Mailing-List: contact cygwin-apps-help AT sourceware DOT cygnus DOT com; run by ezmlm Sender: cygwin-apps-owner AT sourceware DOT cygnus DOT com List-Subscribe: List-Archive: List-Post: List-Help: , Delivered-To: mailing list cygwin-apps AT sources DOT redhat DOT com Message-ID: <023301c0c819$181d87e0$0200a8c0@lifelesswks> From: "Robert Collins" To: "Billinghurst, David \(CRTS\)" , , References: <8D00C32549556B4E977F81DBC24E985DC839 AT crtsmail1 DOT technol_exch DOT corp DOT riotinto DOT org> Subject: Re: gcj && libjava for cygwin Date: Thu, 19 Apr 2001 01:06:04 +1000 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.50.4133.2400 X-MimeOLE: Produced By Microsoft MimeOLE V5.50.4133.2400 X-OriginalArrivalTime: 18 Apr 2001 14:58:38.0629 (UTC) FILETIME=[0CD2DD50:01C0C818] ----- Original Message ----- From: "Billinghurst, David (CRTS)" To: "'Robert Collins'" ; ; Sent: Thursday, April 19, 2001 12:43 AM Subject: RE: gcj && libjava for cygwin > OK. I have duplicated this, with a couple of variations on the way. > * a couple more patches below > * -lpthreads are defined by configure.in. Will submit patch for this > * I think Ican work around IP6 issues > * WinMain AT 16 problem comes from linking -lcygwin to early > > What I don't understand where _Jv_Sjlj_Throw and JNI_OnLoad are defined. > > > Needed one strncasecmp in gcc/java/parse.c (or more correctly parse.y) > > --- parse.c.orig Tue Apr 17 23:55:39 2001 > +++ parse.c Wed Apr 18 19:47:12 2001 > @@ -6026,7 +6026,8 @@ > ; > if (f[0] == '/' || f[0] == DIR_SEPARATOR) > f++; > - if (strncmp (IDENTIFIER_POINTER (raw_name), > + /* DB CYGWIN */ > + if (strncasecmp (IDENTIFIER_POINTER (raw_name), > f , IDENTIFIER_LENGTH (raw_name)) || > f [IDENTIFIER_LENGTH (raw_name)] != '.') > parse_error_context > > In libjava/java/lang/natSystem.cc I think this is already fixed in current cygwin snapshots. (If not, I can roll a patch for cygwin in minutes.) Thus no need for the _tzname workaround. Could you try a recent snapshot? > --- natSystem.cc.orig Tue Apr 17 23:23:23 2001 > +++ natSystem.cc Wed Apr 18 20:58:19 2001 > @@ -265,7 +265,11 @@ > // issue exists in java/util/natGregorianCalendar.cc. > tzoffset = 0L; > #endif > +#ifdef __CYGWIN__ > + tzinfo = _tzname; > +#else > tzinfo = tzname; > +#endif > > if ((tzoffset % 3600) == 0) > tzoffset = tzoffset / 3600; > Rob