<HTML><HEAD>
<META http-equiv=Content-Type content="text/html; charset=iso-8859-15">
<META content="MSHTML 6.00.2800.1505" name=GENERATOR></HEAD>
<BODY style="MARGIN: 4px 4px 1px; FONT: 10pt Comic Sans MS">
<DIV>Tamas,</DIV>
<DIV> </DIV>
<DIV>I'm wondering about the real need for XML version of MapFiles as well. Before going to town with this I think I need to plpay around with some alternatives to see what might/might not work for our needs, which is basically to build a GUI editor for Cartography, and another future item, is to be able to manage many layers in an abstracted mode. We use sepearte MapFiles now for each layer in GeoMoose for example.</DIV>
<DIV> </DIV>
<DIV>I've looked at some of the editors, but they don't entirely come up to snuff and seem hard to keep inline with development.</DIV>
<DIV> </DIV>
<DIV>My first thought was to use Shell scripting to do the XML writing. But I agree with your thought about not all issues are being taken into account.</DIV>
<DIV> </DIV>
<DIV>I wonder if it makes sense to try and limit the scope of the XML definitions, maybe a subset of the params in the LAYER block or some such. Still pondering . . . .</DIV>
<DIV>bobb</DIV>
<DIV><BR><BR>>>> "Tamas Szekeres" <szekerest@gmail.com> wrote:<BR></DIV>
<DIV style="PADDING-LEFT: 7px; MARGIN: 0px 0px 0px 15px; BORDER-LEFT: #050505 1px solid; BACKGROUND-COLOR: #f3f3f3">Bob,<BR><BR>Here`s a demo example to show what the XML writer would do:<BR><BR>typedef struct object_info {<BR>char* name; // may be the corresponding mapfile item name<BR>void* value; // reference to the object or value (string)<BR>char* type; // type of the item<BR>bool isValueType;<BR>int index;<BR>} objectInfo;<BR><BR>int WriteObjectAsXML(objectInfo i, msIOContext *context)<BR>{<BR> char buffer[1024];<BR> objectInfo s;<BR> if (i.isValueType)<BR> {<BR> sprintf(buffer, "<%s type=%s>%s</%s>", i.name, i.type,<BR>(char*)i.value, i.name);<BR> msIO_contextWrite(context, buffer, strlen(buffer));<BR> }<BR> else<BR> {<BR> sprintf(buffer, "<%s type=%s>", i.name, i.type);<BR> msIO_contextWrite(context, buffer, strlen(buffer));<BR> while (s = GetNextSubItem(i))<BR> WriteObjectAsXML(s, context);<BR> sprintf(buffer, "</%s>", i.name);<BR> msIO_contextWrite(context, buffer, strlen(buffer));<BR> }<BR>}<BR><BR>int WriteMapAsXML(mapObj map, msIOContext *context)<BR>{<BR> objectInfo i, s;<BR> i.name = "MAP"<BR> i.value = map;<BR> i.type = "mapObj";<BR> WriteObjectAsXML(i, context);<BR>}<BR><BR>However I don`t think that all of the possible issues have been taken<BR>into account. Just for something to think about.<BR><BR>Best regards,<BR><BR>Tamas<BR><BR><BR><BR>2008/6/11 Bob Basques <Bob.Basques@ci.stpaul.mn.us>:<BR>> Tamas,<BR>><BR>> I know this would be easier to add after an XML conversion process were<BR>> arrived at, but it's going further than what I was intending, at least in<BR>> the near term.<BR>><BR>> It does get me thinking about things in more detail though, how would your<BR>> ideas translate into the XML itself, I mean would there be things that would<BR>> be bad to use in the XML that might mitigate these ideas?<BR>><BR>> I subscribed to the DEV list, but no Email confirmation yet.