On 2020-02-21 13:15, Luboš Luňák wrote: > On Thursday 20 of February 2020, Stephan Bergmann wrote: >> Why can Windows --enable-pch apparently not cope with >> <https://gerrit.libreoffice.org/plugins/gitiles/core/+diff/a7be9479a04de2f0 >> 059c5769bbd9dc354c84136b%5E%21> "Move Boost.Spirit usage away from legacy >> namespace"? > ... >>> C:/cygwin/home/tdf/lode/jenkins/workspace/lo_tb_master_win_dbg/slideshow/ >>> source/engine/smilfunctionparser.cxx: warning C4651: >>> '/DBOOST_SPIRIT_USE_OLD_NAMESPACE' specified for precompiled header but >>> not for current compile > > The build is apparently incremental, so the PCH is still built > with -DBOOST_SPIRIT_USE_OLD_NAMESPACE from a previous build. > I have no experience playing with PCH in LO build, but seeing this problem, and also similar problems when compiler updates, might this possibly be used: if we have a special compilation unit for PCH, could it be touched every time this module needs to recompile anything; then be compiled first time, checking for errors/warnings; then if failed, PCH deleted and the unit re-built again. Hopefully that could make it more robust, and in the same time would not impose a significant overhead (if my assumption that re-building the PCH unit does not automatically re-build PCH itself, thus typically being an almost no-op when PCH is up-to-date)? -- Best regards, Mike Kaganski
Attachment:
pEpkey.asc
Description: application/pgp-keys
_______________________________________________ LibreOffice mailing list LibreOffice@xxxxxxxxxxxxxxxxxxxxx https://lists.freedesktop.org/mailman/listinfo/libreoffice