Commit Graph

266 Commits

Author SHA1 Message Date
kurt
a5bc96319a - use HW_{USER|PHYS}MEM64. okay naddy@
- on amd64 only do implicit null optimizations on offsets landing in the
first page of memory.
2008-03-17 03:21:25 +00:00
kurt
b5be1f69c8 - use HW_PHYSMEM64. okay naddy@ 2008-02-28 14:54:56 +00:00
kurt
43fe843d56 - mark as BROKEN 'Work-in-progress and outdated' so users understand this
port is not finished yet. also remove a MASTER_SITE. okay naddy@
2008-02-26 18:26:23 +00:00
kurt
6284762954 - update to BSD patchset 8 based off Sun's JRL licensed Update 14 source:
http://java.sun.com/j2se/1.5.0/ReleaseNotes.html#150_14
- NOTE: patchset 8 includes amd64 plugin support now.
- adjust pkg/MESSAGE-* files for change in plugin support and add missing
  JRL licensing blurb for the jre package.
2008-02-15 15:14:37 +00:00
kurt
b3296e52b5 - remove 1.3-linux now that it is no longer needed for building 1.4
- update 1.3+ MODJAVA_RUN_DEPENDS to account for 1.3-linux removal
okay ian@ pval@
2008-02-14 21:48:43 +00:00
kurt
e0c8972d72 - switch to kaffe bootstrap. linux emul no longer needed to build.
- update to patchset 8. mostly just removal of patches committed
  upstream a long time ago.
okay brad@
2008-02-13 13:53:50 +00:00
kurt
d658fc57aa don't install libnative_chmod.so. it is an unused linux shared lib
okay ajacoutot@
2008-01-18 16:08:05 +00:00
kurt
54a9930f43 - don't install libnative_chmod.so. it is an unused linux shared lib
- install JRL LICENSE in second location in JDK
okay ajacoutot@
2008-01-18 15:52:55 +00:00
kurt
447d54c5a1 update fetch manually url xalan. reported by James Hartley 2007-11-30 14:57:23 +00:00
kurt
ccfa8aa068 add 1.6 2007-11-19 03:20:55 +00:00
kurt
5d62878859 Initial import of Java (TM) 2 SDK, Standard Edition Version 1.6.0 with
amd64 plugin support. 64-bit plugin hacking by myself and Jung-uk Kim.
Credit for the majority of the 1.6 BSD porting goes to Greg Lewis who
back ported my 1.7 hotspot patches to 1.6 and forward ported the 1.5
j2se patches to 1.6. Thanks to several testers who provided helpful
feedback. okay ian@
2007-11-19 03:19:43 +00:00
kurt
c7b570efa8 s/ERRORS/BROKEN/ to fix mirror-maker per espie@ and nikolay@ 2007-11-09 15:27:55 +00:00
kurt
d03e52ac71 - another master site url 2007-11-07 17:17:41 +00:00
kurt
8778311747 Remove interactive license verification since it causes issues due to
lack of stdin availability when built indirectly as a depend of another
port. Instead of interactive license verification solely use the
ACCEPT_JRL_LICENSE var. If it's not correctly set error out and let the
user know what to do. Cluestick and okay espie@
2007-11-07 17:04:36 +00:00
kurt
a51d7f9f59 - migrate forward a few necessary networking patches from 1.5 that fell
through the cracks.
2007-11-07 03:53:53 +00:00
kurt
f35ddca1d5 Exit prompt loop when there's nothing to read from stdin. This can happen
when stdin is redirected someplace unanticipated. This fixes the infinite
prompt loop when the port is built as a build depend of another port
(reported by Martin Toft <mt at martintoft.dk>). Still to do is fix
bsd.port.mk to not redirect stdin for build depends...
2007-10-29 01:22:34 +00:00
kurt
c347040e92 - update to BSD patchset 7 based off Sun's JRL licensed Update 13 source:
http://java.sun.com/j2se/1.5.0/ReleaseNotes.html#150_13
- Do a better job respecting Sun's JRL license by using interactive
  license verification, installing the correct license and include a
  license blurb in the pkg MESSAGEs. (Interactive license verification
  script from FreeBSD/Greg Lewis)
- fix a packing problem with the no_web FLAVOR.

