Dimitry Naldayev wrote:
In that case, you cannot enter Russian either, because the system will treat it as unprintable characters. Please stop using uClibc until it is fixed.Declan Moriarty <[EMAIL PROTECTED]> writes:Recently, Somebody Somewhere wrote these wordsHello I have built uClibc HLFS version SVN-20060122 The build logs aviable here: http://linux.perm.ru/dimitry/hlfs/ Are there general recommendations about building software from blfs book on the top of the uClibc HLFS system? ie do I need modify build instruction from BLFS book ?Only when thing won't compile :-PAs I notice, HLFS (at least uClibc version) tend to do sed -e 's/^CFLAGS .*$/& -pie -fpie/' -i Makefile.inor something like this for practicaly every package (but there was a few exceptions as I remember)Do I need apply this to BLFS packages? Are there general rule for this?Also , you may have noticed the recent discussion of UTF-8.I probably miss this discusion ;-/ but I do not UTF-8 at present time.Usually I use ru_RU.KOI8-R but I can live with pure POSIX locale (hmm.... in last case it probably better build system without nls at all)
And just for a ru_RU.KOI8-R breakage testcase: attached is a bzipped POSIX tar archive containing a folder with Russian characters in its name. Attempt to list this archive: tar tf tmp.tar.bz2, in the ru_RU.KOI8-R locale. Glibc-based systems do this correctly. uClibc-based HLFS doesn't. uClibc-based HLFS with libiconv preloaded also works correctly, but I am not sure if this really fixes, not masks, the real problem.
-- Alexander E. Patrakov
tmp.tar.bz2
Description: Binary data
-- http://linuxfromscratch.org/mailman/listinfo/hlfs-dev FAQ: http://www.linuxfromscratch.org/faq/ Unsubscribe: See the above information page
