summaryrefslogtreecommitdiff
path: root/solenv/gbuild
AgeCommit message (Collapse)Author
2013-02-28fix make <module> for BUILDDIR <> SRCDIRMatúš Kukan
Change-Id: I63528f351685167c1d67551d92afc2826a8b42ac
2013-02-27gbuild: add dependencies on makedepend to rules that use itMichael Stahl
Change-Id: I25d23cd9a526b0d0be37ea0a0211e407266e6475
2013-02-27sal: get rid of CustomTarget_generated.mkMichael Stahl
Change-Id: I2347f60ef922d32201beda47e59e0b345d07a3b1
2013-02-27pack .cfg and .tree filesDavid Tardon
Change-Id: I2a7936c9c71d65482281816720b20b920c2e3e74
2013-02-27actually pack the help jarDavid Tardon
Change-Id: Iec3ca941d16cdfd52711a3effb96c865cbd24069
2013-02-27make sure that outdir existsDavid Tardon
Change-Id: Id481219cae8af1cea90dbc0f1d28f6a6483927d8
2013-02-27move instsetoo_native to gbuildBjoern Michaelsen
Change-Id: Ie4deddc426356636b76ad4fc27552c6d3a1dd815 Reviewed-on: https://gerrit.libreoffice.org/2429 Tested-by: LibreOffice gerrit bot <gerrit@libreoffice.org> Tested-by: Caolán McNamara <caolanm@redhat.com> Reviewed-by: Caolán McNamara <caolanm@redhat.com>
2013-02-27Do not depend on (not-yet) delivered .xcs files in --stringparam schemaRootStephan Bergmann
.xcs/.xcu files can reference other .xcs files via oor:component attributes, and the various xsltproc scripts called from solenv/gbuild/Configure.mk try to load the referenced .xcs files, locating them via --stringparam schemaRoot arguments. These references among .xcs/.xcu files are not tracked via dependencies in the build system. When building the primary registry in module officecfg, schemaRoot must therefore not reference solver, as any referenced .xcs files need not yet be delivered. In the old dmake build system, this was solved by basing the schemaRoot argument on an ABSXCSROOT variable that was set differently in officecfg/util/makefile.pmk (pointing into officecfg/ itself) than in other modules (like in connectivity/makefile.pmk, pointing into solver/). Reproduced that hack in the new gbuild system now, too (which already had that hack partially implemented for .xcs files). Change-Id: Ib2300686c4de144f0b1e18d0e985e6fc7d9f64ae
2013-02-27Clean up deliver, showdeliverables, showmodules targetsStephan Bergmann
- deliver was only relevant for dmake - showdeliverables works only per-module - "'showmodules' is now pointless, it was used to identify modules from tail_build" [<http://lists.freedesktop.org/archives/libreoffice/2013-February/046510.html> "Re: Show module dependencies"] Change-Id: I400ec5bed4397a12ce4c5b604253c74ef7de0ce7
2013-02-27gbuild: add ZipPackage targetMichael Stahl
This is basically a Zip and a Package in one, so it's not necessary to enumerate the content twice. The zip is sadly the easiest way to get a large number of files through scp2 and into instsets. Change-Id: Idadcced76bf10317fc348ac91203c15c47e8762d
2013-02-26fixes for building in a separate dirPeter Foley
Change-Id: I244fa0c54c8d98827ab0a987ba353dedf733e695
2013-02-26postprocess does not care for _any_ instsetoo_native or postprocess packagesBjoern Michaelsen
Change-Id: Ieb80176a59f8a41a89c1926827f3190d006a2fd1
2013-02-26move postprocess to gbuildBjoern Michaelsen
the gb_Postprocess* foo could also be in gb_Module* as it is conceptionally close ('do things globally/productwide'). OTOH I see no obvious reason for e.g. signing not to be done right along with building a lib/executable anyway instead of in postprocess. The same is likely true for the other stuff too. Change-Id: I9c8f569564c056643af7ca39bfe038ed228dcd3d Reviewed-on: https://gerrit.libreoffice.org/2426 Reviewed-by: Matúš Kukan <matus.kukan@gmail.com> Reviewed-by: Fridrich Strba <fridrich@documentfoundation.org> Tested-by: Fridrich Strba <fridrich@documentfoundation.org>
