First Last Prev Next    No search results available      Search page      Enter new bug
Bug#: 113589
Alias:
Product:
Component:
Status: RESOLVED
Resolution: FIXED
Assigned To: Java team <java@gentoo.org>
Hardware:
OS:
Version:
Priority:
Severity:
Reporter: bware <bWare@iWare.co.uk>
Add CC:
CC:
Remove selected CCs
URL:
Summary:
Status Whiteboard:
Keywords:

Filename Description Type Creator Created Size Actions
ebuild.diff Mainly updated dependencies for the new version patch bware 2005-11-25 14:14 0000 1.37 KB Details | Diff
ebuild.diff Corrected some mistakes in first diff. patch bware 2005-12-19 06:17 0000 1.60 KB Details | Diff
ebuild.diff Opps, should have used ${PN} for the dobin to I assume patch bware 2005-12-19 06:21 0000 1.60 KB Details | Diff
ebuild.diff Opps, should have used ${PN} for the dobin to I assume patch bware 2005-12-19 06:21 0000 1.60 KB Details | Diff
ebuild.patch Updated ebuild for new beta release patch bware 2005-12-31 07:12 0000 1.73 KB Details | Diff
ebuild.patch Added optional mathml support patch bware 2006-02-06 06:07 0000 2.15 KB Details | Diff
index.fo an example fo document that causes my fop 0.91 beta to fail application/octet-stream Chris Chiasson 2006-02-06 09:38 0000 23.53 KB Details
index.fo Updated example application/xml bware 2006-02-06 10:09 0000 23.47 KB Details
0.91_beta-startscript.patch Startup script patch so we don't need to modify classpath patch bware 2006-02-06 10:36 0000 1.45 KB Details | Diff
ebuild.patch Updated ebuild to apply state script patch patch bware 2006-02-06 10:38 0000 2.12 KB Details | Diff
0.91_beta-startscript.patch Updated state script patch (had included batik twice?) patch bware 2006-02-06 10:45 0000 1.44 KB Details | Diff
Engineering_Optimization.fo fo file with mathml that causes the no layout maker error application/octet-stream Chris Chiasson 2006-02-06 21:00 0000 59.71 KB Details
fop-0.92_beta.ebuild Updated ebuild for .92 Beta text/plain bware 2006-05-13 07:49 0000 2.03 KB Details
0.92_beta-startscript.patch Patch the start script so fop can find it's dependencies text/plain bware 2006-05-13 07:50 0000 1.43 KB Details
fop-1.0_beta.ebuild dev-java/fop/fop-1.0_beta.ebuild text/plain Jiri Tyr 2006-05-19 09:17 0000 2.27 KB Details
1.0_beta-startscript.patch files/1.0_beta-startscript.patch patch Jiri Tyr 2006-05-19 09:18 0000 1.43 KB Details | Diff
fop-0.93_beta.ebuild fop-0.93_beta.ebuild text/plain Jiri Tyr 2007-01-09 11:00 0000 2.22 KB Details
fop-0.93_beta.tar.gz fop-0.93.ebuild and startscript patch text/plain Michael Kefeder 2007-01-20 20:41 0000 1.87 KB Details
fop-0.93_beta.ebuild fop-0.93_beta.ebuild text/plain Jiri Tyr 2007-01-22 14:11 0000 2.26 KB Details
0.93_beta-startscript.patch 0.93_beta-startscript.patch patch Jiri Tyr 2007-01-22 14:13 0000 1.76 KB Details | Diff
fop-1.0_beta.ebuild fop-1.0_beta.ebuild text/plain Jiri Tyr 2007-05-07 12:07 0000 2.37 KB Details
fop-0.94.ebuild fop-0.94.ebuild text/plain Jiri Tyr 2007-10-02 12:36 0000 3.47 KB Details
fop-1.0_beta.ebuild fop-1.0_beta.ebuild text/plain Jiri Tyr 2007-10-04 09:56 0000 3.54 KB Details
Create a New Attachment (proposed patch, testcase, etc.) View All

Bug 113589 depends on: Show dependency tree
Bug 113589 blocks:
Votes: 0    Show votes for this bug    Vote for this bug

Additional Comments: (this is where you put emerge --info)


Not eligible to see or edit group visibility for this bug.