<BR>><BR>> bobb<BR>><BR>>>>> "Tamas Szekeres" <szekerest@gmail.com> wrote:<BR>> Hi Bob and All,<BR>><BR>> I would support having an interface in mapserver that would provide<BR>> low level access to the object hierarchy, like:<BR>><BR>> typedef struct object_info {<BR>> char* name;<BR>> void* value;<BR>> char* type;<BR>> int index;<BR>> } objectInfo;<BR>><BR>> objectInfo* GetNextSubItem(objectInfo* object);<BR>><BR>> By using this, each persistence provider (MapFile/XML/JSON) could go<BR>> through the object hierarchy easily<BR>> In order to support reading the files, the mapObj properties in the<BR>> hierarchy could also be set in an unique manner, like:<BR>><BR>> int AddSubItem(objectInfo parent, objectInfo child);<BR>><BR>> We should also utilize an msIO like functionality so that the<BR>> providers would support reading from and writing to various kind of<BR>> streams/storage types.<BR>><BR>> We should probably switch to the -dev list with further implementation<BR>> details.<BR>><BR>> Best regards,<BR>><BR>> Tamas<BR>><BR>><BR>><BR>> 2008/6/10 Bob Basques <Bob.Basques@ci.stpaul.mn.us>:<BR>>> All,<BR>>><BR>>> We've been pondering some sort of alternative to the Mapfiles for a few<BR>>> years now. A preferable approach would be something that could be stored<BR>>> in<BR>>> a DB in some fashion for querying/assembly processes. It seems on the<BR>>> surface like a DB schema could be developed to handle the MapFile storage<BR>>> aspects.<BR>>><BR>>> A first step would be in how to best approach moving into an XML way of<BR>>> life<BR>>> for the MapFiles. Would it make any sense in the beginning to just build<BR>>> a<BR>>> MapFile2XML convertor (I would imagine this would be needed before anyone<BR>>> would sign up for XML) and once something like this is a state close to<BR>>> production, the innards of MapServer would then be made to parse the XML<BR>>> directly?<BR>>><BR>>> Another approach might be to do something like MapFile2SQL first, and then<BR>>> the MapFile2XML. This might save some time and seem like it would make<BR>>> standardization easier, since it would need to be inside of the DB fist.<BR>>><BR>>> Has anyone tried putting together any requirements list along these lines<BR>>> at<BR>>> all? Is it going to be something where we just need to jump in and build<BR>>> something even if it might be a wrong approach to begin with?<BR>>><BR>>> This thread just got me thinking is all . . .you know how dangerous that<BR>>> can<BR>>> be . . . :c)<BR>>><BR>>> bobb<BR>>><BR>>><BR>>>>>> "Tamas Szekeres" <szekerest@gmail.com> wrote:<BR>>> Hi,<BR>>><BR>>> MapServer currently doesn`t support any other persitence<BR>>> representation than the mapfiles. There have been some initial plans<BR>>> related to an XML format in this list, but no one had any motivation<BR>>> to implement that.<BR>>> So I think the best what you can do at the moment is to use reflection<BR>>> to access the properties along with some other members of the objects<BR>>> and serialize the values manually.<BR>>><BR>>> Best regards,<BR>>><BR>>> Tamas<BR>>><BR>>><BR>>> 2008/6/9 BrainDrain <paulborodaev@gmail.com>:<BR>>>><BR>>>> Is there any 'standart' fast(!) way/method to serialize (xml/json/other<BR>>>> markup) mapObj in C#? I need it for using server mapObj as JSON on rich<BR>>>> client app running on browser. Look at my method (using reflection):<BR>>>><BR>>>> public static ListDictionary PartialSerialize(object instance, Stack<BR>>>> callerTypes, Type[] excludeTypes)<BR>>>> {<BR>>>> ListDictionary result = new ListDictionary();<BR>>>> object val;<BR>>>><BR>>>> callerTypes.Push(instance.GetType());<BR>>>> PropertyInfo[] pis = instance.GetType().GetProperties();<BR>>>> foreach (PropertyInfo pi in pis)<BR>>>> {<BR>>>> if (pi.PropertyType.IsSerializable &&<BR>>>> !pi.PropertyType.IsArray)<BR>>>> result[pi.Name] = pi.GetValue(instance, new<BR>>>> object[0]);<BR>>>> else<BR>>>> {<BR>>>> //preventing useless nesting<BR>>>> if (!callerTypes.Contains(pi.PropertyType) &&<BR>>>> !