Bug#879801: ftbfs with icu from experimental

2017-12-10 Thread Norbert Preining
> Not sure if it is useful: I've replaced "rm -f" by "rm -fv". Now the > build runs OK and reports which files are really found (and removed). Very useful! I think I have removed now all useless rm statements (the fmtutil.cnf/updmap.cfg and the long with ep2all etc ... interesting). Thanks a lot

Bug#879801: ftbfs with icu from experimental

2017-12-10 Thread Norbert Preining
Hi Hilmar, thanks for testing. > (new log attached). I've tried to document all not existing files in my > debian/rules file. Maybe I've overdone it. Yes, that is the problem. Look for example into the first patch @@ -168,14 +168,14 @@ # remove various man pages for i in latex pd

Bug#879801: ftbfs with icu from experimental

2017-12-10 Thread Norbert Preining
> Please have a look at the patch, look for the removed (and inserted) > lines. Build log is attached. Ah and this is BTW not correct. Because files like latex.1 and pdflatex.1 are actually installed. Only fmtutil(-sys).1 etc are not. That was my question, which of the files in the for loop. Th

Bug#879801: ftbfs with icu from experimental

2017-12-10 Thread Norbert Preining
Very interesting ... and very surprising ... it seems that the forced reautoconf in all directories *did* change something ... We need to compare the actual file lists in -7 and -8 before we can do these kind of things ... Ahh ... wait ... I guess I have an idea ... there is the patch debian/p

Bug#879801: ftbfs with icu from experimental

2017-12-09 Thread Hilmar Preuße
On 09.12.2017 16:14, Norbert Preining wrote: Hilmar >> installation. I can reproduce the problem on both systems. > > Which files are tried to be removed but are not there. Not all of them I > guess, right? > Please have a look at the patch, look for the removed (and inserted) lines. Build log

Bug#879801: ftbfs with icu from experimental

2017-12-09 Thread Norbert Preining
> installation. I can reproduce the problem on both systems. Which files are tried to be removed but are not there. Not all of them I guess, right? Norbert -- PREINING Norbert http://www.preining.info Accelia Inc. +JAIST +TeX Live +Debian Dev

Bug#879801: ftbfs with icu from experimental

2017-12-09 Thread Hilmar Preuße
On 09.12.2017 13:52, Norbert Preining wrote: Hi Norbert, >> were you able to build the package? For me it failed, b/c some rm commands >> in the install stage failed. For now I replaced rm by rm -f to ignore the > > Strange. Before uploading I will check, but rm -f is not the solution. > I want

Bug#879801: ftbfs with icu from experimental

2017-12-09 Thread Norbert Preining
Hi Hilmar, > were you able to build the package? For me it failed, b/c some rm commands > in the install stage failed. For now I replaced rm by rm -f to ignore the Strange. Before uploading I will check, but rm -f is not the solution. I want that what is written in debian/rules is actually carrie

Bug#879801: ftbfs with icu from experimental

2017-12-08 Thread Hilmar Preuße
Am 07.12.2017 um 15:42 teilte Norbert Preining mit: Hi Norbert, were you able to build the package? For me it failed, b/c some rm commands in the install stage failed. For now I replaced rm by rm -f to ignore the exit code (patch attached). Hilmar Found the bug ... dh_autoreconf does not r

Bug#879801: ftbfs with icu from experimental

2017-12-07 Thread Norbert Preining
> Hmm, I'd expect that dh_autoreconf does its rebuild recursively. I did expect it, too, but it seems it doesn't. Looking at the reautoconf script distributed with TL it calls with --no-recursive but gives a long list of directories. Probably because the autoconf code is not setting the "subdirs"

Bug#879801: ftbfs with icu from experimental

2017-12-07 Thread Hilmar Preuße
Am 07.12.2017 um 15:42 teilte Norbert Preining mit: Hi, Found the bug ... dh_autoreconf does not rebuild the files down in texk/web2c, working on a fix now by adding all dirs to debian/autoreconf Norbert Hmm, I'd expect that dh_autoreconf does its rebuild recursively. NAME dh_autore