View Bug Activity   |   Format For Printing   |   XML   |   Clone This Bug


Description:   Opened: 2005-11-25 14:12 0000
Apache have finally made a released from the new FOP trunk. This is quite a
major rewrite and defiantly unstable but it would be nice to have an ebuild.

As a starting point I will attache an ebuild diff that can build and install the
new version.

This is dependent on dev-java/commons-io-1.1 which I have submited a seperate
request for [Bug 113585].

The new build does not like having a previous version of Batik in my classpath,
and I get out of memory errors if I try and build the documentation.

Reproducible: Always
Steps to Reproduce:

------- Comment #1 From bware 2005-11-25 14:14:25 0000 -------
Created an attachment (id=73614) [edit]
Mainly updated dependencies for the new version

------- Comment #2 From Josh Nichols (RETIRED) 2005-12-15 18:49:21 0000 -------
From the announcemnt of the new version:
This release is the first after a big redesign effort on the whole FOP codebase.
This release is to be considered ALPHA quality and it is intended as a preview
release encouraging people to take a look at the new version and to provide
feedback to the developers. Please not only report to us problems you might
experience but also tell us if it works for you. If you find out that this
version works fine for you, you're welcome to use it but please test it
thoroughly as we don't consider this release ready for every production environment.

Because it is alpha, this is going to get a real low priority. Additionally, if
it does get added to the tree, it would probably be hard-masked until it is out
of the alphas.

A few comments on the diff:
* java-pkg_dojar build/fop.jar should be changed to java-pkg_dojar build/${PN}.jar
  I know the old ebuild was like this, but it is worth noting that you should
reuse P, PN, and PV whenever possible.
* Don't need to use newbin when the filename is the same as what will be
installed. Use dobin instead.
* Use >=dev-java/xerces-2.6.2 in depend. 2.7.0 is in portage now, and in the
same slot.

------- Comment #3 From bware 2005-12-19 06:17:05 0000 -------
Created an attachment (id=75091) [edit]
Corrected some mistakes in first diff.

Corrected some mistakes in first diff.

Here is an updated patch with you changes. Thanks for the help, still very new
to editing  ebuilds.

It also implements the important licence change I missed first time around.

I agree that the new version should be hard masked (especial as all the
previous versions are marked unstable). However I do think it is worth
including as FOP are not releasing bug fixes for the 0.20.5 branch, and the new
version is much more standards compliant.

One outstanding bug is no PDF generation under Java 1.5. (Though the is a patch
I haven't tried: http://issues.apache.org/bugzilla/show_bug.cgi?id=37108;
should we add this to the current ebuild?)

------- Comment #4 From bware 2005-12-19 06:21:31 0000 -------
Created an attachment (id=75092) [edit]
Opps, should have used ${PN} for the dobin to I assume

------- Comment #5 From bware 2005-12-19 06:21:34 0000 -------
Created an attachment (id=75093) [edit]
Opps, should have used ${PN} for the dobin to I assume

------- Comment #6 From Josh Nichols (RETIRED) 2005-12-26 15:32:49 0000 -------
0.91_beta was released yesterday.

------- Comment #7 From bware 2005-12-27 04:27:56 0000 -------
The beta seems to just be an SVN tag, for now the ebuild will still have to use
the alpha release.

------- Comment #8 From bware 2005-12-31 07:12:59 0000 -------
Created an attachment (id=75869) [edit]
Updated ebuild for new beta release

Sorry - didn't look hard enough. They have changed the location of thier
releases and have forgoten to updated the link to the source from their
website.

The new patch is just for the change of SRC_URI and HOMEPAGE, the build still
seems to work the same.

------- Comment #9 From Chris Chiasson 2006-02-05 12:19:23 0000 -------
(In reply to comment #8)
> Created an attachment (id=75869) [edit]
> Updated ebuild for new beta release
> 
> Sorry - didn't look hard enough. They have changed the location of thier
> releases and have forgoten to updated the link to the source from their
> website.
> 
> The new patch is just for the change of SRC_URI and HOMEPAGE, the build still
> seems to work the same.
> 

Can you add (optional?) ebuild support for the MathML extension that is in the
fop 0.91beta archive's examples directory? It requires jeuclid.jar to be added
to the examples/mathml/lib directory. I was able to download and compile the
jeuclid install from sourceforge, but there is no ebuild for jeuclid :/ I still
wasn't able to get the mathml extension to build though :[

------- Comment #10 From bware 2006-02-06 06:07:50 0000 -------
Created an attachment (id=79025) [edit]
Added optional mathml support

Created an initial ebuild for JEculid in bug [121822] and updated this ebuild
to optionally depend on it.

