site_archiver@lists.apple.com Delivered-To: darwin-dev@lists.apple.com Thread-index: AclJ0PsBduHLe/k36EG8mrPL5L+FEg== Thread-topic: Leopard sed(1) behavior hangs... User-agent: Microsoft-Entourage/12.14.0.081024 [had submitted this to the Unix-porting list; it was suggested I try this list] I have been moving my build environment to Leopard (fm Tiger) for the past couple of weeks on a part-time basis. The Xcode projects (fm 2.5.1) all required changes - to deal with SRCFILE, and some new gcc visibility warnings. But I finally got that all cleaned up. What is killing me now are the numerous sh scripts that we use for testing on multiple platforms. I got beyond the tr(1) issues - where new complaints about illegal bytes sequences were encountered. Nice. But sed(1) seems to completely lockup in many cases; the process list tells that it is hanging and consuming 99% of its particular cpu. It appears that it must be hung" upon the initial open - cuz its output stream has produced nothing. I initially assumed that this was a similar issue as tr(1) (ie, text encoding) - and I tried a few things to normalize the incoming data stream to a particular text encoding (using iconv(1)). While that would seem to fix some hangs; other tests would then lock up. Sampling sed (using Shark) did not tell me anything; there were no symbol names revealed... Anyone have any comments on this behavior? Suggestions to remedy it? -harry _______________________________________________ Do not post admin requests to the list. They will be ignored. Darwin-dev mailing list (Darwin-dev@lists.apple.com) Help/Unsubscribe/Update your Subscription: http://lists.apple.com/mailman/options/darwin-dev/site_archiver%40lists.appl... This email sent to site_archiver@lists.apple.com