home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!ames!saimiri.primate.wisc.edu!sdd.hp.com!elroy.jpl.nasa.gov!usc!news!netlabs!lwall
- From: lwall@netlabs.com (Larry Wall)
- Newsgroups: comp.lang.perl
- Subject: Re: Perl bug using read
- Message-ID: <1992Nov7.001313.25650@netlabs.com>
- Date: 7 Nov 92 00:13:13 GMT
- References: <9830012@hpfcso.FC.HP.COM> <1992Oct29.141034.2649@dcs.qmw.ac.uk> <MERLYN.92Oct29095217@romulus.reed.edu>
- Sender: news@netlabs.com
- Organization: NetLabs, Inc.
- Lines: 25
- Nntp-Posting-Host: scalpel.netlabs.com
-
- In article <MERLYN.92Oct29095217@romulus.reed.edu> merlyn@ora.com (Randal L. Schwartz) writes:
- : The number and string value of a scalar are held separately. Ideally,
- : every assignment to a string invalidates the number, and vice versa.
- : But there have been bugs in the past where the cache was not properly
- : blown, and this is probably an example of that.
-
- Bingo.
-
- : Don't expect a patch for 4.035 from Larry on this... He's too deep
- : into Perl 5.0, so we'll have a whole new set of bugs to clean. :-)
-
- As it happens, I'd carried this particular bug over. It'll be fixed in 5.
-
- : (Larry tells me that nearly every line of Perl 5.0 is *new*.)
-
- If you count "heavily munged" as "new"...
-
- Essentially, the tokener is about half munged, the parser is totally new
- except for some of the yacc grammar, and the run-time system is an
- unrecognizable transmogrification of the old, with maybe 1/3 new code.
- So far it actually parses the entire TEST program, and executes the
- first 20 lines or so before blowing up. When I get it back to where
- it'll pass the test suite, then I can *really* start having fun.
-
- Larry
-