elsn4000@mailszrz.zrz.tu-berlin.de (Frank Elsner) (06/11/91)
Why doesn't ELM tolerate the specification of DSt in the "From " line ? The specification is one field too much for elm, he reports the folder being corrupt. After deleting the string "DST " all works perfect. Is this behavier a must ? Please let ELm be a little more tolerant. -- Frank Elsner (TU Berlin, Postmaster)
syd@DSI.COM (Syd Weinstein) (06/11/91)
elsn4000@mailszrz.zrz.tu-berlin.de (Frank Elsner) writes: >Why doesn't ELM tolerate the specification of DSt in the "From " line ? >The specification is one field too much for elm, he reports the folder >being corrupt. After deleting the string "DST " all works perfect. Why Elm doesn't is a matter of history and having no spec.... There is no 'official' specification for the "From " line in UNIX mailbox format. Its not part of the RFC's because its only in the UNIX mailbox format. And its not in the SVID. The 'unofficial' format is for the line to have the form 'From ' return-path Date-Time where return-path is one word, and Date-Time is a valid timestamp. Now, UNIX timestamps used to be of the format: Day Mon dd hh:mm:ss TZ year and in this format there are a particular number of fields.... 6 this the From line has exactly 8 fields. Some sites never defined a TZ and all of a sudden the format "Day Mon dd hh:mm:ss year" started showing up. This was also allowed in Elm and a 7 field from line was then also allowed. (I think this is also wrong, but....) Now, some sites in Europe have taken to making the timestamp: Day Mon dd hh:mm:ss TZ [TZmod] year where the TZ does not change for daylight/summer time, but is followed by another field, containing dst in some form of capitalization for that flag. This is then a 6 or 7, depending on the time of the year. This format of the date is NOT supported elsewhere in the RFCs for a valid date. The TZ is always a single field, either a recognized mnemonic, or the number of hours of deviation from UT (with a counter- intiutive sign). It is an easy hack to make Elm tolerant of this also, and I recommend that you make the hack if your site runs into this. The more important question, to me, is whether this is a proper format or not. I just want someone to write the spec, and then have everyone adhere to it (should there be a :-) there???). >Is this behavier a must ? Please let ELm be a little more tolerant. As to why Elm even checks at all... This is to prevent falsing on 'From ' lines that are not marked with a >From in the body. This is currently very important, as 2.3 does not support the Content-Length header. On SVR4 systems that use the Content-Length header, the 'From ' is NOT escaped to '>From ' and could be falsed. (2.4 will support Content-Length). There will not be any 'official' patch to 2.3 for this, so make any local ones you see fit. In fact, baring any catastrophic problems with Elm there will be no more patches to 2.3 at all. 2.4 is too close to want to patch 2.3. (having said that, no one will finish their work on 2.4 and it will be at least a year till it comes out 1/2 :-).). -- ===================================================================== Sydney S. Weinstein, CDP, CCP Elm Coordinator Datacomp Systems, Inc. Voice: (215) 947-9900 syd@DSI.COM or dsinc!syd FAX: (215) 938-0235