NOTE: This port is now IS_INTERACTIVE=Yes. After agreeing to the JRL
license, users can put ACCEPT_JRL_LICENSE=Yes into /etc/mk.conf to
automatically accept the license and make the port IS_INTERACTIVE=No.
Bulk builders will need to do this so things like OpenOffice get built.
okay naddy@
2007-10-25 20:51:41 +00:00
kurt
d9e46b09e2 - fix patch borked by cvs changing cvs tag upon commit (i.e. $Revision$).
breakage reported by RD Thrush rd at thrush.com
2007-09-17 13:31:20 +00:00
merdely
376b75df06 Remove quotes surrounding COMMENT*/PERMIT_*/BROKEN 2007-09-16 02:52:56 +00:00
kurt
324a878467 Remove BROKEN and update port to b19. The port is fully 1.7 now (both
hotspot and j2se). The j2se patches are from the bsd java repositories
1.5 -> 1.6 (performed by Greg Lewis) -> 1.7 by myself. j2se patches are
in sync with 1.6 patchset 1 plus the Runtime.exec fix from Michiel Boland
via FreeBSD PR http://www.freebsd.org/cgi/query-pr.cgi?pr=115557.

Also from FreeBSD: interactive JRL license verification. By default the
port is IS_INTERACTIVE=Yes. Add ACCEPT_JRL_LICENSE=Yes to /etc/mk.conf
to permanently accept the JRL license.

At some point in the future I will work on switching this back to
OpenJDK/GPLv2 src. That largely depends on Sun's progress at removing more
of the binary plug bits so that stubbing out the remainder is less
cumbersome and time consuming.

Testing, comments and okay ian@
2007-09-06 15:46:18 +00:00
kurt
1afa07f15b let's not give the impression 4.2 has a working 1.7 port. mark it BROKEN
for now. okay ian@ naddy@
2007-08-07 16:07:22 +00:00
kurt
c882a87b87 - upgrade to Update 12 srcs which fixes the following bugs:
http://java.sun.com/j2se/1.5.0/ReleaseNotes.html#150_12
- remove patches I committed into patchset 6
2007-07-27 19:29:51 +00:00
kurt
dacf31c5b4 make this work with motif > 2.1 too 2007-07-18 17:53:07 +00:00
kurt
f71ecfc678 regen patches 2007-07-18 15:09:52 +00:00
kurt
515f2a2739 regen patches 2007-07-18 14:19:27 +00:00
kurt
d46a2a67a5 - remove patch that should have been removed with last commit 2007-07-13 23:25:30 +00:00
kurt
6b86e4383e - update to build 15
- switch to the JRL licensed src. why? because it will allow me to complete
the port. OpenJDK/GPLv2 src isn't complete and the parts that we don't have
are only released in binary form with a defend and indemnify clause. When
all of the jdk is GPLv2 I will switch back to OpenJDK/GPLv2 src.
- copy the Linux src files for the rest of the jdk so i can begin to patch
them. One src file is based on solaris code since that suits us better as
a starting point.
2007-07-13 22:52:35 +00:00
kurt
ca0e01e493 - don't link jvmti demo native libs with $ORIGIN in RPATH since ld.so
does not support it yet.
- fix writes to HOME complaints
- clean up post-build target
2007-07-06 01:35:57 +00:00
kurt
ff364b1c3b - actually enable version scripts on libjvm which fixes a symbol conflict
with recent xulrunner which is used by eclipse's internal swt-browser.
(patch missed in last update)
- adjust url to update 11 download which has been move to the archive.
2007-07-01 20:34:02 +00:00
ian
614ac9a90a Better url in FETCH_MANUALLY, ok kurt@ 2007-06-28 03:20:20 +00:00
kurt
450f231ad4 - update port to be based of Sun's 1.5.0 Update 11 src which includes many
bug fixes since the initial release of the 1.5 jdk. I'm sure some of the
fixes are security related but I wasn't able to find a concise list. All
the bugs fixes listed here from update 11 and lower should be fixed by
this update: http://java.sun.com/j2se/1.5.0/ReleaseNotes.html#150_11
- NOTE: License change: used to be SCSL, now is JRL (Java Research
License). Still no packages with this license but at least we get the bug
fixes and security fixes now: https://tiger.dev.java.net/
- enable version scripts on libjvm which fixes a symbol conflict with
recent xulrunner which is used by eclipse's internal swt-browser.
2007-06-27 21:51:09 +00:00
kurt
01ba206cd1 expand BUILD_DEPENDS to allow building ports with jdk-1.5.0 update 11 2007-06-27 21:33:33 +00:00
steven
05565e86a6 linux-getcwd -> linux-__getcwd after the change in linux compat.
fixes build of jdk 1.4 with systrace.