2013-02-26gbuild: Zip/Jar targets: depend on makefile to rebuild on removalMichael Stahl
Change-Id: I0b79b6c9018e4513b9f0b16be3e57f620d12ea23
2013-02-26add support for building help packs into gbuildDavid Tardon
Change-Id: I1cc5600ab0f6236c451d07bfb5ad9ee42ea4754a
2013-02-26allow to create non-delivered Zip targetsDavid Tardon
Change-Id: I3496d815d4f2fedf04ca2c1727f37362c7395296
2013-02-25gbuild: ComponentTarget: rebuild if library names changeMichael Stahl
Stale component files result in failure to load libraries in incremental builds; because rebuilding these is fast just depend on the 2 Repository makefiles that define the library file names. Change-Id: Ia72a0460d3bb8bceb0e17334415ca3dde0401c24
2013-02-25Revert inadvertent changeStephan Bergmann
Change-Id: I203ad92e8a006d1f262203852f05108706a90ee4
2013-02-25Last use of obsolete gb_*_add_api is long goneStephan Bergmann
Change-Id: Ia6c938ef0e45ec86c0c202e95ef28d49fccd3497
2013-02-25add solenv to tail_buildPeter Foley
Change-Id: Ib8987f0d45ed5360d49f3589b418209488267863 Reviewed-on: https://gerrit.libreoffice.org/2373 Tested-by: LibreOffice gerrit bot <gerrit@libreoffice.org> Reviewed-by: David Ostrovsky <David.Ostrovsky@gmx.de> Tested-by: David Ostrovsky <David.Ostrovsky@gmx.de>
2013-02-24fix HAVE_CXX0X case on MinGWDavid Ostrovsky
During configure phase HAVE_CXX0X is determined and exposed if the current gcc version supports it. With gcc 4.7.2 this is the case. Activate it during compilation phase too. Change-Id: I7b738de083fa337a422c20bdf88d9dca5f8303ed
2013-02-24document generic <class>_<target> idiom in gbuildDavid Ostrovsky
Change-Id: I8b746c3786eda42c3c91c104daee2018ba20b975
2013-02-24convert moz to gbuild and add to tail_buildPeter Foley
Change-Id: I2d3f5d127f831ee9abc020d46558f7cfbe8c9380 Reviewed-on: https://gerrit.libreoffice.org/2349 Reviewed-by: David Ostrovsky <David.Ostrovsky@gmx.de> Tested-by: David Ostrovsky <David.Ostrovsky@gmx.de>
2013-02-23convert odk to gbuild and add to tail_buildPeter Foley
Change-Id: I8e8de7f2bb87cce7916c7c2df24c1b0ddaea55c0 Reviewed-on: https://gerrit.libreoffice.org/2288 Reviewed-by: Peter Foley <pefoley2@verizon.net> Tested-by: Peter Foley <pefoley2@verizon.net>
2013-02-22quiet external module build log unless failureNorbert Thiebaud
ExternalProject usually involve a configure and a make step that produce a bunch of output usually irrelevant including a large number of warning and other mess. now that everything is pretty much in tail_build these output get interleaved with useful output from the build of the product and actually drown them in a logorrhea of messy noise. This store the output of external modules in a log file and only print them as a whole if the module failed do build. on a non-verbose build. Change-Id: I3abfcccd6d16821a9e061a71e031b427cc283647 Reviewed-on: https://gerrit.libreoffice.org/2304 Reviewed-by: Norbert Thiebaud <nthiebaud@gmail.com> Tested-by: Norbert Thiebaud <nthiebaud@gmail.com>
2013-02-17convert sysui to gbuild and add to tail_buildPeter Foley
Change-Id: Ia32e51f0d95e001bcf07766f6340398e0ab1bf6a Reviewed-on: https://gerrit.libreoffice.org/2192 Tested-by: LibreOffice gerrit bot <gerrit@libreoffice.org> Reviewed-by: David Ostrovsky <David.Ostrovsky@gmx.de> Tested-by: David Ostrovsky <David.Ostrovsky@gmx.de>
