You are viewing a plain text version of this content. The canonical link for it is here.
Posted to derby-dev@db.apache.org by Kathey Marsden <km...@sbcglobal.net> on 2008/02/04 19:39:13 UTC
Bug fixes in trunk that are not in 10.3 branch
I used a slightly modified version of Rick's Jira VTI demo (Thanks
Rick!) to take a report of the bugs affecting 10.3 that are fixed in the
trunk but not 10.3 and thought it might be of general interest to the
community. In general I think we have done a good job of backporting
fixes so users can benefit right away.
Kathey
DERBY-3366 |[DERBY-3366] Various formatting erros in L10N property
files
DERBY-3350 |[DERBY-3350] SQL CAST always marks its type as nullable
even if the expression to be cast is not nullable
DERBY-3308 |[DERBY-3308] Broken synchronization for event handling in
ClientPooledConnection40
DERBY-3298 |[DERBY-3298] Getting Started manual needs clearer
introduction to connection
URL
DERBY-3278 |[DERBY-3278] Developer's Guide topic cdevdvlp51654 has
duplicate information on connection URL
DERBY-3274 |[DERBY-3274] Developer's Guide has duplicate information on
database connections
DERBY-3262 |[DERBY-3262] Documentation is missing cross-references for
connection URL attributes
DERBY-3226 |[DERBY-3226] BlobLocatorInputStream.read() and
ClobLocatorInputStream.read() don't mask out sign
bits
DERBY-3215 |[DERBY-3215] Potential NullPointerException in CachedPage
class
DERBY-3168 |[DERBY-3168] Reference Manual lacks topics on trace-related
connection URL attributes
DERBY-3160 |[DERBY-3160] SYSCS_GET_USER_ACCESS incorrectly treats the
passed in user name as a SQL identifier and thus can reports the wron&
DERBY-3084 |[DERBY-3084] CREATE SCHEMA in refman does not contain
material on restrictions in sqlAuthorization mode
DERBY-3079 |[DERBY-3079] Database name is printed twice in derby.log on
rollbacks when logStatementText is enabled
DERBY-3060 |[DERBY-3060] Network Server incorrectly assumes that all
SQLExceptions with error code 08004 are caused by an authentication fa&
DERBY-2948 |[DERBY-2948] Reoccurrence of DERBY-2842:
derbynet/testconnection.java fails with '...Connection refused: connect'
expecting 'Co&
DERBY-2935 |[DERBY-2935] DDMReader.readLengthAndCodePoint() decodes
long integer incorrectly
DERBY-2815 |[DERBY-2815] ij doesn't start with J2ME / JSR169 / weme6.1
because attempting to find java.sql.Driver if ij.protocol property i&
DERBY-2733 |[DERBY-2733] ij rolls through NullPointerException (NPE)
with J2ME/JSR169/WEME 6.1.
DERBY-2270 |[DERBY-2270] '18 ' (18 followed by four blanks) needs three
more blanks
DERBY-1823 |[DERBY-1823] Derby Developer's Guide - Issues w/ User
authentication and authorization extended examples section/paragraph
building demo code, was: Bug fixes in trunk that are not in 10.3 branch
Posted by Rick Hillegas <Ri...@Sun.COM>.
Dyre.Tjeldvoll@Sun.COM wrote:
> Kathey Marsden <km...@sbcglobal.net> writes:
>
>
>> I had to build the demo with javac -g -d $WS/classes. It did not seem
>> to build automatically as part of the build. I don't know if that is a
>> problem.
>>
>
> I think there is a separate ant target for it, vti-demo that you can
> run. But as you say, that target is not run automatically
>
>
>
I think that this target can be wired into the general build now. This
target was removed from the ordinary build because the demo code
requires Java 5. Since then we decided to require Java 5 so I believe
that the impediment has been removed. I haven't gotten around to
changing the build. Feel free to!
However, even if we make this change, the demo classes will be built
into the classtree but not into any jar files. Part of me thinks that
some of these table functions (e.g., the PropertyFileVTI) are generic
and mature enough that they could migrate into the tools subtree.
