RouteConverter and Mac OS X

[asp]

Do the coordinates of the dragged position change in the position list if the flag is duplicated?

Yes. At least the length coordinate, there seems to be another small bug that the width coordinate is not updated in the list. (But it's updated when I change to the route and then back to the waypoints, so it is saved internally.)

The red flags also stay when I clear the routes etc. by File / New. Maybe this is related.

[hvdwolf]

Well done! Thank you very much!


Some quick tests:
O = OSX Snow Leopard 10.6 (64 bit), java 1.6.0_20
U = Ubuntu 10.04 32bit, java 1.6.0_18

- No error message (O and U).
- Inserting points works fine (O and U).
- switching from waypoints to route to waypoints and back works fine, even multi-point complex routes are drawn instantaneously (O and U).
- When dragging points the Red marker remains on the old spot (O occasionally, U often -> layer(s) not correctly updated?).
- When dragging points in Route mode the old route remains displayed (O + U occasionally-> layer(s) not correctly updated?).
- Opening new panes like file->open, Extras->Preferences takes minutes and sometimes make the application hang (or at it takes more than 3-6 minutes and I don't want to wait that long) (O only and always).

I will do further tests with better analysis tonight.

[routeconverter]

Yes. At least the length coordinate, there seems to be another small bug that the width coordinate is not updated in the list. (But it's updated when I change to the route and then back to the waypoints, so it is saved internally.)

I think I've fixed that width bug - but the app may be a bit slower and there might be other new bugs now... that Java Swing eventing stuff is hard to optimize and debug.

The red flags also stay when I clear the routes etc. by File / New. Maybe this is related.

Please test the prerelease I've just uploaded.

[routeconverter]

- When dragging points the Red marker remains on the old spot (O occasionally, U often -> layer(s) not correctly updated?).
- When dragging points in Route mode the old route remains displayed (O + U occasionally-> layer(s) not correctly updated?).

Thank you for testing, the update bugs seem to be the as those that asp reported. Maybe the latest prerelease I've just uploaded helps, too?

- Opening new panes like file->open, Extras->Preferences takes minutes and sometimes make the application hang (or at it takes more than 3-6 minutes and I don't want to wait that long) (O only and always).

Sounds like dozens of Exceptions might be printed on the console... or the log?

[hvdwolf]

OSX test only: The Red dots now work fine. Routes are also updated correctly after an "event" has happened. I see the same behaviour in google maps itself. Some things are immediately updated, some others a "little" later.

- Opening new panes like file->open, Extras->Preferences takes minutes and sometimes make the application hang (or at it takes more than 3-6 minutes and I don't want to wait that long) (O only and always).

Sounds like dozens of Exceptions might be printed on the console... or the log?

The only entry I get when trying to save a route and before it "hangs" is the following entry in the system log

Code:

/usr/bin/java[352]: GetDYLDEntryPointWithImage(/System/Library/Frameworks/AppKit.framework/Versions/Current/AppKit,_NSCreateAppKitServicesMenu) failed.

When trying to open a route, the file dialog appears (without files) and then it hangs.
Nothing is written to the terminal or the system log.

Does this have something to do with this thread, page 3, post 27 and on page 4, post 33 I mentioned the thing with

It starts with an error "Can't open <blah blah> due to unknow file type" (note:translated back to english from dutch) for the source. This is due to the fact that the application somehow tries to open a folder instead of a file.

Some users reported this to me and I'm confused... do you have an idea what can cause this error?

On OSX the file handle can be: "file", "folder", "file+folder". I assume the third is used but I don't know java. Only AppleScript and some basic C.

and then you fixed it

[routeconverter]

The only entry I get when trying to save a route and before it "hangs" is the following entry in the system log

Code:

/usr/bin/java[352]: GetDYLDEntryPointWithImage(/System/Library/Frameworks/AppKit.framework/Versions/Current/AppKit,_NSCreateAppKitServicesMenu) failed.

That sounds ugly... 9 Google hits: http://www.google.de/search?q=GetDYLDEnt...tWithImage

I completely clueless why your "File/Open..."-does-not-work fails. Any ideas?