To run FOP with MathML support you need to ensure that JEculid is in your
classpath. To get usable results you also need to install the required fonts.

------- Comment #11 From bware 2006-02-06 06:09:49 0000 -------
http://bugs.gentoo.org/show_bug.cgi?id=121822

------- Comment #12 From Chris Chiasson 2006-02-06 09:38:49 0000 -------
Created an attachment (id=79037) [edit]
an example fo document that causes my fop 0.91 beta to fail

I'm getting this error on all fo documents I compile, no matter whether I use
the MathML enabled ebuild or not. The command executes successfully using fop
0.20

/usr/bin/fop ../path/to/index.fo -pdf ../path/to/index.pdf
Exception in thread "main" java.lang.NoClassDefFoundError:
org/apache/fop/cli/Main

My use flags from make.conf:
USE="unicode apache2 jimi jai mathml"

BTW, much thanks for making the MathML enabled install, bware!

Dependencies
(not quite sure why this command was able to also eliminate the 3rd and lower
level deps, but whatever - it worked)
equery g dev-java/fop | grep -v '^  `*'
dev-java/fop-0.91_beta:
`-- dev-java/fop-0.91_beta
 `-- dev-java/blackdown-jre-1.4.2.03 (virtual/jre-1.4)
 `-- dev-java/sun-jai-bin-1.1.2.01-r1 [ jai ]
 `-- dev-java/sun-jimi-1.0-r1 [ jimi ]
 `-- dev-java/avalon-framework-4.2.0
 `-- dev-java/batik-1.6
 `-- dev-java/xalan-2.7.0-r1
 `-- dev-java/commons-io-1.1
 `-- dev-java/commons-logging-1.0.4-r1
 `-- dev-java/servletapi-2.2-r2
 `-- dev-java/jeuclid-2.0 [ mathml ]
 `-- dev-java/ant-1.6.2-r6

------- Comment #13 From bware 2006-02-06 09:56:02 0000 -------
This is because FOP isn't in your class path. Currently we are still using the
fop launch script strait from the distribution, this probably needs patching to
be more Gentoo.

In the mean time either make sure that FOP (any any other stuff you need like
JEuclid, and Batik) are in you class path using:

java-config -A
fop,jeuclid,batik-1.6,commons-io-1,avalon-framework-4.2,commons-logging
&&
/usr/sbin/env-update
&&
source /etc/profile

------- Comment #14 From bware 2006-02-06 10:09:41 0000 -------
Created an attachment (id=79040) [edit]
Updated example

Note, FOP has also dropped the proprietary fox:outline in favour of the
standard fo:bookmark, you will need to change this before your example will
work.

------- Comment #15 From Josh Nichols (RETIRED) 2006-02-06 10:29:22 0000 -------
(In reply to comment #13)
> This is because FOP isn't in your class path. Currently we are still using the
> fop launch script strait from the distribution, this probably needs patching to
> be more Gentoo.
> 
> In the mean time either make sure that FOP (any any other stuff you need like
> JEuclid, and Batik) are in you class path using:
> 
> java-config -A
> fop,jeuclid,batik-1.6,commons-io-1,avalon-framework-4.2,commons-logging
> &&
> /usr/sbin/env-update
> &&
> source /etc/profile
> 

Adding things to the system classpath can cause unexpected problems when
emerging things. This usually happens when you have something slotted, like
bati-1.6, and you emerge something that wants batik-1.5. The package would
likely fail with batik-1.6 on the classpath before batik-1.5, because 1.6 has a
different api than 1.5.

Instead, it is preferable to patch the launcher script to add these to its
classpath.

------- Comment #16 From bware 2006-02-06 10:30:22 0000 -------
working on it ;-)

------- Comment #17 From bware 2006-02-06 10:36:41 0000 -------
Created an attachment (id=79046) [edit]
Startup script patch so we don't need to modify classpath

------- Comment #18 From bware 2006-02-06 10:38:49 0000 -------
Created an attachment (id=79047) [edit]
Updated ebuild to apply state script patch

------- Comment #19 From bware 2006-02-06 10:45:40 0000 -------
Created an attachment (id=79048) [edit]
Updated state script patch (had included batik twice?)