Regards,
-Rick
Re: Bug fixes in trunk that are not in 10.3 branch
Posted by Dy...@Sun.COM.
Kathey Marsden <km...@sbcglobal.net> writes:
>
> I had to build the demo with javac -g -d $WS/classes. It did not seem
> to build automatically as part of the build. I don't know if that is a
> problem.
I think there is a separate ant target for it, vti-demo that you can
run. But as you say, that target is not run automatically
--
dt
Re: Bug fixes in trunk that are not in 10.3 branch
Posted by Kim Haase <Ca...@Sun.COM>.
I notice that at least for doc bugs, there have been some cases that
this report hasn't caught. Three bugs that I know of that were fixed for
10.4 have empty version fields, and two of these have empty fixversion
fields too:
DERBY-2240 (no version)
DERBY-3072, DERBY-3297 (no version or fixversion)
Two of these are closed, one is resolved. Would it make sense to fill in
these fields? 3072 was fixed by Rick Hillegas, the other two by me.
There may be other such bugs; I don't know.
Thanks,
Kim
Kathey Marsden wrote:
> Myrna asked me how I generated this report.
> 1) I patched the demo to include a few extra columns.
> Index: java/demo/vtis/java/org/apache/derbyDemo/vtis/example/VTIs.java
>
> ===================================================================
>
> --- java/demo/vtis/java/org/apache/derbyDemo/vtis/example/VTIs.java
> (revision 618330)
>
> +++ java/demo/vtis/java/org/apache/derbyDemo/vtis/example/VTIs.java
> (working copy)
>
> @@ -94,8 +94,8 @@
>
> @XMLRow
> (
> rowTag = "item",
> - childTags = { "key", "type", "priority", "status",
> "component", "title" },
> - childTypes = { "varchar(12)", "varchar(10)", "varchar(10)",
> "varchar(10)", "varchar(50)", "varchar(200)" },
> + childTags = { "key", "type", "assignee", "priority", "status",
> "component", "version","fixVersion", "title","resolution" },
> + childTypes = { "varchar(12)", "varchar(10)", "varchar(50)",
> "varchar(10)", "varchar(10)", "varchar(50)", "varchar(150)",
> "varchar(150)", "varchar(200)","varchar(20)" },
> vtiClassName = "org.apache.derbyDemo.vtis.core.XmlVTI"
> )
> public static ResultSet apacheVanillaJiraReport( String
> xmlResource ) throws SQLException
>
> I had to build the demo with javac -g -d $WS/classes. It did not seem to
> build automatically as part of the build. I don't know if that is a
> problem.
>
> 2) I ran the part of the demoFileVtis.sql file related to the JiraVTI
>
> connect 'jdbc:derby:vtitest;create=true';
>
> ----------------------------------------------------------------------------------------
>
> --
> -- Drop and recreate the database procedures and tables needed
> -- by this demonstration script.
> --
> ----------------------------------------------------------------------------------------
>
>
> --
> -- Drop procedures and tables
> --
> drop procedure registerXMLRowVTIs;
>
> --
> -- Drop miscellaneous table functions
> --
> drop function svnLogReader;
> drop function propertyFileVTI;
>
>
> --
> -- Recreate procedures
> --
> create procedure registerXMLRowVTIs( className varchar( 32672 ) )
> language java
> parameter style java
> modifies sql data
> external name 'org.apache.derbyDemo.vtis.core.XmlVTI.registerXMLRowVTIs'
> ;
>
> ----------------------------------------------------------------------------------------
>
> --
> -- Declare the table functions.
> --
> ----------------------------------------------------------------------------------------
>
>
> --
> -- Register the table functions in the VTIs class
> --
> call registerXMLRowVTIs( 'org.apache.derbyDemo.vtis.example.VTIs' );
>
> --
>
>
> 3) I ran the report.