"makes sense" sturm@
2007-06-06 04:33:33 +00:00
kurt
0d2c9182de . update to b12 sources (just building hotspot)
. adjust SIGRTMIN SIGRTMAX defines - from Jung-uk Kim <jkim@freebsd.org>
. use SIGSEGV || SIGBUS in amd64 signal handling code
2007-05-09 15:48:31 +00:00
kurt
a6b2a31df6 Initial import of jdk 1.7 (a.k.a. OpenJDK 7) work in progress. So far Sun
has released the hotspot portion of 1.7 under GPLv2. j2se and deploy
(plugin & webstart) will come later. No packages yet since j2se is copied
in from 1.5.

Note: hotspot for BSD in previous versions of the jdk were based on a
mixture of Solaris code and Linux code (for amd64). In this version I have
selected Linux as the base and only ported of small parts of the Solaris
code as needed. Using Linux code as the base has significantly reduced the
code divergence for BSD support and should make maintaining the port
easier.

okay ian@, go for it espie@
2007-05-08 17:06:18 +00:00
kurt
7083038c36 fix bug in the filtering of corrupt environment vars 2007-05-01 02:05:46 +00:00
kurt
3cb2b7ac3b - work-around libpthread limitations in javaForkAndExec() by adjusting the
child process to not set file descriptors back to blocking when closing
  them. This fixes the netbeans build failure since a libpthread fix is not
  evident.
- update the North America time zone file to account for the tz changes
  (from ian@)
- use X11BASE
- SHARED_ONLY=Yes
2007-04-13 00:29:48 +00:00
kurt
3a2afa990f - fix the build for good on systems with >= 4G memory by limiting
the jdk's view of physical memory to datasize rlimit max
- remove and adjust old memory adjustments for OpenBSD
- shrink overhead on amd64 making more memory available for use
ok wilfried@
2007-04-08 12:43:06 +00:00
espie
470294650d base64 distinfo with SHA256 2007-04-05 15:37:40 +00:00
kurt
d26de13317 install javaws man page for amd64 too. fixes a packaging problem on amd64
found by wilfried@. okay wilfried@
2007-02-04 15:57:06 +00:00
kurt
9203a7311b - fix mistake that broke OOo build. the jdk needs to search ${PREFIX}/lib
by default.
- regen plist and move javaws man page to no-no_web PFRAG files where it
belongs.
2007-02-01 21:00:52 +00:00
kurt
064d60b115 Security fix for CVE-2007-0243:
Buffer overflow allows applets to gain privileges via a GIF image with a
block with a 0 width field, which triggers memory corruption and possible
execution of arbitrary code.
2007-01-31 18:14:36 +00:00
kurt
10c96fb157 Security fix for CVE-2007-0243:
Buffer overflow allows applets to gain privileges via a GIF image with a
block with a 0 width field, which triggers memory corruption and possible
execution of arbitrary code.
2007-01-28 04:45:04 +00:00
kurt
a5d5e8fec5 distfile rerolled with milestone bumped. 2007-01-24 21:05:10 +00:00
kurt
20ef89247d Update to patchset 4 which includes a security fix for CVE-2007-0243
Buffer overflow allows applets to gain privileges via a GIF image with a
block with a 0 width field, which triggers memory corruption and possible
execution of arbitrary code.

