On Wednesday, June 1, 2016 at 8:10:26 AM UTC+1, Dima Pasechnik wrote:
>
>
>
> On Wednesday, June 1, 2016 at 12:26:31 AM UTC+1, Volker Braun wrote:
>>
>> Yes, there is not much we can do about the case-sensitivity "feature" in 
>> OSX. Probably the best course of action would be to raise an error as soon 
>> as possible instead of attempting to build.
>>
>
> I just asked : 
> http://apple.stackexchange.com/questions/240245/checking-path-capitalization-of-pwd
>  
> <http://www.google.com/url?q=http%3A%2F%2Fapple.stackexchange.com%2Fquestions%2F240245%2Fchecking-path-capitalization-of-pwd&sa=D&sntz=1&usg=AFQjCNFMPlc6Ve1wyXmVLHbqO1_-DfsADA>
>
> We'll see, if there will be no reply, we'd work out ourselves a quick 
> check that we can do in the top makefile or some more appropriate way.
>

OK, so they suggest to use `/bin/pwd -P` as a way to get the correct 
capitalisation, and it does work.
So, perhaps adding comparison of this output with the one from `pwd` is a 
good idea, to insert somewhere
as a sanity check.


Sanitising all the Sage scripts to use this instead of plain pwd is perhaps 
too much trouble...