>
> select s."key", s."title", s."assignee"
> from table( "apacheVanillaJiraReport"(
> 'file:///kmarsden/projects/tablefunction/all_JIRA_ISSUES.xml' ) ) s
> where "type" = 'Bug' AND "version" LIKE '%10.3%' AND "fixVersion" LIKE
> '%10.4%' AND NOT "fixVersion" LIKE '%10.3%' AND "resolution" = 'Fixed';
>
>
> The version and the fixversion values are kind of weird with all the
> versions smashed together, but the worked ok for my purposes. I don't
> know whether it would be appropriate to add them to the demo though in
> that format.
>
> Kathey
>
Re: Bug fixes in trunk that are not in 10.3 branch
Posted by Kathey Marsden <km...@sbcglobal.net>.
Myrna asked me how I generated this report.
1) I patched the demo to include a few extra columns.
Index: java/demo/vtis/java/org/apache/derbyDemo/vtis/example/VTIs.java
===================================================================
--- java/demo/vtis/java/org/apache/derbyDemo/vtis/example/VTIs.java
(revision 618330)
+++ java/demo/vtis/java/org/apache/derbyDemo/vtis/example/VTIs.java
(working copy)
@@ -94,8 +94,8 @@
@XMLRow
(
rowTag = "item",
- childTags = { "key", "type", "priority", "status",
"component", "title" },
- childTypes = { "varchar(12)", "varchar(10)", "varchar(10)",
"varchar(10)", "varchar(50)", "varchar(200)" },
+ childTags = { "key", "type", "assignee", "priority", "status",
"component", "version","fixVersion", "title","resolution" },
+ childTypes = { "varchar(12)", "varchar(10)", "varchar(50)",
"varchar(10)", "varchar(10)", "varchar(50)", "varchar(150)",
"varchar(150)", "varchar(200)","varchar(20)" },
vtiClassName = "org.apache.derbyDemo.vtis.core.XmlVTI"
)
public static ResultSet apacheVanillaJiraReport( String
xmlResource ) throws SQLException
I had to build the demo with javac -g -d $WS/classes. It did not seem to
build automatically as part of the build. I don't know if that is a problem.
2) I ran the part of the demoFileVtis.sql file related to the JiraVTI
connect 'jdbc:derby:vtitest;create=true';
----------------------------------------------------------------------------------------
--
-- Drop and recreate the database procedures and tables needed
-- by this demonstration script.
--
----------------------------------------------------------------------------------------
--
-- Drop procedures and tables
--
drop procedure registerXMLRowVTIs;
--
-- Drop miscellaneous table functions
--
drop function svnLogReader;
drop function propertyFileVTI;
--
-- Recreate procedures
--
create procedure registerXMLRowVTIs( className varchar( 32672 ) )
language java
parameter style java
modifies sql data
external name 'org.apache.derbyDemo.vtis.core.XmlVTI.registerXMLRowVTIs'
;
----------------------------------------------------------------------------------------
--
-- Declare the table functions.
--
----------------------------------------------------------------------------------------
--
-- Register the table functions in the VTIs class
--
call registerXMLRowVTIs( 'org.apache.derbyDemo.vtis.example.VTIs' );
--
3) I ran the report.
select s."key", s."title", s."assignee"
from table( "apacheVanillaJiraReport"(
'file:///kmarsden/projects/tablefunction/all_JIRA_ISSUES.xml' ) ) s
where "type" = 'Bug' AND "version" LIKE '%10.3%' AND "fixVersion" LIKE
'%10.4%' AND NOT "fixVersion" LIKE '%10.3%' AND "resolution" = 'Fixed';
The version and the fixversion values are kind of weird with all the
versions smashed together, but the worked ok for my purposes. I don't
know whether it would be appropriate to add them to the demo though in
that format.
Kathey
Re: Bug fixes in trunk that are not in 10.3 branch
Posted by Kathey Marsden <km...@sbcglobal.net>.
Thanks Kristian for the reformatted report.
Kristian Waagan wrote:
>
> I see that I worked on two of the issues (3308 and 3079), but I didn't
> backport them because I wasn't quite sure about which guidelines to
> follow.
