[GRASSLIST:4272] Re: compiling Grass5pre5 on Cygwin (wrap up)

[Glynn Clements]

Thomas Dewez wrote:

> Configure didn't find the inconv() function. I don't know what it does but I
> assumed it wouldn't be too important (right?). The inconv.exe file in /bin
> and the libinconv.a is in /lib. Don't know what package inconv works with so
> I couldn't link it.

It's "iconv", not "inconv". On my Cygwin installation (which hasn't
been updated for a few months), the function is in libintl.a.

The purpose of iconv() is to convert text between different encodings. 
The only program which uses it is d.text.freetype. d.text.freetype
should compile and run without iconv(); however, you will only be able
to display text which uses the ISO-8859-1 encoding (maybe Windows-1252
as well).

> Configure couldn't find odbc (hence the without odbc switch) though again it
> existed in /lib/w32api and /usr/include/w32api. Adding those paths didn't
> seem satisfactory in --with-odbc-includes and --with-odbc-libs. Don't know
> why but I skipped it as a result.

GRASS' ODBC functionality uses the UnixODBC library. It won't work
with the Win32 ODBC library (or iODBC, which is the other popular Unix
ODBC implementation).

> Due to a problem between Cygwin and XP Pro (this was a fresh install of the
> latest Cygwin) I had to redirect the standard output to a log file. I don't
> know whether other Cygwin users have observed this before but when Cygwin
> produces a large amount of output to the screen, the System process
> (monitored with the Windows Task manager) takes over more than 95% of the
> CPU power and sometimes crashes (Blue Screen of Death with error 0x96 or
> error 0xA). This is not the case when the output is directed to a file,
> though this doesn't say why this happens. One of the symptoms of imminent
> crash is the fan speeding up.

Is this using Windows' "command prompt" window, or Cygwin's rxvt? You
could try changing to see if the problem is specific to one of them.

> Finally, Grass5pre5 is working quite happily in XWin as long as I am not
> trying to do any intensive computing (especially r.surf.rst on 10^6 sites).
> Though I was expecting r.surf.rst to be significantly improved in speed and
> efficiency on version pre5, Cygwin and/or XP are still not happy with large
> volumes of data. I got again a Blue Screen of Death with error 0xA in
> writing (third info is 0x1 which means writing error) just after lunch. Any
> hint on this would be hugely appreciated.

Issues which are specific to running GRASS on Cygwin might be better
addressed to the wingrass list (wingrass at grass.itc.it; you have to be
subscribed to post), or even a Cygwin list (XP is fairly new, so I'd
expect it to take some time to update Cygwin accordingly).

-- 
Glynn Clements <glynn.clements at virgin.net>