brian@cdp.UUCP (03/08/91)
i have just discoverd that the binary search routine in smail 2.5 is faulty. it seems it can miss (at least) the last line in the paths file. the version i have is: ** @(#)defs.h 2.5 (smail) 9/15/87 my paths file looks like: ---------------------file starts below here------------ cdp cdp!%s 100 cdp2 %s 0 smart-host cdp!uunet!%s 100 ---------------------file ends above here-------------- if i try to feed it "smail -dR anyhost!anyuser" i get: resolve: parse address 'anyuser' = 'anyuser' @ '' (LOCAL) resolve: parse address 'anyhost!anyuser' = 'anyuser' @ 'anyhost' (UUCP) getpath: looking for '.anyhost' getpath: looking for 'anyhost' getpath: looking for 'smart-host' route 'anyhost' failed resolve 'anyhost!anyuser' = 'anyuser' @ 'anyhost' (UUCP) COMMAND: /usr/bin/uux -aroot -r - anyhost!rmail '(anyuser)' has there been a patch to fix this? thanks Brian J. Coan Community Data Processing (CdP): 415-322-9069 Internet: brian@cdp.igc.org Bitnet: cdp!brian%labrea@stanford UUCP: uunet!cdp!brian DASNet: [DE3MIR]brian
woods@eci386.uucp (Greg A. Woods) (03/14/91)
In article <135800005@cdp> brian@cdp.UUCP writes: > i have just discoverd that the binary search routine in > smail 2.5 is faulty. it seems it can miss (at least) the > last line in the paths file. the version i have is: > > ** @(#)defs.h 2.5 (smail) 9/15/87 > > my paths file looks like: > ---------------------file starts below here------------ > cdp cdp!%s 100 > cdp2 %s 0 > smart-host cdp!uunet!%s 100 > ---------------------file ends above here-------------- > > if i try to feed it "smail -dR anyhost!anyuser" i get: [a failure to find [smart-host] I think this bug has been previously discovered, and perhaps there is a patch for it. I know I ran into it at one time. However, if I remember correctly, the bug is only triggered on very small files, and only on files such as yours where your smart-host entry is the last thing in the file. I know that my current copy of smail, with the same SCCS-id and date as you show above, can correctly find smart-host in the paths file, and can correctly find the last entry in the file. Our current paths file has an even number of records though (23014). -- Greg A. Woods woods@{eci386,gate,robohack,ontmoh,tmsoft}.UUCP ECI and UniForum Canada +1-416-443-1734 [h] +1-416-595-5425 [w] VE3TCP Toronto, Ontario CANADA Political speech and writing are largely the defense of the indefensible-ORWELL
rob@array.UUCP (Rob Marchand) (03/14/91)
In article <1991Mar13.211418.20877@eci386.uucp> woods@eci386.UUCP (Greg A. Woods) writes: >In article <135800005@cdp> brian@cdp.UUCP writes: >> i have just discoverd that the binary search routine in >> smail 2.5 is faulty. it seems it can miss (at least) the >> last line in the paths file. the version i have is: >> [ Example of problem deleted ] > >I think this bug has been previously discovered, and perhaps there is >a patch for it. I know I ran into it at one time. > For those who are interested, I have a copy of the patch to smail2.5 which corrects the binary search problem (well, `a' binary search problem anyways :-) Send mail if you desire a copy.... Cheers! Rob -- Rob Marchand UUCP : uunet!attcan!lsuc!array!rob Array Systems Computing ARPA : rob%array.UUCP@uunet.UU.NET 401 Magnetic Drive, Unit 24 Phone : +1(416)736-0900 Fax: (416)736-4715 Downsview, Ont CANADA M3J 3H9 Telex : 063666 (CNCP EOS TOR) .TO 21:ARY001
barrett@Daisy.EE.UND.AC.ZA (Alan P. Barrett) (03/19/91)
In article <1991Mar13.211418.20877@eci386.uucp>, woods@eci386.UUCP (Greg A. Woods) writes: > In article <135800005@cdp> brian@cdp.UUCP writes: > > i have just discoverd that the binary search routine in > > smail 2.5 is faulty. it seems it can miss (at least) the > > last line in the paths file. the version i have is: There are (at least) two binary search routines in smail2.5. One deals with the paths file and one deals with the fullnames file. Both had bugs. Both are fixed in the apbmje900117 patches to smail2.5, which I posted to alt.sources earlier this year. (However, I have a nasty suspicion that my posting did not propagate very well -- I don't think it made it onto the list in alt.sources.index. Should I post again?) > I think this bug has been previously discovered, and perhaps there is > a patch for it. I know I ran into it at one time. > > However, if I remember correctly, the bug is only triggered on very > small files, and only on files such as yours where your smart-host > entry is the last thing in the file. I am not sure, but I seem to remember that the bug could bite at just about any time, although small files are indeed more likely to trigger it. It had to do with the relationship between the length of a line in the file being searched, and the distance between the start and end pointers used during the binary search. I am sure that smart-host was not the only string that missed being found, although of course you are more likely to notice if it misses the smart-host than if it misses some other entry. --apb Alan Barrett, Dept. of Electronic Eng., Univ. of Natal, Durban, South Africa Internet: barrett@ee.und.ac.za UUCP: m2xenix!quagga!undeed!barrett