Julia installation on Mac Pro fails - julia

I am using Mac Pro with El Capitan. I tried installing Julia through homebrew but it fails with the following stack trace
==> make FC=/usr/local/bin/gfortran
Last 15 lines from /Users/vishalsharma/Library/Logs/Homebrew/openblas-julia/01.make:
printf("#define DLOCAL_BUFFER_SIZE\t%ld\n", (DGEMM_DEFAULT_Q * DGEMM_DEFAULT_UNROLL_N * 2 * 1 * sizeof(double)));
^
getarch_2nd.c:69:50: error: use of undeclared identifier 'CGEMM_DEFAULT_Q'
printf("#define CLOCAL_BUFFER_SIZE\t%ld\n", (CGEMM_DEFAULT_Q * CGEMM_DEFAULT_UNROLL_N * 4 * 2 * sizeof(float)));
^
getarch_2nd.c:69:68: error: use of undeclared identifier 'CGEMM_DEFAULT_UNROLL_N'
printf("#define CLOCAL_BUFFER_SIZE\t%ld\n", (CGEMM_DEFAULT_Q * CGEMM_DEFAULT_UNROLL_N * 4 * 2 * sizeof(float)));
^
getarch_2nd.c:70:50: error: use of undeclared identifier 'ZGEMM_DEFAULT_Q'
printf("#define ZLOCAL_BUFFER_SIZE\t%ld\n", (ZGEMM_DEFAULT_Q * ZGEMM_DEFAULT_UNROLL_N * 2 * 2 * sizeof(double)));
^
fatal error: too many errors emitted, stopping now [-ferror-limit=]
20 errors generated.
make: *** [getarch_2nd] Error 1
Makefile:131: *** OpenBLAS: Detecting CPU failed. Please set TARGET explicitly, e.g. make TARGET=your_cpu_target. Please read README for the detail.. Stop.

It looks like OpenBLAS can't detect your CPU architecture. Do you have a Nehalem CPU? You'll need to create a Make.inc file with OPENBLAS_TARGET_ARCH=NEHALEM inside it, and run make TARGET=NEHALEM FC=... or similar (adjust NEHALEM to HASWELL or your arch as appropriate). See the recent comments on this GitHub issue.

Related

New 2020 GNAT Ada installation failing to build an old 2015 project

I recently installed GNAT Ada (2020) and GNAVI GWindows on a new PC.
On trying to build an old project developed under the 2015 version, I get this:
No candidate interpretations match the actuals:
Too many arguments in call to "Put"
expected private type "Printer_Canvas_Type" defined at gwindows-drawing.ads:603
found private type "Canvas_Type" defined at gwindows-drawing.ads:96
This is one of the several similar pieces of code that produce the same result (MapCanvas is declared elsewhere as Canvas_Type):
Put (MapCanvas, (DATUM_BASE_X +
(x * 10 * GRID_MONAD_SPACING)) +
(5 * GRID_MONAD_SPACING),
y - (GRID_MONAD_SPACING + 5),
Integer'Image (x));
Possibly related, I also get this in regard to the last argument in the call to Put - Integer'Image(x)
expected type "Standard.Wide_String"
found type"Standard.String"
As an experiment, I tried converting the last argument using To_Wide_String (Integer'Image (x)) but the result was the same.
Elsewhere, similar code with a literal compiles ok:
Put (MapCanvas, (DATUM_BASE_X - 1 +
(GRID_MONAD_SPACING / 2) +
(x * 10 * GRID_MONAD_SPACING)),
y + 20 + (60 * GRID_MONAD_SPACING),
"0 2 4 6 8 ");
On the previous PC with GNAT Ada 2015, everything compiled. I've compared the declarations
of Put in gwindows-drawing.ads in the old and new installations of GWindows and they are identical.
Unfortunately I can no longer build on the old PC without a lot of backtracking - the old project drive
is in use elsewhere, though I do have it all backed up.
I'd appreciate any ideas on where to look for the cause of the problem.
GWindows has two string modes matching Windows API: ANSI (8-bit character) and Unicode.
The type GString is either a String or a Wide_String.
You can switch the framework's mode with ansi.cmd and unicode.cmd .
Obviously your old project was in ANSI mode.

how to use a brew formula from github?

I got the following error for https://github.com/sorin-ionescu/homebrew-personal/blob/master/sqlite.rb
Is this because this formula is too old and does not support my OS version?
$ brew install sorin-ionescu/homebrew-personal/sqlite
==> Tapping sorin-ionescu/personal
Cloning into '/usr/local/Homebrew/Library/Taps/sorin-ionescu/homebrew-personal'...
remote: Enumerating objects: 19, done.
remote: Counting objects: 100% (19/19), done.
remote: Compressing objects: 100% (19/19), done.
remote: Total 19 (delta 1), reused 6 (delta 0), pack-reused 0
Unpacking objects: 100% (19/19), done.
Error: Invalid formula: /usr/local/Homebrew/Library/Taps/sorin-ionescu/homebrew-personal/dvtm.rb
dvtm: undefined method `sha1' for #<BottleSpecification:0x00007f8d5a983ef8>
Error: Invalid formula: /usr/local/Homebrew/Library/Taps/sorin-ionescu/homebrew-personal/osxfuse.rb
osxfuse: unknown version :snow_leopard
Error: Invalid formula: /usr/local/Homebrew/Library/Taps/sorin-ionescu/homebrew-personal/httpie.rb
httpie: Unsupported special dependency :python
Error: Invalid formula: /usr/local/Homebrew/Library/Taps/sorin-ionescu/homebrew-personal/minidlna.rb
minidlna: undefined method `sha1' for #<Resource::PatchResource:0x00007f8d5a0b4250>
Error: Cannot tap sorin-ionescu/personal: invalid syntax in tap!
This very custom tap is outdated. The command you used is totally right.