2013-02-15handle more than one lib correctly :/Bjoern Michaelsen
Change-Id: I1c6478e0ed69c7678598eb0d65cd5043d6c8dd99
2013-02-15Remove the need to manually register libs in scp2Bjoern Michaelsen
- we have all the needed information in gbuild - reduce the amount of double and triple bookkeeping - hopefully the generated files for scp2 are only a intermediate step and we can do whatever perl voodoo scp2 does directly in gbuild - currently using basctl as an example - this also means we have to be more strict in Repository.mk with what we register: - if we dont build a libs, it shouldnt be registered (otherwise scp2 will want the nonexisting lib) - in the end, this is a Good Thing(tm): we dont want anyone trying to e.g. link against a lib we dont build - should now hopefully work on all platforms - added module bookkeeping to gbuild Change-Id: I241ca67edbfabae83859274f9aa35d2d26165ef6
2013-02-14Build DESKTOP stuff for BUILD only if HOST is of type DESKTOPTor Lillqvist
The helplinker library has grown a dependency on the svt and tl libraries. But the helpcompiler module is needed to be built for the BUILD platform only if the HOST platform's BUILD_TYPE includes DESKTOP (or only if its BUILD_TYPE includes HELP?). We used to build helpcompiler and some other things for the BUILD platform depending on whether the BUILD platform's BUILD_TYPE included DESKTOP. Also, codemaker should surely be built for BUILD unconditionally, ant xmlhelp presumably not at all. Change-Id: I3df995ce6b92da4ce9eb35115cf3e8560e06f292
2013-02-09make it possible to limit what files will be modified by a compiler pluginLuboš Luňák
Change-Id: I4e3e8f5ca5e5b5b59b1bd067281f90940dd893b1
2013-02-09treat all source files as not up to date only with FORCE_COMPILE_ALLLuboš Luňák
So running the compiler plugin can be done either with FORCE_COMPILE_ALL=1 for simplicity, running the plugin on all source files for the target, or it's possible to use make's -W option (or touch files manually) to run it only for selected files. Change-Id: Ifa82bbe91dc1e179af9db8fbf99271bcf03a1d47
2013-02-06pagein: learn to respect --enable-mergelibs caseMatúš Kukan
Add libmerged to pagein-common and remove merged libraries from there. Also fixes few misspelled library names. This should make pagein usefull also for libmerged and speed up cold start by pre-loading it. Change-Id: I80ffb263ed5f399dc639e0d054f782ecb85001d6
2013-02-06Propagate also ENABLE_MACOSX_SANDBOX to compilationsTor Lillqvist
Change-Id: I6250ec56c7b6d8f1da94b98e494f483c4076844c
2013-02-05Add --enable-readonly-installset optionTor Lillqvist
Not yet implemented in the code, but my idea is that any functionality that modifies the system-wide installation will go away in this case. Automatically set if --disable-externsions, or if building a sandboxed LO for OS X. Should probably be set automatically also when just building a signed (but not necesssarily sandboxed) LO for OS X? Surely installing a system-wide extension should count as tampering with the app. Especially if we can make also extension (scripts) be signed (by locating them in the Resources folder?) Change-Id: Id654bfaa6331535a66eae1bc6531a756085a3f06
2013-02-02Sign executables right after linkingTor Lillqvist
Change-Id: I27c929deffeb803e103609634b501622e04da033
2013-02-01No need for aliases hereStephan Bergmann
Change-Id: I29dcc26eb13222ef4348ab3281a19b2e6f8c46f2
2013-02-01one more config_host/ breakageLuboš Luňák
2013-02-01gbuild: copy stdlibs into solver again when they are bundled in the installerPetr Mladek