>
I think if a change will likely benefit users and offers low risk of
regression or impact to existing users it is a good candidate for
backport. DERBY-3308 sounds like it might likely hit in a production
scenario and be hard to diagnose, so if you consider it low risk, I
think that would be a good candidate for backport. I doubt DERBY-3079
on the other hand is going to cause anyone too much trouble, so may not
be worth the backport effort. In general it is a personal judgement call.
http://wiki.apache.org/db-derby/MoveAChangeBetweenBranches
offers some guidance about backporting changes. You can also always ask
on the list if you are want additional opinions.
Kathey
> DERBY-3366: Various formatting erros in L10N property files
> DERBY-3350: SQL CAST always marks its type as nullable even if the
> expression to be cast is not nullable
> DERBY-3308: Broken synchronization for event handling in
> ClientPooledConnection40
> DERBY-3298: Getting Started manual needs clearer introduction to
> connection URL
> DERBY-3278: Developer's Guide topic cdevdvlp51654 has duplicate
> information on connection URL
> DERBY-3274: Developer's Guide has duplicate information on database
> connections
> DERBY-3262: Documentation is missing cross-references for connection
> URL attributes
> DERBY-3226: BlobLocatorInputStream.read() and
> ClobLocatorInputStream.read() don't mask out sign bits
> DERBY-3215: Potential NullPointerException in CachedPage class
> DERBY-3168: Reference Manual lacks topics on trace-related connection
> URL attributes
> DERBY-3160: SYSCS_GET_USER_ACCESS incorrectly treats the passed in
> user name as a SQL identifier and thus can reports the wron&
> DERBY-3084: CREATE SCHEMA in refman does not contain material on
> restrictions in sqlAuthorization mode
> DERBY-3079: Database name is printed twice in derby.log on rollbacks
> when logStatementText is enabled
> DERBY-3060: Network Server incorrectly assumes that all SQLExceptions
> with error code 08004 are caused by an authentication fa&
> DERBY-2948: Reoccurrence of DERBY-2842: derbynet/testconnection.java
> fails with '...Connection refused: connect' expecting 'Co&
> DERBY-2935: DDMReader.readLengthAndCodePoint() decodes long integer
> incorrectly
> DERBY-2815: ij doesn't start with J2ME / JSR169 / weme6.1 because
> attempting to find java.sql.Driver if ij.protocol property i&
> DERBY-2733: ij rolls through NullPointerException (NPE) with
> J2ME/JSR169/WEME 6.1.
> DERBY-2270: '18 ' (18 followed by four blanks) needs three more blanks
> DERBY-1823: Derby Developer's Guide - Issues w/ User authentication
> and authorization extended examples section/paragraph
>
>
> regards,
Re: Bug fixes in trunk that are not in 10.3 branch
Posted by Kristian Waagan <Kr...@Sun.COM>.
Kathey Marsden skrev:
> I used a slightly modified version of Rick's Jira VTI demo (Thanks
> Rick!) to take a report of the bugs affecting 10.3 that are fixed in
> the trunk but not 10.3 and thought it might be of general interest to
> the community. In general I think we have done a good job of
> backporting fixes so users can benefit right away.
Hi Kathey,
Thanks for looking into this. The listing was a bit garbled in my mail
client, so I took the liberty to reformat it for easier reading.
I see that I worked on two of the issues (3308 and 3079), but I didn't
backport them because I wasn't quite sure about which guidelines to follow.