Wrong directory for sipcfg.sip_module_dir

By trying to compile QGIS from sources on Ubuntu, there is the file /opt/QGIS/cmake/FindSIP.py which content is:
import sipconfig
sipcfg = sipconfig.Configuration()
print("sip_version:%06.0x" % sipcfg.sip_version)
print("sip_version_num:%d" % sipcfg.sip_version)
print("sip_version_str:%s" % sipcfg.sip_version_str)
print("sip_bin:%s" % sipcfg.sip_bin)
print("default_sip_dir:%s" % sipcfg.default_sip_dir)
print("sip_inc_dir:%s" % sipcfg.sip_inc_dir)
# SIP 4.19.10+ has new sipcfg.sip_module_dir
if hasattr(sipcfg, "sip_module_dir"):
print("sip_module_dir:%s" % sipcfg.sip_module_dir)
else:
print("sip_module_dir:%s" % sipcfg.sip_mod_dir)
In Python3.6, the last if/else statement prints:
sip_module_dir:/usr/lib/python3.6/dist-packages
But the string /usr/lib/python3.6/dist-packages doesn't match an existing directory (it is probably part of why I encountered the error:
python/CMakeFiles/python_module_qgis__core.dir/build.make:537: recipe for target 'python/core/sip_corepart0.cpp' failed' failed during build process).
I either have:
/usr/lib/python3/dist-packages
or
/usr/lib/python3.6/site-packages
And it's only in /usr/lib/python3/dist-packages that I have some 'sip' related files (the other directory gives no results):
$ find . -iname "*sip*"
./twisted/protocols/__pycache__/sip.cpython-36.pyc
./twisted/protocols/sip.py
./twisted/test/test_sip.py
./twisted/test/__pycache__/test_sip.cpython-36.pyc
./sipconfig.py
./sip.pyi
./sipconfig_nd6.py
./sipdistutils.py
./__pycache__/sipconfig_nd6.cpython-36.pyc
./__pycache__/sipdistutils.cpython-36.pyc
./__pycache__/sipconfig.cpython-36.pyc
./sip.cpython-36m-x86_64-linux-gnu.so
I guess there is something to fix within 'sip' itself but don't know where exactly, neither how to do that.
More information;
OS: Ubuntu 16.04 64 bits
Python: 3.6.7
Sip: 4.19.7

Syntax error caused by using different version of IDL

I am new in IDL. When I use IDL version 8, my program run successfully. But when I use IDL Version 7.1 I get a syntax error that you can see below.
How can I fix this issue?
IDL> .com schwmodfirstiter_box.pro
str +=pars[i].name+strn(p[i],format='('+pars[i].sformat+')')
^
% Syntax error.
At: /home/behzad/schwmodfirstiter_box.pro, Line 5
% 1 Compilation error(s) in module SCHMAKEPARSTRING.
str +=pars[i].name+strn(p[i],format='('+pars[i].sformat+')')
^
% Syntax error.
At: /home/behzad/schwmodfirstiter_box.pro, Line 14
% 1 Compilation error(s) in module SCHMAKEPARSTRING0.
mlstr=pars[0].name+strn(ml[k],format='('+pars[0].sformat+')')+'/'
^
% Syntax error.
At: /home/behzad/schwmodfirstiter_box.pro, Line 63
% 1 Compilation error(s) in module SCHMAKEMODEL.
% Compiled module: SUBMITORBLIB.
% Compiled module: COMPILENNLS.
% Compiled module: SCHWMODFIRSTITER.
IDL>
% Interrupt encountered.
strn.pro is not part of IDL but is likely from
https://idlastro.gsfc.nasa.gov/ftp/pro/misc/strn.pro
Make sure you have this procedure in your !Path and can compile it.

How do I troubleshoot an "Error #1090" trying to generate ASDoc?

I've got a reasonably large Flex project and I'm trying to generate ASDoc documentation for it. When I run asdoc, it gives me this error:
An unexpected error occurred.
Error #1090: XML parser failure: element is malformed.
(Location of error unknown)XSLT Error (javax.xml.transform.TransformerException): java.io.FileNotFoundException: C:\projects\projectname\asdoc\toplevel_classes.xml (The system cannot find the file specified)
I found a closed as could-not-reproduce bug in Jira about single-quotes in a double-quoted public field initialiser, but I don't have any of those. Any clues?
#Amarghosh prompted me to try running ASDoc individually on my classes, and it turns out that the error is due to a malformed ASDoc comment in my source:
/**
* #see Blah.blah
*
* Furbles the whatsit.
*/
needs to be:
/**
* Furbles the whatsit.
*
* #see Blah.blah
*/

Resources