[mkgmap-dev] LBL subfile details

Thu Sep 24 14:37:08 BST 2009

> From: Steve Ratcliffe <steve at parabola.me.uk>
> On 23/09/09 15:52, Anton Todorov wrote:
> > Here are some findings that i can not manage to code in java... In
> > mkgmap the text between the header and the first lbl section is
> > hardcoded to the string 'mkgmap'. This section is defined at the end
> 
> > original files the text is 'American'. In the source it is commented
> > as "// Sort descriptor ??? ...". I could speculate that this is kind
> > of dictionary name(like 'American English'). Another finding is
> 
> So that would explain what Sort Descriptor meant.  All cgpsmapper
> created files just have a string like 'map created with cgpsmapper'
> which is why mkgmap ones just have the hardwired string in them.
> 
As I wrote it is my toughts about the meaning of the exact text, but in all original files it is same. Even on an old MG Spain for example...

> > (when the texts are coded in 6bits) that the first byte of LBL1 is
> > 0xFF(in Metro Guide) which is decoded as 0x3F == 'End Marker'. In
> > mkgmap it is 0x00 that i think is not correct. All elements that
> > have no name have lbl1 pointer to this end marker so it should be
> > 0xFF. [...]
> 
> OK, I believe this.  But, have you noticed any problems caused by it though?
> 
I am trying to code an perl decoder to help me understand the file format. I am testing primary with original files so my routine for reading lbl1 is not working propperly...
My assumption is that mkgmap should create files closer to the original ones than to ones created by other compilers because it is possible to hit problem not clearly solved in the other compilers...

> > [...] If there are points for the cities and LBL1,2,3,4 are filled
> > in the overview map img, menu Find->Find Places has working City tab
> > in MapSource. Maybe overview img participate in the indexing in some
> 
> Ah ha, that is interesting.  I did recently notice that the trip and
> waypoint map that came with my GPS had city search even though it does
> not have the global index file.
> 
> The work I've started on the global index will also allow us to create
> a better overview map too so we could do this then.
> 
Cool! I will target my decoder to dump and analyse mdr files...

BR,
Anton Todorov