------- Comment #20 From Chris Chiasson 2006-02-06 21:00:43 0000 -------
Created an attachment (id=79084) [edit]
fo file with mathml that causes the no layout maker error

I patched the fop script as per the last post, but when processing a document
with math ml, I recieved the following error.

xsltproc --output
../live/chris.chiasson.name/Engineering_Optimization/Engineering_Optimization.fo
--xinclude /home/chris/sites/source/xsl/alpha/fo.xsl
../live/chris.chiasson.name/Engineering_Optimization/Engineering_Optimization.chunkbook
Making portrait pages on USletter paper (8.5inx11in)
/usr/bin/fop
../live/chris.chiasson.name/Engineering_Optimization/Engineering_Optimization.fo
-pdf
../live/chris.chiasson.name/Engineering_Optimization/Engineering_Optimization.pdf
Couldn't find hyphenation pattern en
No LayoutManager maker for class class org.apache.fop.mathml.MathMLElement

I figure the output of this command might be usefull (I didn't include
batik-1.6 because it is so verbose).

java-config -p 'commons-io-1,commons-logging,avalon-framework-4.2,jeuclid,fop' 
   
/usr/share/commons-io-1/lib/commons-io.jar:/usr/share/commons-logging/lib/commons-logging.jar:/usr/share/commons-logging/lib/commons-logging-api.jar:/usr/share/avalon-framework-4.2/lib/avalon-framework.jar:/usr/share/jeuclid/lib/jeuclid.jar:/usr/share/fop/lib/mathml-fop.jar:/usr/share/fop/lib/fop.jar

------- Comment #21 From bware 2006-02-07 04:09:28 0000 -------
I am afraid I am not a expert on the new version of FOP, it is a BETA release
and you may have more luck asking on the FOP mailing lists.

Can you render the MathML example:

fop /usr/share/doc/fop-0.91_beta/examples/examples/mathml/docs/mathml.fo
test.pdf

I get the external MathML shown correctly, but the inline examples are skiped.

As far as I can tell this isn't gentoo related, does anyone now if inline
mathml is ment to work in this release of FOP?

------- Comment #22 From Jiri Tyr 2006-04-07 10:34:38 0000 -------
Hi,

I have tryed your ebuild and I have found this problems:

1) If are some errors in FOP compilation, but the compilation is on the end
SUCCESFULL, then I recomended to remove '|| die "complie faild"' in the ebuild.

2) If you do not add mathml USE flag, then in /usr/bin/fop is jeuclid and FOP
report error on the begin of running.

3) I would like to use my owen userconfig.xml, so what if you add USE flag
userconfig and insert the default config from fop.jar for example to the
/etc/fop/conf and add it into /usr/bin/fop like default config by parameter -c?

Bwt. it is good job this your ebuild! Thank you!

Jiri

------- Comment #23 From Jiri Tyr 2006-05-02 07:36:19 0000 -------
The fop-0.92 was released. Can you make new ebuild?

I have also problems with MathML. It doesn't works at all. It give me only this
message:

fop -fo /usr/share/doc/fop-0.91_beta/examples/examples/mathml/docs/mathml.fo
./mathml.pdf
Base URL set to: file:///tmp
Source resolution set to: 72.0dpi (px2mm=0.35277778)
Target resolution set to: 72.0dpi (px2mm=0.35277778)
Default page-height set to: 11in
Default page-width set to: 8.26in
Warning element mpadded is not implemented
Warning element mpadded is not implemented
Some XML content will be ignored. No handler defined for XML:
http://www.w3.org/1998/Math/MathML

------- Comment #24 From Almad 2006-05-13 05:44:12 0000 -------
With 0.92_beta out, 0.91 has been deleted from mirrors :(

I can't build 0.92 (havent tried 91) as startscript.patch fail; if proceed
without patching, build fails with     [javac]
/mnt/pordisk/portage/tmp/portage/fop-0.92_beta/work/fop-0.92beta/src/java/org/apache/fop/image/XmlGraphicsCommonsImage.java:73:
cannot resolve symbol
    [javac] symbol  : class MemoryCacheSeekableStream 
    [javac] location: class org.apache.fop.image.XmlGraphicsCommonsImage
    [javac]                 seekableInput = new
MemoryCacheSeekableStream(inputStream);
    [javac]                                     ^
    [javac]
/mnt/pordisk/portage/tmp/portage/fop-0.92_beta/work/fop-0.92beta/src/java/org/apache/fop/layoutmgr/table/TableCellLayoutManager.java:340:
warning:
addBorders(org.apache.fop.area.Area,org.apache.fop.fo.properties.CommonBorderPaddingBackground,org.apache.fop.datatypes.PercentBaseContext)
in org.apache.fop.layoutmgr.TraitSetter has been deprecated
    [javac]                 TraitSetter.addBorders(curBlockArea, 
    [javac]                            ^
    [javac] 100 errors
    [javac] 9 warnings


(few hundred lines stripped)

Pretty please, could someone more skilled provide a patch? 

