Michael wrote: "In the second case, it would be cool if the plugin could prepare a bug report (plugin name, input geometry(ies) in WKT, stacktrace...) and propose the user to send it to the list."
I'm going to catch any exceptions from JTS and store them. I'll put the unprocessed features in there own layer, but hopefully I can dump the stack trace and other information from the exceptions in a log. I need to think some more about how to do this. But I agree the exception information is important, and shouldn't be lost. This project has got me thinking more about the best way to handle errors in plug-ins when calling outside libraries, especially JTS. It is an interesting problem with several possible solutions. I'll have to check my code. I didn't see the error catching code you mentioned, so I may be working with an older version of the class. The Sunburned Surveyor 2010/10/29 Michaël Michaud <michael.mich...@free.fr>: > Hi, >> Michael wrote: "May I ask you what you plan to fix on UnionByAttribute." >> >> Certainly. >> >> Currently, if the input features to the plug-in processing trigger an >> exception from JTS, the plug-in chokes. I'm tweaking the plug-in so >> that it reports exceptions from the underlying JTS operations and >> continues processing the remaining input features. > OK, I agree that OpenJUMP should be able to report topology errors > instead of throwing exceptions. >> I'm curious what was broken in the plug-in that caused Nacho problems? >> Was it a problem with the union operation in JTS, with his data, or >> with the plug-in code? Do you remember? > Nacho dataset contained invalid geometries (linestrings made of two > identical points). > Invalid geometry input is one of the most frequent case for getting > topologic exception. > For this plugin, I just catched invalid geometries (at least this case) > before processing them and I reported their id in the html output window. > > Could have put them in a new layer instead of reporting their ids. Feel > free to do this way if you think it's better... > > More difficult are the cases where input data is valid and processing > them generates topology errors. It may come from a plugin misconception > or from JTS. In the second case, it would be cool if the plugin could > prepare a bug report (plugin name, input geometry(ies) in WKT, > stacktrace...) and propose the user to send it to the list. > > Michaël > >> The Sunburned Surveyor >> >> 2010/10/28 Michaël Michaud<michael.mich...@free.fr>: >>> Hi, >>>> Just to clarify, the bug where the fill from the initial vertice size >>>> still appears within resized nodes has nothing to do with the color >>>> theming style. I apologize if I didn't make this clear in my bug >>>> report. Thanks for adding the additional comments to the bug report >>>> Michael. >>> There are probably two slightly different bugs. The one you described >>> succinctly, and the one arnd described after I asked for clarification. >>> So, by chance, it will help to fix two bugs instead of one ;-) >>>> I plan to work on the vertex styling bugs when I'm done with my error >>>> reporting fix to the UnionByAttribute plug-in. Rendering in OJ is not >>>> my strong point, so I may be asking for some advice when I get to the >>>> vertex styling bugs. :] >>> May I ask you what you plan to fix on UnionByAttribute. I have done a >>> few modifications about 6 month ago and I can union the shapefile Nacho >>> has sent about one year ago without exception (with an error reporting 4 >>> invalid geometries). >>> >>> Michaël >>>> The Sunburned Surveyor >>>> >>>> 2010/10/28 Michaël Michaud<michael.mich...@free.fr>: >>>>> Hi, >>>>> >>>>> Thanks for clarification. >>>>> I now see several buggy behaviours with Vertex style : >>>>> - VertexStyle is still active (and use the basic style color) when a >>>>> ColorTheming is added (what you say) >>>>> - VertexStyle disappears when you click on ChangeStyle menu or button >>>>> (and reappear after if you don't change anything) >>>>> - LineString vertices are not drawn with just a BasicStyle, but if you >>>>> add Vertex, you can have up to 3 representations of the vertex (ex. with >>>>> linewidth = 10 and vertex size = 15, you get a default small square, a >>>>> vertex with the shape you choosed and the line width size, a vertex with >>>>> the shape you choosed and the vertex size !) >>>>> >>>>> Currently, ColorTheming style can only use Basic style (fill >>>>> color/pattern, line color/pattern/width, transparency). I don't know how >>>>> much work it would be to extend this. >>>>> >>>>> I'll add these precision in the bug report >>>>> Thanks for help >>>>> >>>>> Michaël >>>>> >>>>> Le 28/10/2010 10:00, Arnd Kielhorn a écrit : >>>>>> Hello, >>>>>> >>>>>> I think what Sunburned mean is that: >>>>>> >>>>>> With the Change Style dialog You change the vertices size (esp. point >>>>>> layer). But when You also want to make a colour theming it is whithour >>>>>> an effect...You only see that vertices with larger size with their >>>>>> origin uniform colour. But when You zoom or look very contrated to the >>>>>> vertices You can see