 
>
>>
>>
>> On Tuesday, May 31, 2016 at 9:05:37 PM UTC+2, Dima Pasechnik wrote:
>>>
>>> in more detail, build scrips routinely use `pwd` to figure out the 
>>> current directory. Now, watch this (on OSX):
>>>
>>> nash:~ dima$ cd desktop
>>>
>>> nash:desktop dima$ pwd
>>>
>>> /Users/dima/desktop
>>>
>>> nash:desktop dima$ cd    
>>>
>>> nash:~ dima$ cd Desktop 
>>>
>>> nash:Desktop dima$ pwd
>>> /Users/dima/Desktop 
>>>
>>> No wonder these "slightly wrong" paths get into configurations, and then 
>>> some tools, which are case-sensitive, choke on them...
>>>
>>>
>>> On Tuesday, May 31, 2016 at 7:48:16 PM UTC+1, Dima Pasechnik wrote:
>>>>
>>>>
>>>>
>>>> On Tuesday, May 31, 2016 at 7:37:09 PM UTC+1, paulmasson wrote:
>>>>>
>>>>> I have a partial answer as to what is causing these errors. Before 
>>>>> starting 'make' I switched to the Sage directory using 'cd desktop/sage', 
>>>>> as you can see by the small 'd' in 'desktop'. On a Mac that directory is 
>>>>> most properly 'Desktop' with a capital 'D'. I revisited the build that 
>>>>> halted by switching to the Sage directory with 'cd Desktop/sage' and 
>>>>> again 
>>>>> running 'make', and now everything copies and builds just fine.
>>>>>
>>>>
>>>> oy gevalt :-( Well, OSX case-insensitive file systems (legacy of OSY 
>>>> with Y<10) lend you a very helping hand to shoot yourself in the foot.
>>>>
>>>>>
>>>>> That leaves open the question of why the build process is allowed to 
>>>>> mostly work with the slightly wrong directory name, and how to flag the 
>>>>> error before building starts.
>>>>>
>>>>
>>>> well, it's because of inability to distinguish right from wrong, 
>>>> apparently, due to oh so helpful case-insensitivity "feature".
>>>>
>>>>>
>>>>>
>>>>> On Sunday, May 29, 2016 at 3:02:31 PM UTC-7, paulmasson wrote:
>>>>>>
>>>>>> Turns out this issue has nothing to do with the Github copy. I 
>>>>>> downloaded a 7.2 tarball and had the same issue a total of nine times. 
>>>>>> Here 
>>>>>> are the successive error messages:
>>>>>>
>>>>>> /Users/Masson/desktop/sage/src/build/cythonized/sage/combinat/partitions.cpp:323:26:
>>>>>>  
>>>>>> fatal error: partitions_c.h: No such file or directory
>>>>>>
>>>>>>  #include "partitions_c.h"
>>>>>>
>>>>>> /Users/Masson/desktop/sage/src/build/cythonized/sage/combinat/matrices/dancing_links.cpp:319:29:
>>>>>>  
>>>>>> fatal error: dancing_links_c.h: No such file or directory
>>>>>>
>>>>>>  #include "dancing_links_c.h"
>>>>>>
>>>>>> /Users/Masson/desktop/sage/src/build/cythonized/sage/graphs/base/boost_graph.cpp:314:31:
>>>>>>  
>>>>>> fatal error: boost_interface.cpp: No such file or directory
>>>>>>
>>>>>>  #include "boost_interface.cpp"
>>>>>>
>>>>>> /Users/Masson/desktop/sage/src/build/cythonized/sage/libs/ecl.c:324:20: 
>>>>>> fatal error: eclsig.h: No such file or directory
>>>>>>
>>>>>>  #include "eclsig.h"
>>>>>>
>>>>>> /Users/Masson/desktop/sage/src/build/cythonized/sage/libs/ppl.cpp:336:23:
>>>>>>  
>>>>>> fatal error: ppl_shim.hh: No such file or directory
>>>>>>
>>>>>>  #include "ppl_shim.hh"
>>>>>>
>>>>>> /Users/Masson/desktop/sage/src/build/cythonized/sage/libs/eclib/mwrank.cpp:356:20:
>>>>>>  
>>>>>> fatal error: wrap.cpp: No such file or directory
>>>>>>
>>>>>>  #include "wrap.cpp"
>>>>>>
>>>>>> /Users/Masson/desktop/sage/src/build/cythonized/sage/matroids/matroid.c:296:22:
>>>>>>  
>>>>>> fatal error: minorfix.h: No such file or directory
>>>>>>
>>>>>>  #include "minorfix.h"
>>>>>>
>>>>>> /Users/Masson/desktop/sage/src/build/cythonized/sage/misc/darwin_utilities.c:290:33:
>>>>>>  
>>>>>> fatal error: darwin_memory_usage.h: No such file or directory
>>>>>>
>>>>>>  #include "darwin_memory_usage.h"
>>>>>>
>>>>>> /Users/Masson/desktop/sage/src/build/cythonized/sage/symbolic/pynac.cpp:334:22:
>>>>>>  
>>>>>> fatal error: pynac_cc.h: No such file or directory
>>>>>>
>>>>>>  #include "pynac_cc.h"
>>>>>>
>>>>>> After each failed build I manually copied each missing file to the 
>>>>>> appropriate directory and ran 'make' again. I now have a running build 
>>>>>> of 
>>>>>> 7.2. Yippee!
>>>>>>
>>>>>> A friend has suggested that the errors might be caused by bad line 
>>>>>> endings that only show up on a Mac. Any other explanations for such 
>>>>>> apparently random behavior?
>>>>>>
>>>>>>
>>>>>> On Saturday, May 28, 2016 at 5:55:04 PM UTC-7, paulmasson wrote:
>>>>>>>
>>>>>>> Running OS X El Capitan on a year-old Macbook.
>>>>>>>
>>>>>>> Deleted my entire Sage directory. Did a clean clone from 
>>>>>>> github.com/sagemath/sage.git. Immediately changed to Sage directory 
>>>>>>> and ran make. Same error as previously:
>>>>>>>
>>>>>>> error: command 'gcc' failed with exit status 1
>>>>>>>
>>>>>>> /Users/Masson/downloads/github/sage/src/build/cythonized/sage/combinat/partitions.cpp:323:26:
>>>>>>>  
>>>>>>> fatal error: partitions_c.h: No such file or directory
>>>>>>>
>>>>>>>  #include "partitions_c.h"
>>>>>>>
>>>>>>>                           ^
>>>>>>>
>>>>>>> compilation terminated.
>>>>>>>
>>>>>>> make[3]: *** [sage] Error 1
>>>>>>>
>>>>>>>
>>>>>>> Is there a problem building from the github copy?
>>>>>>>
>>>>>>

-- 
You received this message because you are subscribed to the Google Groups 
"sage-support" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-support+unsubscr...@googlegroups.com.
To post to this group, send email to sage-support@googlegroups.com.
Visit this group at https://groups.google.com/group/sage-support.
For more options, visit https://groups.google.com/d/optout.

Reply via email to