and let's understand
by this the number of spawn processes, or threads) to a fix number. So
they will
just return an error back. This is actually not a nice thing
to do. And I presume, people there, writing the draft,
tried to be nice: well, if I get a request, but,
I don't have time to give a response right away, because
I am busy, let's store the request and tell the person to try
again sometimes later. 
Perhaps there is also the possibility
that your clock is at that moment not available, (I would
like to believe that there will be TSA's out there that won't read
the time from the local system, like I do at the moment...) or maybe
other resource... how could I know??  :)  
In any case you have to take cautions about the overflooding with
requests (or even pending requests, that havn't been answered yet)

Well, at least this is the reason I can imagine. Perhaps
there are also some other (dark!) reasons, but I would like
to hear/read about them from the TSP gurus. :))

Kindly regards,

-----Original Message-----
From: libel paris md [mailto:libel paris md]
Sent: Mittwoch, 12. September 2001 10:19
To: ietf-pkix imc org
Subject: Question about TSP (rfc 3161)


I would like to know what are the reasons for introducing the 
flags "pollReq",
"pollResp" and "negPollRep" in the socket based protocol 
(section 3.3).

It would mean that a tsa server can divide the der code he 
calculated for the
response. But why would it do that?



Get your firstname lastname email for FREE at

Version: PGPfreeware 6.0.2i


[Date Prev][Date Next]   [Thread Prev][Thread Next]   [Thread Index] [Date Index] [Author Index]