[mkgmap-dev] TYP files and character encoding

Fri Jan 17 16:26:53 GMT 2020

Hi Gerd

That looks fine.

I'll add something to ./doc/typ-compiler.txt sometime soon that
describes the source/destination charset/codepage behaviour.

Ticker  

On Fri, 2020-01-17 at 15:21 +0000, Gerd Petermann wrote:
> Hi Ticker,
> 
> thanks, I've committed the patch with 4423, please check if my svn
> log message.
> 
> Gerd
> 
> 
> ________________________________________
> Von: mkgmap-dev <mkgmap-dev-bounces at lists.mkgmap.org.uk> im Auftrag
> von Ticker Berkin <rwb-mkgmap at jagit.co.uk>
> Gesendet: Freitag, 17. Januar 2020 13:43
> An: mkgmap development
> Betreff: Re: [mkgmap-dev] TYP files and character encoding
> 
> Hi Gerd
> 
> Here is typCodePage_v4 that uses try () in both CharsetProbe and
> compile, getting rid of Utils.closeFile().
> 
> The patch includes the change in typCodePage-test.patch
> 
> My javac doesn't seem to have an option to detect unused imports, but
> when I run it with -Xlint I get a variety of errors - I've attached
> the
> log.
> 
> Ticker
> 
> On Fri, 2020-01-17 at 11:13 +0000, Gerd Petermann wrote:
> > Hi Ticker,
> > 
> > I use Eclipse with customized settings in Preferences -> Java
> > ->Compiler-> Error/Warnings as well as the SonarLint plugin.
> > 
> > My understanding is that the InputStream is only closed if
> > everything
> > goes well. The nature of unit tests is that they produce special
> > cases".
> > The try-with-ressources was introduced to handle this.
> > Maybe you can post a v4 which uses try-with-ressources in class
> > CharsetProbe (as the unpatched version does)?
> > 
> > Gerd
> > 
> > 
> > ________________________________________
> > Von: Ticker Berkin <rwb-mkgmap at jagit.co.uk>
> > Gesendet: Freitag, 17. Januar 2020 11:54
> > An: Gerd Petermann
> > Betreff: Re: AW: AW: [mkgmap-dev] TYP files and character encoding
> > 
> > Hi Gerd
> > 
> > I have another patch almost ready for the StandardCharset utf8 /
> > try
> > (with-resources) etc, but this is quite wide-ranging and unrelated
> > to
> > the need for the typCodePage patch.
> > 
> > I wanted to get the typCodePage patches committed first so that I
> > can
> > get on with mapnik.txt patches and also improve some bits of
> > TypCompiler as the last part of the utf8 patch.
> > 
> > My compilation system doesn't warn about unused imports - what
> > options/tool do you use for this?
> > 
> > Concerning the close, the FileInputSteam is closed, which should
> > release any OS file handle; it's just the InputSteamReader &
> > BufferedReader that arn't, but these are just java data structures.
> > 
> > Ticker
> > 
> > On Fri, 2020-01-17 at 10:27 +0000, Gerd Petermann wrote:
> > > Hi Ticker,
> > > 
> > > ah, seems I got another post wrong. I thought you'd work on a
> > > typCodePage_v4.patch which would use try-with-ressources.
> > > With typCodePage_v3.patch  and typCodePage-test.patch I still see
> > > some warnings for TypCompiler:
> > > - unused imports
> > > - br is not closed (line 232)
> > > 
> > > Gerd
> > > 
> > > ________________________________________
> > > Von: Ticker Berkin <rwb-mkgmap at jagit.co.uk>
> > > Gesendet: Freitag, 17. Januar 2020 10:17
> > > An: Gerd Petermann
> > > Betreff: Re: AW: [mkgmap-dev] TYP files and character encoding
> > > 
> > > Hi Gerd
> > > 
> > > Yes. Sorry - I didn't explain it at all well. This needs to be
> > > applied
> > > at the same time as typCodePage_v3.patch from 14-Jan
> > > 
> > > Ticker
> > > 
> > > On Fri, 2020-01-17 at 08:48 +0000, Gerd Petermann wrote:
> > > > Hi Ticker,
> > > > 
> > > > I don't understand this patch. Do I have to use it in
> > > > combination
> > > > with another one?
> > > > 
> > > > Gerd
> > > > 
> > > > ________________________________________
> > > > Von: mkgmap-dev <mkgmap-dev-bounces at lists.mkgmap.org.uk> im
> > > > Auftrag
> > > > von Ticker Berkin <rwb-mkgmap at jagit.co.uk>
> > > > Gesendet: Freitag, 17. Januar 2020 00:02
> > > > An: Development list for mkgmap
> > > > Betreff: Re: [mkgmap-dev] TYP files and character encoding
> > > > 
> > > > Hi Gerd
> > > > 
> > > > I've just noticed that a change to a function profile stopped a
> > > > test
> > > > from compiling, so here is a patch for that
> > > > 
> > > > Ticker