Also included in patchset 4 are corrections for building the jdk on amd64
systems with 4G+ memory and the inclusion of the mozilla headers needed to
build the plugin without an external dependency on the old mozilla suite.
2007-01-23 14:14:37 +00:00
kurt
80d3cd2c49 Change the default package for 1.3+ and 1.4+ ports to devel/jdk/1.5 on
i386. This fixed a 'FORCE_UPDATE=yes make install' issue noticed by brad@
and further encourages users to use the 1.5 jdk which works the best.
2007-01-16 02:47:14 +00:00
steven
b6d439743e use BASE_PKGPATH instead of homegrown variable, no bumps needed 2006-12-14 09:07:55 +00:00
kurt
f6da9a94ff On i386 all source built java ports will use the 1.5 jdk to build. Runtime
compatibility is preserved through the use of the javac -source and
-target arguments. These arguments are now required for any new source
built java ports.
2006-11-28 14:28:25 +00:00
steven
4db7fc3254 take flavors into account in @pkgpath 2006-11-24 21:35:29 +00:00
steven
330f0193b5 take flavors into account in @pkgpath
ok espie@
2006-11-24 21:30:13 +00:00
kurt
d3fbad8b83 new MULTI_PACKAGES 2006-11-22 19:42:54 +00:00
kurt
aa6ff6836d new MULTI_PACKAGES 2006-11-22 14:58:03 +00:00
espie
23011e3982 tweak for new MULTI_PACKAGES 2006-11-20 19:28:18 +00:00
kurt
3ebfead8e1 - attempt to fully send/receive packets in socketTransport.c by handling
EINTR and continue to send/receive the rest of the packet. corrects:
"ERROR: transport error 202: recv error: Interrupted system call
["transport.c",L41]" seen while debugging large projects in eclipse.
from niklas@
- Prevent problems opening RandomAccessFile with "rws" as the mode by
defining O_SYNC and O_DSYNC if they aren't defined.  In particular we
were previously defining O_SYNC bogusly to the same value as O_EXCL.
from FreeBSD
- close two more pthread_addr leaks
2006-11-17 20:19:33 +00:00
kurt
1ec8b79489 fix ownership. okay steven@ 2006-11-16 15:06:18 +00:00
kurt
7a8823ecec use system libz instead of internal libz 2006-11-15 17:27:40 +00:00
steven
87a90c296d USE_X11 2006-11-11 21:04:27 +00:00
kurt
0f22d97d61 adjust depends so that 1.3+, 1.4 & 1.4+ level ports must use libz free
jdk/1.4 from now on.
2006-11-01 20:34:24 +00:00
kurt
7df1fb7f89 Back-port some jdk/1.5 changes:
- use system libz instead of internal one. internal libz conflicts with
system libz, for example apps using JNI & libz will fail (OOo).
- enable parallel hotspot compiles on mp systems (one per cpu). no pthreads
involved so its safe to do here.
- socket option corrections. may help Active MQ work.
2006-10-31 19:48:12 +00:00
sturm
6e3e04ec3b switch to fedora_base and adapt systrace.policy files accordingly 2006-10-19 09:39:58 +00:00
kurt
545e82d800 - remove disabling J malloc option now that pthreads stacks are mmap'ed
this is no longer a concern upon forking.
- use closefrom(2) after forking instead of less efficient heursitics
approach.
2006-10-03 13:36:28 +00:00
kurt
0f8014e857 - Fix jdk build when timezone is GMT+0. Reported by Schöberle Dániel
<Schoeberle.Daniel at aamtech.hu>
- Fix short Thread.sleep() issue. Reported by Mr David SH Rosenthal
<dshr at stanford.edu>
2006-09-26 17:48:19 +00:00
kurt
e7d3816ed9 expand RUN_DEPENDS to include kaffe and jamvm where appropriate so if
someone wants to run a java port with them they don't need to first build
one of Sun's jdks.  okay nikolay@
2006-08-05 16:38:20 +00:00
espie
057dd9afe4 new libspecs 2006-08-01 22:55:52 +00:00
kurt
e8329de2c0 Prevent the jvm from crashing when it encounters a font with an
illegal glyph name (like in x11/freefonts). Problem reported by Jan Zach.
2006-07-28 13:18:45 +00:00
kurt
15f1fe1f7e indent defines too so that this is more readable. 'looks sane' nikolay@ 2006-07-28 13:06:24 +00:00
kurt
aa74a6e1f4 - speed up builds on mp systems by using parallel compiles (one per cpu)
- increase default thread stack size on amd64
- filter out corrupt env vars with no name
- Properly deal with error cases in transferTo0(). Corrects non-blocking
IO case.
All but first one from FreeBSD
2006-07-10 13:17:39 +00:00
steven
4db61bb851 add client vm files in jre package. this unbreaks the browser plugin.
ok kurt@
2006-07-01 09:06:00 +00:00
kurt
4b5dcd0f44 - use /dev/srandom in NativePRNG replacing /dev/random.
/dev/random use noticed by Sigfred Håversen.
2006-06-20 02:02:48 +00:00
kurt
7209ade622 - fix reoccurance of KOI8_U charset NPE's for ALT_BOOTSTRAP (default).
since KOI8_U doesn't come with the Solaris class files, make sure it is
built in the recompile stage. Reported by Matt Jibson.
- add a useable cacerts file and use that instead of current empty one.
charset portions tested by and okay ian@.
2006-06-06 03:02:42 +00:00
kurt
bdb052bb9e add amd64 support. tested & ok ian@ 2006-05-28 00:14:39 +00:00
steven
dbd3e01ddf add a missing man page, update-plist, bump PKGNAME. okay kurt@ 2006-05-26 17:32:21 +00:00
ian
a5b4364f7f Remove extra line committed in error, unbreaks JRE packaging 2006-05-16 03:39:41 +00:00
ian
31e312eb64 Enable JDK/JRE build on amd64. Switch to bootstrapping using Kaffe + Solaris JDK binary,
obviating need to build with JDK Linux compat (which was i386 only anyway).
Add native_bootstrap pseudo-flavor so you can build with a native Sun JDK if you
have one (tested by building this way using JDK built with default Kaffe-based build).
Work by kurt@, tested by him on i386 and me on amd64. OK kurt@.
2006-05-15 18:11:22 +00:00
ian
7b0be3a97d fix typo; 'ok kurt@; no pkg bump needed' 2006-04-28 22:57:11 +00:00
kurt
d5dbd1d194 - add a new pseudo flavor called alt_bootstrap which uses kaffe, Xalan and
solaris class files to bootstrap the jdk build. this avoids the need for
devel/jdk/1.4 to bootstrap and will provide a way to bootstrap the jdk on
amd64 and sparc64 when kaffe runs there.
- fix a keyboard entry problem when numlock is on
- fixup socket option code similar to FreeBSD & NetBSD
2006-04-28 20:56:15 +00:00
kurt
d20d578fd8 Update to patchset 3. Patchset 3 includes several bug fixes that improve
compatiabily and stability which came from work sponsored by the FreeBSD
Foundation.
2006-04-11 00:45:41 +00:00
sturm
645edfce2d fix build with systrace 2006-04-01 13:58:49 +00:00
kurt
bc308600a9 add sparc to ONLY_FOR_ARCHS for ports with MODJAVA_VER=1.3/1.3+ 2006-03-20 14:42:37 +00:00
kurt
107e7546a3 - enter 1.3.1 on sparc. due to the small datasize ulimit on sparc -Xmx60
is about as high as you can go. tomcat v4 and demos work but a bit slowly.
2006-03-20 14:40:24 +00:00
kurt
225683f4ff - don't mmap on top of malloc'ed memory. this fixes a large memory leak
while creating/destroying threads.
- add missing pthread_attr_destroy while creating threads. closes a
small memory leak.