DERBY-3366: Various formatting erros in L10N property files
DERBY-3350: SQL CAST always marks its type as nullable even if the
expression to be cast is not nullable
DERBY-3308: Broken synchronization for event handling in
ClientPooledConnection40
DERBY-3298: Getting Started manual needs clearer introduction to
connection URL
DERBY-3278: Developer's Guide topic cdevdvlp51654 has duplicate
information on connection URL
DERBY-3274: Developer's Guide has duplicate information on database
connections
DERBY-3262: Documentation is missing cross-references for connection URL
attributes
DERBY-3226: BlobLocatorInputStream.read() and
ClobLocatorInputStream.read() don't mask out sign bits
DERBY-3215: Potential NullPointerException in CachedPage class
DERBY-3168: Reference Manual lacks topics on trace-related connection
URL attributes
DERBY-3160: SYSCS_GET_USER_ACCESS incorrectly treats the passed in user
name as a SQL identifier and thus can reports the wron&
DERBY-3084: CREATE SCHEMA in refman does not contain material on
restrictions in sqlAuthorization mode
DERBY-3079: Database name is printed twice in derby.log on rollbacks
when logStatementText is enabled
DERBY-3060: Network Server incorrectly assumes that all SQLExceptions
with error code 08004 are caused by an authentication fa&
DERBY-2948: Reoccurrence of DERBY-2842: derbynet/testconnection.java
fails with '...Connection refused: connect' expecting 'Co&
DERBY-2935: DDMReader.readLengthAndCodePoint() decodes long integer
incorrectly
DERBY-2815: ij doesn't start with J2ME / JSR169 / weme6.1 because
attempting to find java.sql.Driver if ij.protocol property i&
DERBY-2733: ij rolls through NullPointerException (NPE) with
J2ME/JSR169/WEME 6.1.
DERBY-2270: '18 ' (18 followed by four blanks) needs three more blanks
DERBY-1823: Derby Developer's Guide - Issues w/ User authentication and
authorization extended examples section/paragraph
regards,
--
Kristian
>
> Kathey
>
>
> DERBY-3366 |[DERBY-3366] Various formatting erros in L10N property
> files
> DERBY-3350 |[DERBY-3350] SQL CAST always marks its type as nullable
> even if the expression to be cast is not
> nullable DERBY-3308 |[DERBY-3308] Broken
> synchronization for event handling in
> ClientPooledConnection40
> DERBY-3298 |[DERBY-3298] Getting Started manual needs clearer
> introduction to connection
> URL DERBY-3278
> |[DERBY-3278] Developer's Guide topic cdevdvlp51654 has duplicate
> information on connection URL
> DERBY-3274 |[DERBY-3274] Developer's Guide has duplicate information
> on database connections
> DERBY-3262 |[DERBY-3262] Documentation is missing cross-references
> for connection URL
> attributes DERBY-3226
> |[DERBY-3226] BlobLocatorInputStream.read() and
> ClobLocatorInputStream.read() don't mask out sign
> bits DERBY-3215 |[DERBY-3215] Potential
> NullPointerException in CachedPage
> class
> DERBY-3168 |[DERBY-3168] Reference Manual lacks topics on
> trace-related connection URL
> attributes DERBY-3160
> |[DERBY-3160] SYSCS_GET_USER_ACCESS incorrectly treats the passed in
> user name as a SQL identifier and thus can reports the wron&
> DERBY-3084 |[DERBY-3084] CREATE SCHEMA in refman does not contain
> material on restrictions in sqlAuthorization
> mode DERBY-3079 |[DERBY-3079] Database name
> is printed twice in derby.log on rollbacks when logStatementText is
> enabled DERBY-3060 |[DERBY-3060] Network
> Server incorrectly assumes that all SQLExceptions with error code
> 08004 are caused by an authentication fa&
> DERBY-2948 |[DERBY-2948] Reoccurrence of DERBY-2842:
> derbynet/testconnection.java fails with '...Connection refused:
> connect' expecting 'Co&
> DERBY-2935 |[DERBY-2935] DDMReader.readLengthAndCodePoint() decodes
> long integer
> incorrectly DERBY-2815
> |[DERBY-2815] ij doesn't start with J2ME / JSR169 / weme6.1 because
> attempting to find java.sql.Driver if ij.protocol property i&
> DERBY-2733 |[DERBY-2733] ij rolls through NullPointerException (NPE)
> with J2ME/JSR169/WEME 6.1.
> DERBY-2270 |[DERBY-2270] '18 ' (18 followed by four blanks) needs
> three more
> blanks
> DERBY-1823 |[DERBY-1823] Derby Developer's Guide - Issues w/ User
> authentication and authorization extended examples section/paragraph