Otherwise the build --without-system-stdlibs fails; This option is used for the universal Linux build Change-Id: I0f7d87bb161d50748fee05f9c3f6a4b87b788aa4 Reviewed-on: https://gerrit.libreoffice.org/1938 Reviewed-by: David Tardon <dtardon@redhat.com> Tested-by: David Tardon <dtardon@redhat.com>
2013-01-31Use separate config_*.h files for BUILD and HOST when cross-compilingTor Lillqvist
We have only been lucky so far that there hasn't been any compilation errors caused by using the HOST platform config_*.h files also for the BUILD platform. But as more and more information found out during configury is stored in config_*.h files, sooner or later it will be a problem. Move config/* to config_host/*. When not cross-compiling, the config_*.h files in config_host expanded from config_*.h.in are all there is. When cross-compiling, the generated config_*.h files for the BUILD platform are put into a config_build folder. Then use config_build when building the cross_toolset and config_host otherwise. Change-Id: I3c28cc83866bffd696b45d00dfb68e00bc2ae538
2013-01-30Get rid of STAR_RESOURCEPATHStephan Bergmann
...the only remaining use was for CppunitTest to find the .res files in $(OUTDIR)/bin/. Get that solved by delivering the .res files also to $(OUTDIR)/unittest/install/program/resource/ (new gb_ResTarget_get_unittest_target). Change-Id: I370e7cab4223256356c7511b88ae63d9c0dbf215
2013-01-30Apparent copy/paste typoStephan Bergmann
Change-Id: Ie4986f36f7f71df64b1c7c7b5e0190f4d5708c64
2013-01-21Remove -Wsign-promoStephan Bergmann
...to avoid unhelpful "passing [enum] chooses 'int' over 'unsigned int'" warnings from those new calls to rtl::OUString::number as introduced with 63bcb139b941a2eff1b5ad367046bca067e7d1f8 "Replaced O[U]String::valueOf( static_cast<> ) with O[U]String::number()." Change-Id: I2adb16e83865bd4cba4a1650eb54bb7f5cee31e0
2013-01-16T_xxxFLAGS_APPEND fixupsLuboš Luňák
http://lists.freedesktop.org/archives/libreoffice/2013-January/043792.html
2013-01-14append library specific compile flags at the endMatúš Kukan
Change-Id: If9d4640b59dd411734a581c66f9310d31ab9fbb7 Reviewed-on: https://gerrit.libreoffice.org/1632 Tested-by: Norbert Thiebaud <nthiebaud@gmail.com> Reviewed-by: Noel Power <noel.power@suse.com> Tested-by: Noel Power <noel.power@suse.com>
2013-01-11Set version number in AC_INIT() in configure.ac onlyTor Lillqvist
Output a solenv/inc/minor.mk based on that. Get rid of RSCVERSION and RSCREVISION. Automate the end of the copyright year range, although I am actually not sure if that is OK. Change-Id: Ibae2d139b1dd468ee80d18a2a41b786a532df809
2013-01-11QUARTZ is equivalent to MACOSXTor Lillqvist
So just check MACOSX in the sources, and don't pass any -DQUARTZ to compilations. Change-Id: Ia1de6024810f81f1eea01b8eb0b885d9d1ee8e20
2013-01-10optimize for build targets in one placeMatúš Kukan
Change-Id: Ib886a323c1ade6dcd455a44d3dfc6702086416e3
2013-01-10remove effectively unused gb_PARTIALBUILDMatúš Kukan
Change-Id: I5d00db54748de0e1ce3de3dd9c99a88a82e7917b
2013-01-10dev-install target is used from toplevel MakefileMatúš Kukan
Change-Id: I5b53046e9389eac2126e910752d131e7b65bf53c
2013-01-10Add configure switch that allows hardlink deliverTomas Chvatal
This switch just makes possible to configure in ln delivery instead of cp based one. The feature was already in the place, we just make it visible as configure option now. Change-Id: I778ef3b03e6b9a57057e66ff6307e7da46bcd935
generated by cgit (git 2.34.1) at 2025-03-06 15:54:40 +0000