((IList)excludeTypes).Contains(pi.PropertyType))<BR>>>> {<BR>>>> val = pi.GetValue(instance, new object[0]);<BR>>>> if (val != null)<BR>>>> result[pi.Name] = PartialSerialize(val,<BR>>>> callerTypes, excludeTypes);<BR>>>> }<BR>>>> }<BR>>>> }<BR>>>> callerTypes.Pop();<BR>>>> return result;<BR>>>> }<BR>>>> ...<BR>>>> So I can convert mapObj on serever to hashtable automatically an then<BR>>>> populate JSON object<BR>>>> (still need to call explicitly getLayer, getClass etc., but this is not a<BR>>>> problem):<BR>>>> ...<BR>>>> layers[i].Properties = Tools.PartialSerialize(layer, new Stack(), new<BR>>>> Type[3] { typeof(mapObj), typeof(hashTableObj), typeof(colorObj)});<BR>>>> ...<BR>>>> classes[j].Properties = Tools.PartialSerialize(layerClass, new Stack(),<BR>>>> new<BR>>>> Type[4] { typeof(layerObj), typeof(labelObj), typeof(hashTableObj),<BR>>>> typeof(colorObj)});<BR>>>> ...<BR>>>> styles[k].Properties = Tools.PartialSerialize(classStyle, new Stack(),<BR>>>> new<BR>>>> Type[2]{typeof(hashTableObj), typeof(colorObj)});<BR>>>> ...<BR>>>> mapStub.Properties = Tools.PartialSerialize(map, new Stack(), new<BR>>>> Type[11]<BR>>>> {<BR>>>> typeof(labelObj), typeof(hashTableObj), typeof(fontSetObj),<BR>>>> typeof(labelCacheObj), typeof(outputFormatObj[]), typeof(queryMapObj),<BR>>>> typeof(referenceMapObj), typeof(scalebarObj), typeof(symbolSetObj),<BR>>>> typeof(colorObj), typeof(legendObj)});<BR>>>> ...<BR>>>> JavaScriptSerializer class object allows to perform convertion to client<BR>>>> More often I use script method in my web service that can do it behind<BR>>>> the<BR>>>> scenes.<BR>>>><BR>>>> How do you do such kind of operation?<BR>>>> --<BR>>>> View this message in context:<BR>>>><BR>>>> <A href="http://www.nabble.com/mapObject">http://www.nabble.com/mapObject</A>-serialization-in-C--(variant)-tp17739919p17739919.html<BR>>>> Sent from the Mapserver - User mailing list archive at Nabble.com.<BR>>>><BR>>>> _______________________________________________<BR>>>> mapserver-users mailing list<BR>>>> mapserver-users@lists.osgeo.org<BR>>>> <A href="http://lists.osgeo.org/mailman/listinfo/mapserver">http://lists.osgeo.org/mailman/listinfo/mapserver</A>-users<BR>>>><BR>>> _______________________________________________<BR>>> mapserver-users mailing list<BR>>> mapserver-users@lists.osgeo.org<BR>>> <A href="http://lists.osgeo.org/mailman/listinfo/mapserver">http://lists.osgeo.org/mailman/listinfo/mapserver</A>-users<BR>>><BR>>> _______________________________________________<BR>>> mapserver-users mailing list<BR>>> mapserver-users@lists.osgeo.org<BR>>> <A href="http://lists.osgeo.org/mailman/listinfo/mapserver">http://lists.osgeo.org/mailman/listinfo/mapserver</A>-users<BR>>><BR>>><BR>> _______________________________________________<BR>> mapserver-users mailing list<BR>> mapserver-users@lists.osgeo.org<BR>> <A href="http://lists.osgeo.org/mailman/listinfo/mapserver">http://lists.osgeo.org/mailman/listinfo/mapserver</A>-users<BR>><BR></DIV></BODY></HTML>