Thanks...

------- Comment #25 From bware 2006-05-13 07:49:11 0000 -------
Created an attachment (id=86704) [edit]
Updated ebuild for .92 Beta

*** Warning I am now using sun-jdk-1.5.0.06 on an amd64 and have not had time
to try these on anything else ***

Currently I can't get Baltik1.6 building under Java 1.5 and the is no email for
the xmlgraphics commons so I have let FOP use the copies included in it's
distribution (this works but is not the Gentoo way)

------- Comment #26 From bware 2006-05-13 07:50:21 0000 -------
Created an attachment (id=86705) [edit]
Patch the start script so fop can find it's dependencies

------- Comment #27 From Jiri Tyr 2006-05-18 08:15:44 0000 -------
(In reply to comment #25)
> *** Warning I am now using sun-jdk-1.5.0.06 on an amd64 and have not had time
> to try these on anything else ***

It works good for me. I have blackdown-jdk-1.4.2.03 and pentium-m. Thank for
the ebuild. Btw. ebuild for the FOP trunk version is not possible to make? It
is much better then this 0.92 beta version.

------- Comment #28 From Jiri Tyr 2006-05-19 09:17:30 0000 -------
Created an attachment (id=87061) [edit]
dev-java/fop/fop-1.0_beta.ebuild

Here is ebuild for fop_trunk from SVN. I have added also xmlunit support.
Enjoy...

------- Comment #29 From Jiri Tyr 2006-05-19 09:18:40 0000 -------
Created an attachment (id=87062) [edit]
files/1.0_beta-startscript.patch

Patch for the fop start script. It's the same like for fop-0.92_beta.

------- Comment #30 From Petteri Räty 2007-01-08 20:58:31 0000 -------
FOP-0.93 that is considered production quality by upstream, has been released
so we should start thinking about putting this to the main tree.

------- Comment #31 From Jiri Tyr 2007-01-09 11:00:08 0000 -------
Created an attachment (id=106183) [edit]
fop-0.93_beta.ebuild

Ebuild for FOP 0.93.

------- Comment #32 From Michael Kefeder 2007-01-20 20:41:31 0000 -------
Created an attachment (id=107587) [edit]
fop-0.93.ebuild and startscript patch

fixed existing fop-0.93_beta.ebuild (0.93 is a svn tag not a branch) and added
and adopted startscript patch from earlier efforts of this thread, else you'll
get the 'Exception in thread "main" java.lang.NoClassDefFoundError:
org/apache/fop/cli/Main' errormessage.
fop 0.93 runs fine, is considered stable (and I use it in production) - 
therefore I agree with Petteri Räty that it should be added to the main tree.

------- Comment #33 From Jiri Tyr 2007-01-21 18:56:51 0000 -------
(In reply to comment #32)

Maybe you should upload the ebuild and the startup script separately and not in
tar.gz file. Otherwise thanks for thous files.

PS: Are you missing the MathML implementation in FOP? I missing that very much!
I would appreciate to somebody implement the Jeuclid into the FOP ;o)

------- Comment #34 From Chris Chiasson 2007-01-22 05:06:54 0000 -------
(In reply to comment #33)
> (In reply to comment #32)
> 
> Maybe you should upload the ebuild and the startup script separately and not in
> tar.gz file. Otherwise thanks for thous files.
> 
> PS: Are you missing the MathML implementation in FOP? I missing that very much!
> I would appreciate to somebody implement the Jeuclid into the FOP ;o)
> 

SVGMath is pretty good at transforming presentation MathML into SVG. It accepts
a .fo file containing embedded MathML and outputs a .fo file with embedded SVGs
and proper baseline shifts for those SVGs. Does Jeuclid do that (I've never
tried it)?

