[111] in Hesiod
Re: hesiod nameserver limitations
daemon@ATHENA.MIT.EDU (Greg Wohletz)
Mon Nov 16 19:01:34 1992
To: Mark Rosenstein <mar@mit.edu>
Cc: hesiod@Athena.MIT.EDU
In-Reply-To: Your message of "Mon, 16 Nov 92 18:20:47 EST."
Date: Mon, 16 Nov 92 15:44:53 -0800
From: Greg Wohletz <greg@duke.cs.unlv.edu>
> Date: Thu, 05 Nov 92 22:47:30 -0800
> From: Greg Wohletz <greg@duke.cs.unlv.edu>
>
> I recently ran into a strange problem when a group entry went beyond a
> certain length (it was < 256 chars, but > 127). I thought the limit
> on TXT records was 256 characters. Am I mistaken, or perhaps are
> there signed chars in the nameserver or resolver code where there
> shouldn't be effectively reducing the limit to 127? Or is something
> else going on? The symptoms were garbled packets (apparently) being
> returned whenever the record for that group was queried.
>
>We successfully send TXT records with about 450 bytes here at MIT. If
>you read the spec, it says that a TXT record is a series of one or
>more of a length byte followed by that many data bytes. The code we
>ship uses these unsigned bytes to put the first 255 bytes of data in
>the first string, and more data in successive strings. This compiles
>and runs correctly on the VAX, which has signed characters, so I'm
>pretty sure we got that part of the code right.
You are correct, you got it right. However various other folks
apparently did not, most notably the program ``dig'' which barfs on
them, and a perl program of mine that was adapted from some dns perl
code you once created. When my perl script stopped working I used dig
to try and see what was going on, when it didn't work I became
convinced that the nameserver was to blame. Guess I should have
tested it via the hes_resolve routine (which I just did and it works
fine)...
--Greg