Memory leaking problems reported by Carsten Raskgaard and also
David SH Rosenthal from the LOCKSS project.
2006-03-16 16:09:01 +00:00
kurt
1fe7cb2239 stablity fix for os_sleep. patchset 2 introduced a buggy version of
os_sleep based on nanosleep(). revert to patchset 1 version based on
poll(). fixes issues with large sleep values. okay pvalchev@
2006-02-22 22:52:44 +00:00
david
9e2fccfe54 typo in DESCR; bump PKGNAME; ok kurt@ 2006-02-16 05:46:21 +00:00
kurt
f4b38ea4b0 remove HAVE_DPS from MAKE_ENV. It appears that the build was ignoring it
since it wasn't in WANTLIB, but bump pkg anyway.
2006-02-01 01:06:40 +00:00
kurt
c421027c71 Don't build with HAVE_DPS now that it is gone. Also remove dps from
WANTLIB. noticed by and okay naddy@
2006-02-01 01:04:29 +00:00
kurt
e7551ce9e3 jlong isn't defined in all cases java_md.h is included. how did this
slip by me?
2006-01-23 05:49:50 +00:00
kurt
88736cd656 - fix a series of datatype assumptions in i486/hotspot (intptr_t != int
on OpenBSD)
- fix implicit declaration's by adding headers or prototypes as needed.
2006-01-22 22:46:30 +00:00
kurt
75b6e7ae64 don't try to enable kern.emul.linux sysctl for FLAVOR=native_bootstrap.
braino from last commit. noticed by Matthias Killian.
2006-01-19 22:02:48 +00:00
kurt
4768d88a57 remove conflicting read() prototype and correct conflicting prototypes for
sysSendTo and sysRecvFrom. unbreaks build.
2006-01-18 15:12:55 +00:00
kurt
7a27922713 - recent w^x changes combined with random malloc/mmap exposed a previously
undiscovered bug in the SignatureHandlerLibrary. Code generated at runtime
was put into malloc()'ed memory and later attempted to execute it without
first making it executable with mprotect(). Debugging suggestions and help
from tom@, weingart@, marc@, millert@. Thx!
- make sure linux emulation is on for the build phase. needed if you reboot
after extracting.

