Some additional notes:
If a JIRA bug is still open, but a fix has been checked in, then the
developer needs to resolve the JIRA bug appropriately - otherwise, it
won't get tested. Look for individual emails from me regarding bugs in
this state
And - we have to be careful of svn commit information that refers to
other bugs. If you reference another bug number in an svn commit
comment, then that svn commit info is linked to that bug (under the
Subversion Commits tab) and it looks like there has been a checkin for
that bug, but there really hasn't been. For example,
http://www.openlaszlo.org/jira/browse/LPP-4506 should not have
Wafflecone set as the Fix field - because the checkin was regarding a
different bug and only the comment (that 4506 was a different bug)
really applied. The original state, 4.1, was correct.
I'll make a pass through all of these.
Thanks!
--- Amy
On Tue, Sep 11, 2007 at 8:33 AM, P T Withington wrote:
Ok, I think I see what your intent is with the 'p is null' -- there
is a work-around for Wafflecone but the bug still needs to be really
fixed in 4.1.
I think this all goes back to the problem with the 'Fix Version'
field being used to say 'we want a fix for this in this version' and
'we have a fix for this version', depending on the state of the bug.
If there were a way to split this field it would be really nice...
On 2007-09-11, at 08:19 EDT, P T Withington wrote:
Careful here: You seem to have marked a number of bugs that are
open for 4.1 as open for Wafflecone. These bugs were triaged out to
4.1 as they were seen as not critical and a workaround was made in
Wafflecone.
For instance, 4620 is a real fix for a bug that was papered over in
Wafflecone (I already reverted that and commented).
All the 'p is null' bugs are real bugs that need to be investigated,
the paper-over in Wafflecone is that we reduced the WARNING to an
INFO and made it possible to silence the INFO messages. So the bugs
should be targeted for 4.1, not Wafflecone.
I have not looked at this whole list to see if there are other bugs
that fall in this category.
On 2007-09-11, at 01:02 EDT, Benjamin Shine wrote:
Generating release notes is educational. It reveals places where
our several interlocking systems disagree. There are around 30 bugs
which have fixes included in wafflecone but were not marked as
fix-version wafflecone. They got to this state in several ways:
1) The fixes were checked into wafflecone but not marked in jira as
fix-version wafflecone.
2) The fixes were checked in to legals and then merged to
wafflecone. I should have marked the bugs as fix-version wafflecone
when I merged them.
Fixes checked into wafflecone but not marked in jira as fix- version
wafflecone:
4427 Clean-up for "basescrollbar"
1952 destroy before resource is loaded leads to full load queue and
hang
4394 "LzGlobalMouse.onmousemove" is way to slow for
"LzView.prototype.getMouse.dependencies" in SWF
4459 Visible property not respected for replication
4474 amazon soap fails in dhtml / wafflecone branch
4504 Calendar demo presents no sample data
4556 Rich text: Can't change font back to Verdana
4584 dataset fires multiple ondata-Events
4610 request' method is deprecated, causing trouble, needs to be
removed for LzDataset
4630 Dataset subclasses doRequest oninit if autorequest="true" and
request="false"
4660 Httpdatarequest has LzLoader value for xmldata when status is
"error"
4450 Add an option for loading raw string data, without system
attempting to parse XML
4464 setVisible(null) warning, but it is never called
4478 Richedittext: WARNING: __LZgetNodes: p is null in
LzDatapointer
4551 debugger warnings emitted from a private LFC function
4592 basedatacombobox-based components with selectfirst="false"
emits Debugger warnings
4401 Namespace separators are not recognized in datapaths - updated
4484 scrollbars do not play well with explicit replicators
4494 XPath `position()` operator incorrect in explicit replication
4506 lzcombobox doesn't set changed when selecting original value
4532 Two typos in "LzHTTPDatasource#doRequest(..)"
4553 basefocusview warnings when canvas size changes
4620 Convert validate attribute to #pragma
4269 No event data in calendar demo for August 2007
Fixes merged in from legals but not marked in jira as fix-version
wafflecone:
2111 Horrendously inefficient looping in layout classes!!
4357 getMouse returns absolute coordinates in DHTML + FF
4582 YouTube demo not working due to changes of Youtube.com FLV url
pattern
4120 LzTrack interferes w/ window drag w/ FF DHTML
I have just marked all of the above as fix-version wafflecone.
....
There are two bugs whose presence in the checkin log does not mean
that they were fixed in wafflecone:
4046 DHTML embedding should respect passed in values, instead of
the canvas values -- this was unblocked, then blocked, and does not
look like it ended up in wafflecone.
4451 - incorrectly mentioned in checkin for another bug!
....
These bugs are marked as retestbranch, and they are not picked up
by the jira release-note generator. I must make sure these get into
the wafflecone release notes as "fixed":
4347 baserichedittext-test.lzx fails (marked as retestbranch)
4348 rich-text: formatfontsize test fails
4352 scrollrichedittext can't create scrollbar
-ben