Bug#879801: ftbfs with icu from experimental

2017-12-07 Thread Norbert Preining
Found the bug ... dh_autoreconf does not rebuild the files down in texk/web2c, working on a fix now by adding all dirs to debian/autoreconf Norbert On December 7, 2017 9:17:22 AM GMT+08:00, Norbert Preining wrote: >Hi Hilmar, > >thanks for testing! > >> XeTeX in the end fails due to missing sy

Bug#879801: ftbfs with icu from experimental

2017-12-06 Thread Norbert Preining
Hi Hilmar, thanks for testing! > XeTeX in the end fails due to missing symbols (last part of build log is I don't see any libicu* in the linker run, though? > hille@amd64-sid:~/...$ pkg-config --libs icu-uc > -licuuc -licudata I guess for security I would use pkg-config --libs icu-uc icu-io

Bug#879801: ftbfs with icu from experimental

2017-12-06 Thread Hilmar Preuße
On 04.12.2017 15:17, Norbert Preining wrote: Hi Norbert, > Trying to mimic this with pkg-config seems to be a bit tricky. > > I am not sure whether replacement of > icu-config --cppflags > with > pkg-config --cflags icu-uc icu-io > and > icu-config --ldflags-searchpath --ldflags-libsonly

Bug#879801: ftbfs with icu from experimental

2017-12-04 Thread Hilmar Preuße
On 04.12.2017 16:25, Matthias Klose wrote: > On 04.12.2017 15:17, Norbert Preining wrote: Hi, >> Just one last idea ... icu-config is *still* shipped with upstream. >> It would be nice if either you or upstream provides a concise guide how >> to transition to pkg-config, instead of unilaterally

Bug#879801: ftbfs with icu from experimental

2017-12-04 Thread Matthias Klose
On 04.12.2017 15:17, Norbert Preining wrote: > HI all, > > I have now installed icu60 in a chroot and try to find what are the > equivalent calls to pkg-config. > >>> [1] http://userguide.icu-project.org/howtouseicu > > Interestingly, and I am far from sure that this is correct, is that > pkg-co

Bug#879801: ftbfs with icu from experimental

2017-12-04 Thread Norbert Preining
HI all, I have now installed icu60 in a chroot and try to find what are the equivalent calls to pkg-config. > > [1] http://userguide.icu-project.org/howtouseicu Interestingly, and I am far from sure that this is correct, is that pkg-config returns *nothing* on --cflags $ pkg-config --cflags ic

Bug#879801: ftbfs with icu from experimental

2017-12-04 Thread Norbert Preining
> Could you address this problem in the meantime @TL upstream? AFAICT there I did it already on the internal list, without much activity but the depressing outlook on c++11 imminent requirement :-( I am not checking it myself, shouldn't be too complicated. Is there a guide *how* icu-config calls

Bug#879801: ftbfs with icu from experimental

2017-12-04 Thread GCS
Hi Hilmar, On Sun, Dec 3, 2017 at 10:24 PM, Hilmar Preuße wrote: > Am 26.10.2017 um 09:42 teilte Matthias Klose mit: >> icu from experimental dropped the icu-config binary, and texlive-bin >> doesn't have a fallback for pkg-config icu-i18n. >> > changelog of icu: [...] > We now that icu-config is

Bug#879801: ftbfs with icu from experimental

2017-12-04 Thread Hilmar Preuße
Am 26.10.2017 um 09:42 teilte Matthias Klose mit: Hi Norbert, icu from experimental dropped the icu-config binary, and texlive-bin doesn't have a fallback for pkg-config icu-i18n. Could you address this problem in the meantime @TL upstream? AFAICT there wasn't any change in TL SVN for this is

Bug#879801: ftbfs with icu from experimental

2017-12-03 Thread Hilmar Preuße
Am 26.10.2017 um 09:42 teilte Matthias Klose mit: Hi Laszlo, icu from experimental dropped the icu-config binary, and texlive-bin doesn't have a fallback for pkg-config icu-i18n. changelog of icu: icu (59.1-1) experimental; urgency=low * New major upstream release. * Remove icu-config f

Bug#879801: ftbfs with icu from experimental

2017-10-26 Thread Matthias Klose
Package: src:texlive-bin Version: 2017.20170613.44572-6 Severity: important Tags: sid buster icu from experimental dropped the icu-config binary, and texlive-bin doesn't have a fallback for pkg-config icu-i18n.