okay tom@
2006-01-18 15:08:16 +00:00
millert
be2c89c95d No need to protect the type-correct prototypes with #ifdef _LP64
Fixes problems caused by types changes.  OK kurt@
2006-01-17 14:28:19 +00:00
kurt
bf8f82ebbc Disable Junk malloc option for java. java mprotects portions of the thread
stack. When forking, the child process deallocates the threads stacks it
inheritated from the parent. The junk malloc option then attempts to write
to the deallocated memory and segfaults because its not all writable. The
child proc stays in a SIGSEGV loop when this happens.  okay ian@
2006-01-02 03:48:40 +00:00
kurt
56509860b6 - fix build for systems with > 2G physical memory (signed int value was
assigned to uint64).
- set some fall-back max-memory values to reasonable values for OpenBSD.
Problem reported and update tested by Douglas Santos.
2005-11-30 02:21:32 +00:00
kurt
a174f4d822 - add plugin and javaws support. disable with no_web flavor.
Demos all work but deadlocking the vm under load has been
noted. Both MToolkit and XToolkit work ok so far.
- fix segfault with MToolkit where processTree was called
with to==NULL.
- fix NPE in plugin paintGrayBox where g==null
- fix javaws startup (from FreeBSD)
- forgot to mention in last commit: Forte Peformance Tools
collector is no longer disabled.
2005-11-23 19:39:13 +00:00
kurt
d616aeb0d0 forgot to bump pkgname in last commit 2005-11-22 14:48:07 +00:00
kurt
7f5e5f0bdf - update to patchset 2. mostly removing patches now included in patchset.
- link awt_robot with -pthread
- disable a runtime Motif version check that incorrectly thought we were
using Motif 1.2 based on Xm major version. problem reported by Laurence
Tratt
- plugin and webstart comming later
2005-11-22 14:43:05 +00:00
kurt
13dc22b0d1 libc is not explicitly linked into jvm shlibs, so using RTLD_NEXT to find
a symbol in libc won't work. dlopen libc and find it directly.
2005-10-21 22:36:30 +00:00
sturm
0136620a75 sync distfile
as this is the only port, where we have a versioned distfile in
systrace.policy AND this jdk is EOL, there is no need for a generic
solution

discussed with and ok kurt@
2005-10-01 12:16:47 +00:00
kurt
2b1ffc1eb7 update to 1.3.1_16 from steven mestdagh <steven.mestdagh@esat.kuleuven.be> 2005-09-22 16:17:15 +00:00
espie
0031945c6e Fix a buglet in bsd.port.mk where the PSEUDO_FLAVORS get encoded into
the FULLPKGPATH, thus providing changes to packing-lists which shouldn't
happen, and making update more difficult.

Accordingly, bump all pkgnames with PSEUDO_FLAVORS, and provide an
update @pkgpath for the bug for most of them (left out the ones with 3
or 4 pseudo flavors for space constraints...)
2005-09-16 09:51:25 +00:00