For the record:

ImportError: /usr/lib/x86_64-linux-gnu/libgsl.so.23: undefined symbol: 
cblas_ctrmv


Am Mittwoch, 15. Januar 2020 14:16:39 UTC+1 schrieb Martin R:
>
> I'm afraid I was celebrating too early.  make build succeeded, but sage 
> doesn't start anymore now.  I guess I should start from scratch. :-(
>
> Am Mittwoch, 15. Januar 2020 14:07:49 UTC+1 schrieb Martin R:
>>
>> BINGO!!!
>>
>> Many many many thanks!
>>
>> Martin
>>
>> Am Mittwoch, 15. Januar 2020 14:00:29 UTC+1 schrieb Dima Pasechnik:
>>>
>>> Note that you need to uninstall bzip2, xz, pcre, readline, curl before 
>>> the corresponding system 
>>> versions are picked up by ./configure (of course all this is not 
>>> needed for fresh builds...) 
>>>
>>> if this does not help please post config.log 
>>>
>>> On Wed, Jan 15, 2020 at 12:53 PM 'Martin R' via sage-release 
>>> <[email protected]> wrote: 
>>> > 
>>> > 
>>> > 
>>> > Am Mittwoch, 15. Januar 2020 13:23:17 UTC+1 schrieb Dima Pasechnik: 
>>> >> 
>>> >> On Wed, Jan 15, 2020 at 12:13 PM 'Martin R' via sage-release 
>>> >> <[email protected]> wrote: 
>>> >> > 
>>> >> > Oh dea, this is becoming strange: 
>>> >> > 
>>> >> > The output of configure says that 
>>> >> > 
>>> >> > atlas, openblas, iconv will NOT be installed, but 
>>> >> > 
>>> >> > bzip2-1.0.6-20150304.p0 
>>> >> > readline-6.3.008.p0 
>>> >> > xz-5.2.2.p0 
>>> >> > pcre-8.40.p2 
>>> >> > curl-7.62.0.p0 
>>> >> > 
>>> >> > However, I certainly have bzip2, xz, pcre and curl on my system, 
>>> and I guess also readline. 
>>> >> > 
>>> >> > Might there be a bug in configure? 
>>> >> 
>>> >> You need development versions of these, see 
>>> >> 
>>> https://doc.sagemath.org/html/en/installation/source.html#linux-recommended-installation
>>>  
>>> >> 
>>> >> So, on Ubuntu: 
>>> >> 
>>> >> libbz2-dev 
>>> >> libreadline-dev (seems to be missing in docs) 
>>> >> liblzma-dev (this is the xz lib, iirc) 
>>> >> libpcre3-dev 
>>> >> libcurl4-openssl-dev 
>>> > 
>>> > 
>>> >> 
>>> >> the last of these, libcurl4-openssl-dev was indeed missing, but 
>>> configure still wants to build r :-( 
>>> > 
>>> > 
>>> > Martin 
>>> >> 
>>> >> 
>>> >> 
>>> >> > 
>>> >> > Martin 
>>> >> > 
>>> >> > Am Mittwoch, 15. Januar 2020 13:00:51 UTC+1 schrieb Dima Pasechnik: 
>>> >> >> 
>>> >> >> On Wed, Jan 15, 2020 at 11:35 AM 'Martin R' via sage-release 
>>> >> >> <[email protected]> wrote: 
>>> >> >> > 
>>> >> >> > Am Mittwoch, 15. Januar 2020 12:02:30 UTC+1 schrieb Dima 
>>> Pasechnik: 
>>> >> >> >> 
>>> >> >> >> On Wed, Jan 15, 2020 at 10:37 AM 'Martin R' via sage-release 
>>> >> >> >> <[email protected]> wrote: 
>>> >> >> >> > 
>>> >> >> >> > Thanks again!  It still tries to build R and fails 
>>> >> >> >> > 
>>> >> >> >> > I checked that system R starts, it does (R version 3.4.4) 
>>> >> >> >> Please also make sure that 
>>> >> >> >> 
>>> >> >> >> pkg-config --modversion libR 
>>> >> >> >> 
>>> >> >> >> outputs the version. 
>>> >> >> >> 
>>> >> >> > Yes, it does. 
>>> >> >> > 
>>> >> >> >> 
>>> >> >> >> However there is just no "checking for R" message in your 
>>> config.log, 
>>> >> >> >> it indicates that there are still remains of R left 
>>> >> >> >> in your installation. 
>>> >> >> >> 
>>> >> >> >> I've opened https://trac.sagemath.org/ticket/29016 
>>> >> >> >> to deal with R being hard to clean away - feel free to try it. 
>>> >> >> >> 
>>> >> >> >> Maybe you also need to run 
>>> >> >> >> 
>>> >> >> >> make r-clean 
>>> >> >> > 
>>> >> >> > 
>>> >> >> > I did that, it says that there is no legacy uninstaller for r, 
>>> and that there is nothing to do.  I checked for the files in the ticket, 
>>> but none of them exists. :-( 
>>> >> >> > 
>>> >> >> > The config.log contains the lines 
>>> >> >> > 
>>> >> >> > configure:18950: checking installing any of atlas openblas iconv 
>>> readline bzip2 xz pcre curl? 
>>> >> >> > configure:18954: result: yes; install r as well 
>>> >> >> > 
>>> >> >> > but I do not know how to find out which of those (atlas openblas 
>>> iconv readline bzip2 xz pcre curl) is the problem. 
>>> >> >> 
>>> >> >> The easiest is to look at the output of ./configure 
>>> >> >> It should contain lines like 
>>> >> >> 
>>> >> >> atlas-3.10.2.p3 will not be installed (configure check) 
>>> >> >> 
>>> >> >> for each of the packages in that list. 
>>> >> >> So for any of these that are not marked this way, please install 
>>> the 
>>> >> >> corresponding system package 
>>> >> >> and re-run configure. 
>>> >> >> 
>>> >> >> 
>>> >> >> 
>>> >> >> > 
>>> >> >> > Martin 
>>> >> >> > 
>>> >> >> > -- 
>>> >> >> > You received this message because you are subscribed to the 
>>> Google Groups "sage-release" group. 
>>> >> >> > To unsubscribe from this group and stop receiving emails from 
>>> it, send an email to [email protected]. 
>>> >> >> > To view this discussion on the web visit 
>>> https://groups.google.com/d/msgid/sage-release/5d105000-cea2-48a5-8840-aeb405a6cb71%40googlegroups.com.
>>>  
>>>
>>> >> > 
>>> >> > -- 
>>> >> > You received this message because you are subscribed to the Google 
>>> Groups "sage-release" group. 
>>> >> > To unsubscribe from this group and stop receiving emails from it, 
>>> send an email to [email protected]. 
>>> >> > To view this discussion on the web visit 
>>> https://groups.google.com/d/msgid/sage-release/ca07b5bd-8fe6-4383-ad3a-eab749b3760e%40googlegroups.com.
>>>  
>>>
>>> > 
>>> > -- 
>>> > You received this message because you are subscribed to the Google 
>>> Groups "sage-release" group. 
>>> > To unsubscribe from this group and stop receiving emails from it, send 
>>> an email to [email protected]. 
>>> > To view this discussion on the web visit 
>>> https://groups.google.com/d/msgid/sage-release/a2512246-f441-4ace-82df-fd1f354bcaf5%40googlegroups.com.
>>>  
>>>
>>>
>>

-- 
You received this message because you are subscribed to the Google Groups 
"sage-release" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-release/1901644c-892d-4151-8b52-4845029e8753%40googlegroups.com.

Reply via email to