We'll get -C soon. Roland prepared a patch long ago to update usr/src/cmd/sort/ with features from the AST sort to get BSD and GNU features and making /usr/bin/sort and /usr/xpg4/bin/sort a lot faster (even faster than GNU sort). After this step we will concentrate on version of AST sort with multi thread support to squeeze even more performance out of it.
Olga On Fri, Apr 30, 2010 at 1:52 AM, Don Cragun <[email protected]> wrote: > On Apr 28, 2010, at 07:25 AM, Richard L. Hamilton wrote: > >>> Why are there 2 separate binaries for /usr/bin/sort >>> and >>> /usr/xpg4/bin/sort? From looking at it the only >>> difference is option >>> -c. >>> sort(1) says: >>> /usr/bin/sort >>> -c Checks that the single input file is ordered >>> as speci- >>> fied by the arguments and the collating >>> sequence of >>> the current locale. The exit code is set >>> and no output >>> is produced unless the file is out of sort. >>> pg4/bin/sort >>> -c Same as /usr/bin/sort except no >>> output is >>> produced under any circumstances. >>> problem? I thought error messages are >>> notaninterface? Or is this the exception of the >>> common rule? > > I thought I had already responded to this issue earlier, but I may have > just responded to Olga instead of the entire list... > > The format and exact contents of diagnostic messages are usually not > specified by the standards (i.e. not-an-interface), but the fact that a > diagnostic message is or is not written to STDERR is an interface. > >> >> The code in question is at >> http://src.opensolaris.org/source/xref/onnv/onnv-gate/usr/src/cmd/sort/common/check.c#203 >> >> which looks like it enables messages for the non-xpg4 in the case of either >> -c >> or >> -c -u >> options. >> >> http://www.opengroup.org/onlinepubs/9699919799/utilities/sort.html >> >> shows something new: -c for error messages and -C (capital C) same but with >> no error messages: >>> -c >>> Check that the single input file is ordered as specified by the arguments >>> and the collating sequence of the current locale. Output shall not be sent >>> to standard output. The exit code shall indicate whether or not disorder >>> was detected or an error occurred. If disorder (or, with -u, a duplicate >>> key) >>> is detected, a warning message shall be sent to standard error indicating >>> where the disorder or duplicate key was found. >>> -C >>> Same as -c, except that a warning message shall not be sent to standard >>> error if disorder or, with -u, a duplicate key is detected. >> >> As far as I can tell, the current code does not implement the -C (capital C) >> option. >> >> The availability of both those options would comply with the latest standard, >> AFAIK, and would resolve the difference between /usr/bin/sort and >> /usr/xpg4/bin/sort (by allowing both behaviors by command line option, and >> with the existing /usr/bin/sort behavior effectively the default (i.e. if one >> uses -c rather than the new -C). >> >> That's my take, anyway... > > Earlier versions of the standard described the -c (lower-case) option as: > Check that the single input file is ordered as specified > by the arguments and the collating sequence of the current > locale. No output shall be produced; only the exit code > shall be affected. > and did not specify the -C (upper-case) option. The "No output shall > be produced" covers both STDOUT and STDERR; so the current source > matches the earlier standards. Oracle has not announced any plans to > implement the changes needed to align with the current standard. (And > Sun laid off its POSIX/UNIX standards conformance group over a year ago.) > > - Don > > _______________________________________________ > shell-discuss mailing list > [email protected] > http://mail.opensolaris.org/mailman/listinfo/shell-discuss > -- , _ _ , { \/`o;====- Olga Kryzhanovska -====;o`\/ } .----'-/`-/ [email protected] \-`\-'----. `'-..-| / Solaris/BSD//C/C++ programmer \ |-..-'` /\/\ /\/\ `--` `--` _______________________________________________ shell-discuss mailing list [email protected] http://mail.opensolaris.org/mailman/listinfo/shell-discuss
