![](https://secure.gravatar.com/avatar/6114858d84b0f40cf715f12237ab67a1.jpg?s=120&d=mm&r=g)
Dear builders, I'm soon going to add XeTeX back to the list of sources/binaries to faciliate testing of some major changes. The binaries will be uploaded to TLContrib. I need to adapt the scripts first and will let you know as soon as it gets ready. Mojca
![](https://secure.gravatar.com/avatar/6114858d84b0f40cf715f12237ab67a1.jpg?s=120&d=mm&r=g)
Hi everyone, I didn't have any luck building XeTeX on i386-linux (too old compiler) and sparc-solaris (some very weird error), but I hope that builds have some chance to work properly under "more normal circumstances". I'm not sure if cross-compiling XeTeX on Mac OS X works either. You are all invited to do "svn up" on http://svn.contextgarden.net/suite/build-binaries/ and run the builds. Please report any problems (failures to build XeTeX itself should be either addressed to Khaled or XeTeX mailing list or this list, but not to me personally). Thanks to everyone, Mojca PS: One problem that I found so far is that make needs to be a GNU make, else you need to set "GNUMAKE=gmake" or freetype2 building would fail. I believe that Khaled will fix that soon. On Tue, Mar 12, 2013 at 9:07 AM, Mojca Miklavec wrote:
Dear builders,
I'm soon going to add XeTeX back to the list of sources/binaries to faciliate testing of some major changes. The binaries will be uploaded to TLContrib. I need to adapt the scripts first and will let you know as soon as it gets ready.
Mojca
![](https://secure.gravatar.com/avatar/d0e94fb08d4620cfbd2f8f59a7868637.jpg?s=120&d=mm&r=g)
On 03/12/2013 01:46 PM, Mojca Miklavec wrote:
Hi everyone,
I didn't have any luck building XeTeX on i386-linux (too old compiler) and sparc-solaris (some very weird error), but I hope that builds have some chance to work properly under "more normal circumstances". I'm not sure if cross-compiling XeTeX on Mac OS X works either.
You are all invited to do "svn up" on http://svn.contextgarden.net/suite/build-binaries/ and run the builds. Please report any problems (failures to build XeTeX itself should be either addressed to Khaled or XeTeX mailing list or this list, but not to me personally).
Thanks to everyone, Mojca
Hi Mojca, looks good an powerpc-linux: Successfully built binaries: metapost: (skipped) xetex: ok luatex: (skipped) luatex_trunk: (skipped) luajittex: ok (the skipped one were unchanged, I guess?) All best Thomas
![](https://secure.gravatar.com/avatar/6114858d84b0f40cf715f12237ab67a1.jpg?s=120&d=mm&r=g)
On Tue, Mar 12, 2013 at 2:31 PM, Thomas A. Schmitzwrote:
Hi Mojca,
looks good an powerpc-linux:
Oh, perfect! The first successful compilation ;) Thank you. I kept annoying Khaled for months to make XeTeX compile on Mac (but my OS is too new to commit binaries), but then I tried to compile for two other platforms and both failed.
Successfully built binaries: metapost: (skipped) xetex: ok luatex: (skipped) luatex_trunk: (skipped) luajittex: ok
(the skipped one were unchanged, I guess?)
Yes. MetaPost and LuaTeX have already been compiled, LuaTeX trunk is never compiled unless you explicitly ask for it. LuaJITTeX underwent a hotfix release and XeTeX is new. So that's expected. Mojca
![](https://secure.gravatar.com/avatar/235081d51aed94adaf6010662fc0c923.jpg?s=120&d=mm&r=g)
Mojca,
I'm compiling now. The i386 binaries for xetex and xdvipdfmx built and were submitted;
the x86_64 binaries are now building. More soon.
Dick
On Mar 12, 2013, at 6:47 AM, Mojca Miklavec
On Tue, Mar 12, 2013 at 2:31 PM, Thomas A. Schmitzwrote:
Hi Mojca,
looks good an powerpc-linux:
Oh, perfect! The first successful compilation ;)
Thank you.
![](https://secure.gravatar.com/avatar/235081d51aed94adaf6010662fc0c923.jpg?s=120&d=mm&r=g)
Mojca, Xetex and xdvipdfmx both compiled fine in i386, x86_64, powerpc, and universal-darwin. The binaries were submitted. luajittex compiled fine in x86_64, didn't compile in i386, and the system didn't even try to compile it in powerpc. All of this is nothing new. The x86_64 binary was submitted. Dick Koch
![](https://secure.gravatar.com/avatar/6114858d84b0f40cf715f12237ab67a1.jpg?s=120&d=mm&r=g)
On Tue, Mar 12, 2013 at 3:12 PM, Richard Koch wrote:
Mojca,
Xetex and xdvipdfmx both compiled fine in i386, x86_64, powerpc, and universal-darwin. The binaries were submitted.