------- Comment #35 From Jiri Tyr 2007-01-22 14:11:47 0000 -------
Created an attachment (id=107788) [edit]
fop-0.93_beta.ebuild

Correction of RDEPEND for dev-java/commons-io.

------- Comment #36 From Jiri Tyr 2007-01-22 14:13:48 0000 -------
Created an attachment (id=107790) [edit]
0.93_beta-startscript.patch

Plain text version of the start script.

------- Comment #37 From Jiri Tyr 2007-01-22 14:33:56 0000 -------
(In reply to comment #34)
> SVGMath is pretty good at transforming presentation MathML into SVG. It accepts
> a .fo file containing embedded MathML and outputs a .fo file with embedded SVGs
> and proper baseline shifts for those SVGs. Does Jeuclid do that (I've never
> tried it)?
> 

It doesn't works very well, but it could be better then nothing ;o) I just
tried the MathML to SVG and GIF. The quality is not the best, but it is better
then nothing. I don't know if there is implemented the baseline shifting, you
can try it ;o)

------- Comment #38 From Petteri Räty 2007-01-22 15:02:24 0000 -------
(In reply to comment #36)
> Created an attachment (id=107790) [edit]
> 0.93_beta-startscript.patch
> 
> Plain text version of the start script.
> 

Startup scripts should be done using java-pkg_dolauncher. Check the eclass
(java-utils-2.eclass) for documentation.

------- Comment #39 From Jiri Tyr 2007-01-22 16:16:50 0000 -------
(In reply to comment #32)
> fixed existing fop-0.93_beta.ebuild (0.93 is a svn tag not a branch) 

I just tried to compile fop 0.93 with the new SVN path and it doesn't work for
me:

>>> Unpacking source...

!!! ERROR: dev-java/fop-0.93_beta failed.
Call stack:
  ebuild.sh, line 1546:   Called dyn_unpack
  ebuild.sh, line 708:   Called src_unpack
  fop-0.93_beta.ebuild, line 36:   Called subversion_src_unpack
  subversion.eclass, line 274:   Called subversion_fetch
  subversion.eclass, line 183:   Called die

!!! subversion.eclass: ESVN_REPO_URI (or specified URI) and working copy's URL
are not matched.
!!! If you need support, post the topmost build error, and the call stack if
relevant.

!!! When you file a bug report, please include the following information:
GENTOO_VM=sun-jdk-1.5  CLASSPATH="" JAVA_HOME="/opt/sun-jdk-1.5.0.10"
JAVACFLAGS="-source 1.5 -target 1.5" COMPILER=""
and of course, the output of emerge --info
!!! This ebuild is from an overlay: '/usr/local/portage'

------- Comment #40 From Jiri Tyr 2007-01-22 16:20:23 0000 -------
(In reply to comment #32)
> fixed existing fop-0.93_beta.ebuild 

Is it really beta? At this page
(http://svn.apache.org/repos/asf/xmlgraphics/fop/tags/) it is not marked like
beta.

------- Comment #41 From Michael Kefeder 2007-01-26 20:51:18 0000 -------
(In reply to comment #40)
> (In reply to comment #32)
> > fixed existing fop-0.93_beta.ebuild 
> 
> Is it really beta? At this page
> (http://svn.apache.org/repos/asf/xmlgraphics/fop/tags/) it is not marked like
> beta.
> 

I did notice that, I mentioned it in my post that fop is considered stable now.
Just wanted to get the new fop running that's why i didn't rename the ebuild.
Feel free to do so.

Sorry for posting a tar.gz btw. About the mathml support - just install jeuclid
and add its .jar to the classpath of the fop startscript, then it should work.
Never used MathML, I always exported from TeX so far ;)

------- Comment #42 From Alistair Bush 2007-01-27 06:49:57 0000 -------
Im going to to preempt this.  but I am planning (and currently working on)
getting this ebuild into the migrated-java-experimental overlay.

Ill also reiterate that fob-0.93 has been released and this ebuild will not
pull the source from subversion.
I'll post the url once its up.

Thanks for the work your done so far

------- Comment #43 From Alistair Bush 2007-01-29 00:24:40 0000 -------
Just to update, a WIP ebuild is in the overlay here

https://overlays.gentoo.org/svn/proj/java/migrated-java-experimental-overlay/dev-java/fop/

the following TODO's remain

1) cleanup
2) all the optional dependencies
3) the launcher script
4) get JAVADOC's to generate :(

------- Comment #44 From Matt Kynx 2007-02-28 14:17:24 0000 -------
Nice :)

Only prob I had was that the /usr/bin/fop startup script doesn't use slots, so
was looking at my old FOP install. Changing the LOCALCLASSPATH line to:

LOCALCLASSPATH="`java-config -d --classpath=fop-0.93`"

And it plays nice.

