DBus font installs [incomplete]

[Dr Andrew John Hughes]

On 15:32 Thu 14 Apr     , Deepak Bhole wrote:
> Please re-post this patch with proper docs for all of the functions and
> comments on what the non-obvious lines of code are doing, especially
> when they call external API.
> 

Thanks for pinging this Deepak.  Got lost in my mail box until now.

> Comments on generic issues below:
> 
> 
> * Andrew Su <asu at redhat.com> [2011-03-02 13:35]:
> > Hello,
> > 
> >   This patch is an attempt at requesting users to install a font that will allow them to display the text being rendered if the glyph is missing. It is still a work in progress. 
> > 
> > Known issues:
> >  - Segmentation fault when rechecking for newly installed fonts to load.
> 
> This is a showstopper. Nothing we add should make the JVM crash under
> any circumstances.

Agreed.  But I don't think this means the code shouldn't go in so it can be worked on, as long as it is disabled
by default.  Having it available to all makes it easier to fix.

I did ask Andrew to post his work so it wasn't lingering on his desk.  I don't expect it to be perfect.

> 
> >  - If characters from different languages are being rendered one at a time it would constantly request user to install the languages not found.
> 
> Can't we maintain a list and not ask back if the user said no once?
> 
> >  - May ask to install a non-standard language (unrecognised by packagekit).
> > 
> 
> What will packagekit try to install if it is non-standard?
> 
> > Attached: 
> >   - addFontInstall.patch: This will add the patch and modify the Makefile accordingly.
> >   - GlyphChecker.c: Used to create libglyphchecker.so (which should be placed in the openjdk's lib directory.
> >   - GlyphChecker.h: Header for using jni.
> > 
> > I used [on a x86_64 machine]
> > 
> > gcc -I/usr/lib/jvm/java-1.6.0-openjdk-1.6.0.0.x86_64/include/ -I/usr/lib/jvm/java-1.6.0-openjdk-1.6.0.0.x86_64/include/linux/ -I/usr/include/dbus-1.0/ -I/usr/include/glib-2.0/ -I/usr/lib64/glib-2.0/include -shared -fPIC -g -O0 -Wall -lfontconfig -ldbus-glib-1 -o libglyphchecker.so GlyphChecker.c
> > 
> > to build the library file. (Replace with where your libs are..) once compiled, place the .so into your jre's lib dir.
> > ex: ${BUILD_DIR}/openjdk.build/j2sdk-image/jre/lib/amd64/
> > 
> > To test that it searches and installs fonts..
> > You can run it as follows:
> >  GlyphChecker.addMissing((char) 0x710);
> >  GlyphChecker.doAll(); // This should return true if you accepted to install
> > 

This needs integrating into the Makefile before it can go in.

> > Please let me know if you see any issues with it.
> > 
> 
> ...
> > #include <fontconfig/fontconfig.h>
> > #include <stdlib.h>
> > #include <dlfcn.h>
> > 
> > FcChar8* getSuperSet(FcLangSet* fls) {
> > 	//	printf("Getting superset.\n");
> >
> 
> Above should be removed. As should all other uncommented printfs

I'd prefer we had something like #ifdef DEBUG round them as opposed to removing all of them altogether.

snip...

> 
> > +         } else {
> > +diff -r 9d90acd9f99f src/solaris/native/sun/awt/fontpath.c
> > +--- openjdk/jdk/src/solaris/native/sun/awt/fontpath.c	Wed Feb 16 09:39:35 2011 -0500
> > ++++ openjdk/jdk/src/solaris/native/sun/awt/fontpath.c	Wed Mar 02 09:24:59 2011 -0500
> > +@@ -517,11 +517,11 @@
> > + }
> > + 
> > + JNIEXPORT jstring JNICALL Java_sun_font_FontManager_getFontPath
> > +-(JNIEnv *env, jclass obj, jboolean noType1) {
> > ++(JNIEnv *env, jclass obj, jboolean noType1, jboolean recheck) {
> 
> We should find some other way to recheck (whatever it is doing).
> Changing a public api, even for a sun.* class is a no-no.
> 

For sun.*, it's possible.  I'd prefer it to some convoluted workaround.

> Deepak

-- 
Andrew :)

Free Java Software Engineer
Red Hat, Inc. (http://www.redhat.com)

Support Free Java!
Contribute to GNU Classpath and IcedTea
http://www.gnu.org/software/classpath
http://icedtea.classpath.org
PGP Key: F5862A37 (https://keys.indymedia.org/)
Fingerprint = EA30 D855 D50F 90CD F54D  0698 0713 C3ED F586 2A37