Gregory Shimansky wrote:
Geir Magnusson Jr. wrote:
I found it - comment below :
Gregory Shimansky wrote:
I've created a patch to classlib which fixes the problem with drlvm.
Now if you could test it with IBM VME, I'll commit it:
Index: modules/luni/src/main/native/luni/shared/luniglob.c
===================================================================
--- modules/luni/src/main/native/luni/shared/luniglob.c (revision
[SNIP]
This puts a bandage on the problem - it seems like we changed the API
of GetSystemProperty() because this code clearly expected not to get
an error if the property wasn't set.
I've read Oliver's comment to r486100 and it didn't look like this code
shouldn't expect an error from GetSystemProperty:
======================================
Previously we just replaced whatever was already in the
org.apache.harmony.boot.class.path property with our own bootclasspath.
We should not assume that this property is empty before we use it - it
depends on the VM's bcp initialisation order and how it utilises this
property internally.
======================================
The "We should not assume" doesn't mean "it should always be this way"
in my understanding.
Right - the thing I was asking if the behavior of GetSystemProperty()
changed.
Looking at Ollie's change in r486100, he seems to think that it's not an
error to not have the property set, but our impl of it does.
That's the issue I'm trying to get to - does the API of
GetSystemProperty() specify that it returns an error code if the
property isn't set?
If so, then Ollie's code is wrong. If not, then it's either that the
API is ambiguous (Ollie's assumption wasn't unreasonable) or our impl of
GetSystemProperty() is wrong.
geir