Thank you. On the bright side, I realized that I forgot to add xelatex (I just added it now) and teckit_compile (which might need another forced build unless Khaled plans a new release soon; and honestly, I don't expect such a high demand for that binary anyway, so there's no need to recompile now). For everyone else: please make sure to do "svn up" before building (but you need to do that anyway).
luajittex compiled fine in x86_64, didn't compile in i386, and the system didn't even try to compile it in powerpc. All of this is nothing new. The x86_64 binary was submitted.
If it is known not to compile, I'll also exclude i386-darwin from the list until the problem is fixed, so that it won't even consume time. I noticed that i386-solaris contains just xdvipdfmx, but no xetex. What exactly went wrong? (Also, it's not clear to me why the whole process didn't return an error and fail to submit that one binary.) I now added darwin binaries to TLContrib. Let's see how many people will complain. To get the binaries to ConTeXt distribution I need to change a few other scripts. Mojca
![](https://secure.gravatar.com/avatar/386a116d129cdc597f2bb6210bbaff7a.jpg?s=120&d=mm&r=g)
On Tue, Mar 12, 2013 at 4:22 PM, Mojca Miklavec
On Tue, Mar 12, 2013 at 3:12 PM, Richard Koch wrote:
luajittex compiled fine in x86_64, didn't compile in i386, and the system didn't even try to compile it in powerpc. All of this is nothing new. The x86_64 binary was submitted.
If it is known not to compile, I'll also exclude i386-darwin from the list until the problem is fixed, so that it won't even consume time. http://luajit.org/install.html says that is ok on OSX 10.4+ x86 (32 bit)
-- luigi
![](https://secure.gravatar.com/avatar/6114858d84b0f40cf715f12237ab67a1.jpg?s=120&d=mm&r=g)
If it is known not to compile, I'll also exclude i386-darwin from the list until the problem is fixed, so that it won't even consume time. http://luajit.org/install.html says that is ok on OSX 10.4+ x86 (32 bit)
From what I recall I'm 90% sure that it's because the building scripts don't take user-defined CFLAGS/CXXFLAGS into account, so one half of
So we need to find out why it fails. libraries ends up as 32-bit and the other one as 64-bit and linking fails. (It's probably only luajittex that fails to build, not luajit itself.) I seem to remember that I successfully built 32-bit LuaJITTeX on Mac OS X (maybe I'm wrong though), but that required modifications in building scripts. Mojca
![](https://secure.gravatar.com/avatar/02a711d78d02355347d6cff73fda2572.jpg?s=120&d=mm&r=g)
On 2013–03–12 Mojca Miklavec wrote:
I noticed that i386-solaris contains just xdvipdfmx, but no xetex. What exactly went wrong? (Also, it's not clear to me why the whole process didn't return an error and fail to submit that one binary.)
I noticed that, too. I get the following error: gmake all-recursive gmake[5]: Entering directory `/home/marco/build-binaries/src/xetex/build/libs/graphite2' /usr/bin/bash ./config.status config.status: creating Makefile config.status: creating include/graphite2/Makefile config.status: creating config.h config.status: config.h is unchanged config.status: executing depfiles commands config.status: executing libtool commands gmake[5]: Leaving directory `/home/marco/build-binaries/src/xetex/build/libs/graphite2' gmake[5]: Entering directory `/home/marco/build-binaries/src/xetex/build/libs/graphite2' Making all in . gmake[6]: Entering directory `/home/marco/build-binaries/src/xetex/build/libs/graphite2' CXX graphite2-1.2.1/src/Bidi.o ../../../source/libs/graphite2/graphite2-1.2.1/src/Bidi.cpp:40:9: error: expected identifier before numeric constant ../../../source/libs/graphite2/graphite2-1.2.1/src/Bidi.cpp:40:9: error: expected ‘}’ before numeric constant ../../../source/libs/graphite2/graphite2-1.2.1/src/Bidi.cpp:40:9: error: expected unqualified-id before numeric constant ../../../source/libs/graphite2/graphite2-1.2.1/src/Bidi.cpp:55:1: error: expected declaration before ‘}’ token gmake[6]: *** [graphite2-1.2.1/src/Bidi.o] Error 1 gmake[6]: Leaving directory `/home/marco/build-binaries/src/xetex/build/libs/graphite2' gmake[5]: *** [all-recursive] Error 1 gmake[5]: Leaving directory `/home/marco/build-binaries/src/xetex/build/libs/graphite2' gmake[4]: *** [all] Error 2 gmake[4]: Leaving directory `/home/marco/build-binaries/src/xetex/build/libs/graphite2' gmake[3]: *** [../../libs/graphite2/include/graphite2/Font.h] Error 2 gmake[3]: Leaving directory `/home/marco/build-binaries/src/xetex/build/libs/harfbuzz' gmake[2]: *** [all-recursive] Error 1 gmake[2]: Leaving directory `/home/marco/build-binaries/src/xetex/build/libs/harfbuzz' gmake[1]: *** [all] Error 2 gmake[1]: Leaving directory `/home/marco/build-binaries/src/xetex/build/libs/harfbuzz' gmake: *** [../../libs/harfbuzz/include/hb.h] Error 2 Marco
![](https://secure.gravatar.com/avatar/02a711d78d02355347d6cff73fda2572.jpg?s=120&d=mm&r=g)
On 2013–03–12 Marco wrote:
I noticed that i386-solaris contains just xdvipdfmx, but no xetex. What exactly went wrong? (Also, it's not clear to me why the whole process didn't return an error and fail to submit that one binary.)