that the thematically colour vertices lay >>>>>> underneath the enlarged vertices like to layers laying one upon the >>>>>> other. >>>>>> The result of this buggy effect for the users is to decide if You want >>>>>> to have resized vertices or thematically coloured. But the aim should be >>>>>> to make it independ, e.g. You can get larger vertices which are >>>>>> themically coloured. >>>>>> >>>>>> Kindly regards >>>>>> Arnd >>>>>> >>>>>>> Hi Sunburned, >>>>>>> >>>>>>> You filled a bug report about vertex styling. >>>>>>> Can you describe a reproductible case where you noticed that vertices >>>>>>> are painted twice. >>>>>>> I cannot see any evidence of double painting by just styling a layer. >>>>>>> >>>>>>> Thanks >>>>>>> >>>>>>> Michaël >>>>>>> >>>>>>> ------------------------------------------------------------------------------ >>>>>>> Nokia and AT&T present the 2010 Calling All Innovators-North America >>>>>>> contest >>>>>>> Create new apps& games for the Nokia N8 for consumers in U.S. and >>>>>>> Canada >>>>>>> $10 million total in prizes - $4M cash, 500 devices, nearly $6M in >>>>>>> marketing >>>>>>> Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store >>>>>>> http://p.sf.net/sfu/nokia-dev2dev >>>>>>> _______________________________________________ >>>>>>> Jump-pilot-devel mailing list >>>>>>> Jump-pilot-devel@lists.sourceforge.net >>>>>>> https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel >>>>> ------------------------------------------------------------------------------ >>>>> Nokia and AT&T present the 2010 Calling All Innovators-North America >>>>> contest >>>>> Create new apps& games for the Nokia N8 for consumers in U.S. and >>>>> Canada >>>>> $10 million total in prizes - $4M cash, 500 devices, nearly $6M in >>>>> marketing >>>>> Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store >>>>> http://p.sf.net/sfu/nokia-dev2dev >>>>> _______________________________________________ >>>>> Jump-pilot-devel mailing list >>>>> Jump-pilot-devel@lists.sourceforge.net >>>>> https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel >>>>> >>>> ------------------------------------------------------------------------------ >>>> Nokia and AT&T present the 2010 Calling All Innovators-North America >>>> contest >>>> Create new apps& games for the Nokia N8 for consumers in U.S. and >>>> Canada >>>> $10 million total in prizes - $4M cash, 500 devices, nearly $6M in >>>> marketing >>>> Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store >>>> http://p.sf.net/sfu/nokia-dev2dev >>>> _______________________________________________ >>>> Jump-pilot-devel mailing list >>>> Jump-pilot-devel@lists.sourceforge.net >>>> https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel >>>> >>>> >>> >>> ------------------------------------------------------------------------------ >>> Nokia and AT&T present the 2010 Calling All Innovators-North America contest >>> Create new apps& games for the Nokia N8 for consumers in U.S. and Canada >>> $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing >>> Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store >>> http://p.sf.net/sfu/nokia-dev2dev >>> _______________________________________________ >>> Jump-pilot-devel mailing list >>> Jump-pilot-devel@lists.sourceforge.net >>> https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel >>> >> ------------------------------------------------------------------------------ >> Nokia and AT&T present the 2010 Calling All Innovators-North America contest >> Create new apps& games for the Nokia N8 for consumers in U.S. and Canada >> $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing >> Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store >> http://p.sf.net/sfu/nokia-dev2dev >> _______________________________________________ >> Jump-pilot-devel mailing list >> Jump-pilot-devel@lists.sourceforge.net >> https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel >> >> > > > ------------------------------------------------------------------------------ > Nokia and AT&T present the 2010 Calling All Innovators-North America contest > Create new apps & games for the Nokia N8 for consumers in U.S. and Canada > $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing > Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store > http://p.sf.net/sfu/nokia-dev2dev > _______________________________________________ > Jump-pilot-devel mailing list > Jump-pilot-devel@lists.sourceforge.net > https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel > ------------------------------------------------------------------------------ Nokia and AT&T present the 2010 Calling All Innovators-North America contest Create new apps & games for the Nokia N8 for consumers in U.S. and Canada $10 million total in prizes - $4M cash, 500 devices, nearly $6M in marketing Develop with Nokia Qt SDK, Web Runtime, or Java and Publish to Ovi Store http://p.sf.net/sfu/nokia-dev2dev _______________________________________________ Jump-pilot-devel mailing list Jump-pilot-devel@lists.sourceforge.net https://lists.sourceforge.net/lists/listinfo/jump-pilot-devel
