"Benson I. Margulies" <Margulies@CISL-SERVICE-MULTICS.ARPA> (02/12/85)
Why in seven hells do these cute-as-a-shithouse-rat mailers insist on touching the text of the destination host name at all? Why not make the rules be "match case insensitive, but preserve what the sender sent?"
"the tty of Geoffrey S. Goodfellow" <Geoff@SRI-CSL.ARPA> (02/13/85)
I'm with you Benson. As I recall, it all started years ago when Dave Crocker thought it was a "good idea" to cutesy-up host names and put code in the MMDF mailer. Since then, the affliction seems to have grown and spread. Can we kill the epidemic before it becomes inter galactically offensive? g
Greg Skinner <Gds@MIT-XX.ARPA> (02/13/85)
I agree with what was said about retaining case insensitivity while also retaining the original header information. A friend of mine was unkindly flamed at by the mailer at ihnp4 which had rejected a legal address sent to me at houxm via ihnp4, excepting the fact that all the host names had been capitalized. The mail software running at ihnp4 was case sensitive, causing the messages to fail, however the sending agent (the VMS mailer on mit-jcf) had no business capitalizing the address my friend supplied, which was in lowercase. --gregbo gds@mit-xx.arpa gregbo%houxm.uucp@harvard.arpa {allegra,cbogsd,ihnp4}!houxm!gregbo -------
lmcl@ukc.UUCP (L.M.McLoughlin) (02/15/85)
In article <8288@brl-tgr.ARPA> "the tty of Geoffrey S. Goodfellow" <Geoff@SRI-CSL.ARPA> writes: >As I recall, it all started years ago when Dave Crocker thought it was >a "good idea" to cutesy-up host names and put code in the MMDF mailer. >Since then, the affliction seems to have grown and spread. Can we >kill the epidemic before it becomes inter galactically offensive? > One mod I *HAD* to make to MMDF in order to run in the UK/Europe was to take out the uppercasing. This was easily done, everywhere in the main system is caseless but will use the case of the official name in the database when asked to rewrite it. I just stopped the database builder from uppercasing it in the first place! So on being given: Kcl-CS kcl-cs or KCL-CS in an address they will all be considered the same as: kcl-cs and will be re-written to that on output. I say *HAD* since nearby sendmail sites stomped all over message headers because it couldn't recognise sites like UKC and KCL-CS. So if you run MMDF and have a neighbour who runs sendmail better check on this.
sjl@ukc.UUCP (S.J.Leviseur) (02/15/85)
There are interesting side affects to removing case conversion from the MMDF database routines. This kills the aliasing of mixed case names. S.J.Leviseur will not now match s.j.leviseur and be remapped, instead the mail will be rejected with an unknown user message. Still, its better than having sendmail sites returning mail because of case sensitivity. sean sjl@ukc.UUCP SERC UK UNIX SUPPORT OFFICER
jim@mcvax.UUCP (Jim McKie) (02/16/85)
In article <4871@ukc.UUCP> sjl@ukc.UUCP (Sean) writes:
........ Still, its better than having sendmail sites returning
mail because of case sensitivity.
Lee McLoughlin also makes the same (wrong) connection between sendmail
and case-sensitivity. When I pointed the problem out to Lee (and to
which he was quick to fix), I thought I made it clear that it was a
fact of life that UUCP is case sensitive, not sendmail. In a sendmail
configuration for a particular mailer, you can specify whether the
case is sensitive or not - for UUCP it is.
I'm no expert, or great fan of sendmail, but it gets a bad enough press
for other things already.
Jim McKie (...*not* at McVax...)
jsdy@hadron.UUCP (Joseph S. D. Yao) (02/26/85)
> Why in seven hells do these cute-as-a-shithouse-rat mailers insist on > touching the text of the destination host name at all? Why not make the > rules be "match case insensitive, but preserve what the sender sent?" Some machines believe in full ASCII, others are only half-Ascii. Some people on the former machines like all one case (their choice), while others like mixed cases, and those on the latter machines get no choice. Unfortunately, not all software is tolerant of this kind of variation, so each system tries to do "the right thing" by its addresses. And, we being the egoless types we are, "the right thing" gets interpreted in some rather interesting ways at different sites. ;-) Joe Yao hadron!jsdy@seismo.{ARPA,UUCP}