The build script should indeed fail if some parts cannot be compiled. Especially the summary in the end is cofusing. It lists XeTeX as successfully compiled and uploaded, which it's not. Even a recompilation fails, since the revision is already present (unless one deletes the bin directory).
../../../source/libs/graphite2/graphite2-1.2.1/src/Bidi.cpp:40:9: error: expected identifier before numeric constant ../../../source/libs/graphite2/graphite2-1.2.1/src/Bidi.cpp:40:9: error: expected ‘}’ before numeric constant ../../../source/libs/graphite2/graphite2-1.2.1/src/Bidi.cpp:40:9: error: expected unqualified-id before numeric constant ../../../source/libs/graphite2/graphite2-1.2.1/src/Bidi.cpp:55:1: error: expected declaration before ‘}’ token
The corresponding lines are: 27 #include "inc/Main.h" 28 #include "inc/Slot.h" 29 #include "inc/Segment.h" 30 31 using namespace graphite2; 32 33 enum DirCode { // Hungarian: dirc 34 Unk = -1, 35 N = 0, // other neutrals (default) - ON 36 L = 1, // left-to-right, strong - L 37 R = 2, // right-to-left, strong - R 38 AL = 3, // Arabic letter, right-to-left, strong, AR 39 EN = 4, // European number, left-to-right, weak - EN 40 ES = 5, // European separator, left-to-right, weak - ES 41 ET = 6, // European number terminator, left-to-right, weak - ET 42 AN = 7, // Arabic number, left-to-right, weak - AN 43 CS = 8, // Common number separator, left-to-right, weak - CS 44 WS = 9, // white space, neutral - WS 45 BN = 10, // boundary neutral - BN 46 47 LRO = 11, // LTR override 48 RLO = 12, // RTL override 49 LRE = 13, // LTR embedding 50 RLE = 14, // RTL embedding 51 PDF = 15, // pop directional format 52 NSM = 16, // non-space mark 53 54 ON = N 55 }; I looked and grepped through the code, but I fail to see why ES is turned into a number. Usually that's the case if there's a #define ES …, but that's not the case here. Any help appreciated. Marco
![](https://secure.gravatar.com/avatar/b34c604e9bcd368d6253dd0860af9fff.jpg?s=120&d=mm&r=g)
MM> Date: Tue, 12 Mar 2013 13:46:28 +0100
MM> From: Mojca Miklavec
![](https://secure.gravatar.com/avatar/6114858d84b0f40cf715f12237ab67a1.jpg?s=120&d=mm&r=g)
Just a quick note to everyone who hasn't built yet: please fetch the latest scripts and latest binaries. For everyone else here's the building summary: 1.) One trivial bug (but with serious consequences) and one minor problem have been discovered in XeTeX repository and fixed since my request to build the binaries: - [minor problem] gmake now gets exported and XeTeX can be built on systems where "make" is not GNU make - [bug] ConTeXt format generation fails because \XeTeXrevision had a weird value 2a.) For the sake of replacing xetex.platform in TeX Live, I forgot to include teckit_compile binary, but that's not really a showstopper. That has been fixed, but only armel-linux includes that binary so far. 2b.) I also forgot symlinks for xelatex, but that's easy to fix. 3.) [not fixed yet] The build doesn't return a non-zero value in case that xetex build fails. Thus we have several platforms with missing xetex binary: i386-linux, i386-solaris, x86_64-linux. I'm curious about the problems on i386-linux and x86_64-linux. 4a.) i386-solaris has a weird compiler issue. 