Hi Alle Meije, Am Dienstag, 22. Juli 2014, 14:59:19 schrieb Alle Meije Wink:
I followed a slightly different recipe this time, first using 'cmd.exe' (in administrator mode) and MSYS bash:
cd %HOMEPATH%\usr\local bash git config --global core.autocrlf true git clone --recursive https://github.com/boostorg/boost.git boost > clone.log exit bootstrap gcc b2 -a --prefix=c:\users\amwink\usr\local –build-type=complete --build-dir=build toolset=gcc link=shared runtime-link=shared threading=multi
which completes without errors.
You should see: ...update XXX targets ...failed YYY targets at the end of the build output. If you don't see "failed", everything build fine.
If I then re-enter my original build command
b2 -a -d+2 -q --prefix=c:\users\amwink\usr\local –build-type=complete --build-dir=build toolset=gcc link=shared runtime-link=shared threading=multi
it stops again at the point where the assembler is needed (see previous messages).
Yes, as "-d+2" just enables command line output, whereas -q activates "quit on first error". In this case you have make-like behaviour.
Does this mean that the build with the first call of b2 is actually successful?
No, the first build just ignores the build errors and happily builds the rest of the libraries. It just the same as running "make -k".
Or does the call without the " -d+2 -q " options compile/check parts of the building process that the first call skips?
No. Note that the offending commandline will be printed in the error anyway, so there is normally no need to add "-d +2". Hope this helps. Yours, Jürgen -- * Dipl.-Math. Jürgen Hunold ! * voice: ++49 4257 300 ! Fährstraße 1 * fax : ++49 4257 300 ! 31609 Balge/Sebbenhausen * jhunold@gmx.eu ! Germany