Apache OpenOffice (AOO) Bugzilla – Full Text Issue Listing |
Summary: | FreeBSD, MacOS X's sed doesn't support -r | ||||||
---|---|---|---|---|---|---|---|
Product: | porting | Reporter: | pavel | ||||
Component: | code | Assignee: | rene | ||||
Status: | CLOSED FIXED | QA Contact: | issues@porting <issues> | ||||
Severity: | Trivial | ||||||
Priority: | P5 (lowest) | CC: | issues, maho.nakata, quetschke | ||||
Version: | current | ||||||
Target Milestone: | OOo 2.0 | ||||||
Hardware: | All | ||||||
OS: | All | ||||||
Issue Type: | DEFECT | Latest Confirmation in: | --- | ||||
Developer Difficulty: | --- | ||||||
Attachments: |
|
Description
pavel
2005-04-06 21:36:37 UTC
yes FreeBSD's sed doesn't support -r (also 5.3-RELEASE) also MacOS X's sed doesn't support it -> JAVA_HOME in configure9 is not set at all. vq: can you look. the stuff athe kaffe check is not important because kaffe doesn not really work anyhow but the $SED -n -r's in the java stuff are important abnd will break FreeBSD/MacOS X? Can that be rewritten to not require -r? This looks like it tries to subsitute exe in all possible lower/uppercase combinations so we can do three SEDs, can't we? I'll have a look ... Created attachment 24818 [details]
Patch to work around sed -r
The previous patch works around that problem, but are we sure that $SED -n "s,//*bin//*java,,p" really does what it is designed to do? If I'm not mistaken that matches /opt//////////bin/////////////javasomething or /usr/bin/java . This is only one of the two places where you added -r.. Anyway; If it accepts that path you mentioned, it's suboptimal but IMHO shouldn't be that big problem since no one will get this path on sane systems anyway... > This is only one of the two places where you added -r.. Oeh, but my patch (issue 46572) only added one -r. The other one is in the KAFFE_VER part: KAFFE_VER=`$JAVAINTERPRETER -version 2>&1 | $EGREP " Version:" | $SED -r "s/.* Version: ([[0-9\.]]*).*/\1/"` I don't know what kaffe outputs, but can you try (replace the $SED part with): $AWK '/.*Version: [[0-9.]]*/ {print gensub (".*Version: ([[0-9.]]+).*","\\1",g);}' (One line) yeah, error. it's two (including the kaffe check, which can be ignored for now) -> fixed in configure9 close issue. |