[hvdwolf]

The only entry I get when trying to save a route and before it "hangs" is the following entry in the system log

Code:

/usr/bin/java[352]: GetDYLDEntryPointWithImage(/System/Library/Frameworks/AppKit.framework/Versions/Current/AppKit,_NSCreateAppKitServicesMenu) failed.

This means that "something" was compiled against some "AppKit.framework" using the function "_NSCreateAppKitServicesMenu". My "AppKit.framework" doesn't contain that function.
I did a:

Code:

nm /System/Library/Frameworks/AppKit.framework/Versions/Current/AppKit | grep -i NSCreateAppKitServices

and nothing was returned.

I finally did a:

Code:

nm /System/Library/Frameworks/AppKit.framework/Versions/Current/AppKit | grep -i NSCreate
000000000072010e T _NSCreateFileContentsPboardType
00000000000b9bc9 T _NSCreateFilenamePboardType
                 U _NSCreateHashTable
                 U _NSCreateMapTable
                 U _NSCreateMapTableWithZone
0000000000b94150 D _NSCreatesSortDescriptorBindingOption
00000000002b01ef T __NSCreateATSUStyleFromAttributes
0000000000428da2 T __NSCreateAttributesFromATSUStyle
00000000002b9f5f t __NSCreateEightBitMeshedBitmap
0000000000629c21 T __NSCreateIconFamilyFromNSImage
0000000000648452 T __NSCreateImageRef
0000000000193142 T __NSCreateImageRef2
000000000064839a t __NSCreateImageRefFromJPEGData
00000000001cce52 T __NSCreateKeyBindingState
00000000003e1d3b T __NSCreateKitImageNamed
00000000006427bd t __NSCreatePatternForGradient
000000000064291a T __NSCreatePatternForMetal
0000000000642a06 T __NSCreatePatternFromContext
0000000000642e16 T __NSCreatePatternFromPDF
0000000000086cef T __NSCreateShapeForRoundedWindow
0000000000b963f1 d __NSCreateWebArchiveOnAppKitThread
000000000001c55b T __NSCreateWindow
000000000064ad3c T __NSCreateWindowWithOpaqueShape
0000000000087161 t __NSCreateWindowWithOpaqueShape2
00000000003b3bb2 t ___NSCreateCollectionWithName
00000000003b3fd6 t ___NSCreateFontMenu
00000000000f1f12 t ___NSCreateRenderingContextForAttributedString
00000000000bf0b7 t ___NSCreateRenderingContextForString
00000000002684a2 t ___NSCreateStringDrawingTextStorageSettings

It is "simply" a library or so compiled against another library/framework having that function, but my library doesn't contain that function.

In previous prereleases it functioned correctly. Do you now if something was upgraded/downgraded in the SWT libs (dylibs) or other dylibs necessary for the jar?

[routeconverter]

In previous prereleases it functioned correctly. Do you now if something was upgraded/downgraded in the SWT libs (dylibs) or other dylibs necessary for the jar?

I've changed the SWT libs at the 29th of May. Could be anything on your machine or different flows in RouteConverter.

Maybe it's time for an update for SWT? Do you typically use the 64-bit jar?

[hvdwolf]

I've changed the SWT libs at the 29th of May. Could be anything on your machine or different flows in RouteConverter.

I still have the prerelease of 19-June and mid July. They work fine.
I have also been looking for an update on my machine as it could well be that "some" update would have caused this but in that case the older prereleases (after 29th may) wouldn't have run either.

Next to that: The "GetDYLDEntryPointWithImage" is not a fatal one as the older prereleases give the same error upon saving but function further correctly.

Maybe it's time for an update for SWT? Do you typically use the 64-bit jar?

I only use the 64bit jar.

I just synced from svn and built the jar myself. That doesn't make a difference.

[routeconverter]

I've changed the SWT libs at the 29th of May. Could be anything on your machine or different flows in RouteConverter.

I still have the prerelease of 19-June and mid July. They work fine.

So it might be a code path change in RouteConverter since mid July - probably connected to menus?.

I just synced from svn and built the jar myself. That doesn't make a difference.

They best way to debug would be if you'd started RouteConverter from an IDE like IntelliJ IDEA or Eclipse, put a breakpoint on ConvertPanel#saveFile()
and start the debugger.