------- Comment #45 From Miroslav Šulc 2007-03-04 14:42:22 0000 -------
(In reply to comment #44)
> Nice :)
> 
> Only prob I had was that the /usr/bin/fop startup script doesn't use slots, so
> was looking at my old FOP install. Changing the LOCALCLASSPATH line to:
> 
> LOCALCLASSPATH="`java-config -d --classpath=fop-0.93`"
> 
> And it plays nice.
> 

Thank you, I corrected that in overlay.

------- Comment #46 From Vlastimil Babka (Caster) 2007-05-07 10:24:54 0000 -------
Commited to main tree in the same slot, not new. So the launcher is fop.
There's no MathML support (we don't have jeuclid) and no hyphenation because
there are no hyphenation xml files in the sources, weird. Launcher is using
dolauncher without extra features that the upstream launcher had, they didn't
seem worth. If anything of this, or some other problem makes you unhappy,
please open new bug, this one is long enough :)

------- Comment #47 From Jiri Tyr 2007-05-07 12:02:27 0000 -------
(In reply to comment #46)
> There's no MathML support (we don't have jeuclid)
There exists ebuild for jeuclid
(http://bugs.gentoo.org/show_bug.cgi?id=121822).
> no hyphenation because there are no hyphenation xml files in the sources
Hyphenation could be included from this page:
http://offo.sourceforge.net/hyphenation/

------- Comment #48 From Jiri Tyr 2007-05-07 12:07:20 0000 -------
Created an attachment (id=118438) [edit]
fop-1.0_beta.ebuild

Updated ebuild for SVN version of FOP. I use it in production and it works very
good for me. May be you could add it into main portage tree as a fop-1.0_beta
or fop-1.0_alpha and mask it hardly (~M).

------- Comment #49 From Vlastimil Babka (Caster) 2007-05-07 12:38:22 0000 -------
This bug should really be left alone finally, too many comments and attachments
to scroll :) new bugs please, I commited this fop mainly for freemind which
took enough pain on its own to be concerned also with jeuclid and stuff. Not
trying to be bureaucratic, it just helps manageability.

(In reply to comment #47)
> (In reply to comment #46)
> > There's no MathML support (we don't have jeuclid)
> There exists ebuild for jeuclid
> (http://bugs.gentoo.org/show_bug.cgi?id=121822).

open a bug blocking 121822 then if you want mathml

> > no hyphenation because there are no hyphenation xml files in the sources
> Hyphenation could be included from this page:
> http://offo.sourceforge.net/hyphenation/

another bug, preferably with patch and USEflag (extra dep) if you want that,
seems your 1.0beta doesn't even have it?

(In reply to comment #48)
> Created an attachment (id=118438) [edit]
> fop-1.0_beta.ebuild
> 
> Updated ebuild for SVN version of FOP. I use it in production and it works very
> good for me. May be you could add it into main portage tree as a fop-1.0_beta
> or fop-1.0_alpha and mask it hardly (~M).

Hmm live svn ebuild for trunk, definitely should have different naming. Another
bug please :) should be useful to more than one person to be in tree.

------- Comment #50 From Jiri Tyr 2007-10-02 12:36:20 0000 -------
Created an attachment (id=132386) [edit]
fop-0.94.ebuild

Ebuild for dev-java/fop-0.94. Included MathML support with jeuclid-fop plugin.

------- Comment #51 From Jiri Tyr 2007-10-04 09:56:50 0000 -------
Created an attachment (id=132532) [edit]
fop-1.0_beta.ebuild

Ebuild for dev-java/fop-1.0_beta with jeuclid-fop-3.0.1 plug-in support for
MathML. Use this version if fop-0.94 gives you this error:

Exception in thread "main" java.lang.NoClassDefFoundError:
org/apache/xml/utils/URI$MalformedURIException

reported here:
http://issues.apache.org/bugzilla/show_bug.cgi?id=43542

which is already corrected in trunk version:
http://svn.apache.org/viewvc?rev=581806&view=rev

First Last Prev Next    No search results available      Search page      Enter new bug