4b.) sparc-solaris has a known compiler bug http://wiki.opencsw.org/porting-faq#toc7 4c.) please let us know what goes wrong on i386-linux/x86_64-linux 5.) I disabled building luajittex on i386-darwin until the sources get fixed. I uploaded some binaries to TLContrib, but they suffer from the bug with \XeTeXrevision, so that all ConTeXt formats fail and a new rebuild will be required. There are the following options: a) you run: svn up ./bin_update_src.sh ./bin_update_svn.sh ./bin_build.sh --force --engines=xetex (--engines=xetex is not needed, but --force is) ./bin_commit.sh b) you can change the version of XeTeX in bin/VERSIONS to any value and then run svn up ./do_all.sh (it should trick the scripts to think that the version has changed) c) wait for the next version (in particular that is true for platforms where xetex failed to build :) Mojca
![](https://secure.gravatar.com/avatar/ac3fbaf0a287c4cffbc9f18fa931586c.jpg?s=120&d=mm&r=g)
On Wed, 13 Mar 2013 01:36:41 +0100
Mojca Miklavec
3.) [not fixed yet] The build doesn't return a non-zero value in case that xetex build fails. Thus we have several platforms with missing xetex binary: i386-linux, i386-solaris, x86_64-linux. I'm curious about the problems on i386-linux and x86_64-linux.
make[1]: Leaving directory `/home/local/src/context/build-binaries/src/xetex/build/libs/poppler' CXX xetexdir/libxetex_a-XeTeXFontInst.o In file included from ../../../source/texk/web2c/xetexdir/XeTeXFontInst.h:51:0, from ../../../source/texk/web2c/xetexdir/XeTeXFontInst.cpp:42: ../../../source/texk/web2c/xetexdir/XeTeXFontMgr.h:41:35: fatal error: fontconfig/fontconfig.h: No such file or directory compilation terminated. make: *** [xetexdir/libxetex_a-XeTeXFontInst.o] Error 1 make: Leaving directory `/home/local/src/context/build-binaries/src/xetex/build/texk/web2c' Successfully built binaries: metapost: (skipped) xetex: ERROR luatex: (skipped) luatex_trunk: (skipped) luajittex: (skipped) There have been some failures: - build of xetex on x86_64-linux -- Alan Braslau CEA DSM-IRAMIS-SPEC CNRS URA 2464 Orme des Merisiers 91191 Gif-sur-Yvette cedex FRANCE tel: +33 1 69 08 73 15 fax: +33 1 69 08 87 86 mailto:alan.braslau@cea.fr
![](https://secure.gravatar.com/avatar/ac3fbaf0a287c4cffbc9f18fa931586c.jpg?s=120&d=mm&r=g)
On Wed, 13 Mar 2013 08:17:15 +0100
Alan BRASLAU
On Wed, 13 Mar 2013 01:36:41 +0100 Mojca Miklavec
wrote: 3.) [not fixed yet] The build doesn't return a non-zero value in case that xetex build fails. Thus we have several platforms with missing xetex binary: i386-linux, i386-solaris, x86_64-linux. I'm curious about the problems on i386-linux and x86_64-linux.
make[1]: Leaving directory `/home/local/src/context/build-binaries/src/xetex/build/libs/poppler' CXX xetexdir/libxetex_a-XeTeXFontInst.o In file included from ../../../source/texk/web2c/xetexdir/XeTeXFontInst.h:51:0, from ../../../source/texk/web2c/xetexdir/XeTeXFontInst.cpp:42: ../../../source/texk/web2c/xetexdir/XeTeXFontMgr.h:41:35: fatal error: fontconfig/fontconfig.h: No such file or directory compilation terminated. make: *** [xetexdir/libxetex_a-XeTeXFontInst.o] Error 1 make: Leaving directory `/home/local/src/context/build-binaries/src/xetex/build/texk/web2c'
Successfully built binaries: metapost: (skipped) xetex: ERROR luatex: (skipped) luatex_trunk: (skipped) luajittex: (skipped) There have been some failures: - build of xetex on x86_64-linux
So I have now installed the package libfontconfig1-dev recompiling... Alan
participants (7)
-
Alan BRASLAU
-
Boris Veytsman
-
luigi scarso
-
Marco
-
Mojca Miklavec
-
Richard Koch
-
Thomas A. Schmitz