From cca!think!ames!aurora!jaw Tue Mar 31 10:17:52 1987 Received: by mirror.TMC.COM (4.12/UUCP-Project/rel-1.0/08-20-86) id AA24639; Tue, 31 Mar 87 10:17:41 est Received: by Think.COM; Tue, 31 Mar 87 01:10:02 EST Received: Mon, 30 Mar 87 22:07:48 PST by ames.arpa (5.45/1.2) Received: Mon, 30 Mar 87 22:08:02 PST by ames-aurora.arpa (5.51/1.2) Date: Mon, 30 Mar 87 22:08:02 PST From: James A. Woods Message-Id: <8703310608.AA00588@ames-aurora.arpa> To: ames!think!cca!mirror!sources Subject: [ef]?grep instructions [resent mail lossage] Status: R please ack receipt of *this* message and two-part shar immediately following. ames!jaw -------- >From jaw Thu Mar 19 17:48:22 1987 Received: Thu, 19 Mar 87 17:48:16 PST by ames-aurora.arpa (5.51/1.2) Date: Thu, 19 Mar 87 17:48:16 PST From: James A. Woods Message-Id: <8703200148.AA20572@ames-aurora.arpa> To: ames!think!cca!mirror!sources Subject: [ef]?grep -- about the next two (sharfile) messages Status: RO egrep.shar.1 is 56129 bytes; it contains [ef]?grep code and documentation. egrep.shar.2 is 55583 bytes; it contains henry spencer's regexp() routines -- it's not any different than what's already in the archive. i'm sending you both so you can just say sh egrep.shar.1 sh egrep.shar.2 make sh eg.shell to test it out (bsd 4.3 people are instructed to apply a diff to go full bore with the -i option). now, if you choose not to re-post shar.2, the make assumes regexp() source is in the same directory as [ef]?grep source. of course, folks will have to piece together the three regexp() fixes together with the orig. from the archive if it's not re-posted, or mail to hoptoad!gnu for a newer one. the only tricky thing about the egrep.shar.1 is that the application of 'sh' will run some 'tr' commands at the end (see last few lines) to change two files (k.pat & k.test) in the sharchive, so any re-packaging you do *must* contain these commands. they "translit" some parity-marked bytes of japanese kanji for tests -- if i included the files raw, the mailers (to you or to the net) choke, and besides, i can't assume people are running the korn shell which deals correctly with eight-bit chars to unshar things. it's the only way i know to send parity bytes without assuming the existence of the improbable, or at least 'encode/decode'. the best way for you is to keep egrep.shar.1 intact, or re-shar only after preserving/cutting/pasting "k.pat/k.test" from the original. my work phone is (415) 694-6436. sincerely, James A. Woods (hplabs!ames!jaw, or jaw@ames.arpa)