Re: [sage-devel] trouble building sage on mac

[Anne Schilling]

Thank you all for your help! Finally I got sage to install. There were 
permission issues with my /usr/local folder, which prevented brew from 
working correctly.

Anne

On Wednesday, July 1, 2020 at 5:55:42 AM UTC-7, Dima Pasechnik wrote:
>
> my working conjecture was that Anne had two conflicting installations 
> of freetype, for I cannot imagine Homebrew serving version 2.3 of 
> freetype. 
>
>
> On Wed, Jul 1, 2020 at 1:36 PM Zachary Scherr  > wrote: 
> > 
> > Also, you might also consider doing something like the following: 
> > 
> > > cd /usr 
> > > find -L . -name 'ftheader.h' 
> > 
> > you should see the following locations: 
> > 
> > ./local/include/freetype2/freetype/config/ftheader.h 
> > 
> ./local/var/homebrew/linked/freetype/include/freetype2/freetype/config/ftheader.h
>  
>
> > ./local/opt/freetype2/include/freetype2/freetype/config/ftheader.h 
> > ./local/opt/freetype/include/freetype2/freetype/config/ftheader.h 
> > 
> ./local/Cellar/freetype/2.10.2/include/freetype2/freetype/config/ftheader.h 
> > 
> > notice though that 
> "/usr/local/Cellar/freetype/2.10.2/include/freetype2/freetype/config/ftheader.h"
>  
> is the real file and the others should all be symlinks of this one. 
> > 
> > On Wednesday, July 1, 2020 at 8:26:58 AM UTC-4 Zachary Scherr wrote: 
> >> 
> >> Hi All, 
> >> 
> >>I just thought I would chime in on the homebrew side of things.  I 
> have freetype installed via homebrew on Catalina 10.15.5 and when I execute 
> >> 
> >> > pkg-config --modversion freetype2 
> >> 
> >> I get back 23.2.17.  Anne seemed to get back version 9.8.3, which means 
> that something is screwy with your freetype. 
> >> 
> >> Some things you could try: 
> >> 
> >> 1). If you have already installed freetype via homebrew then you can 
> try 
> >>  brew unlink freetype && brew link freetype 
> >> 
> >> 2). It's possible that "brew doctor" will pick up on the 
> inconsistencies and will give recommendations as to how to fix it. 
> >> 
> >> Your log file specifically mentioned that the system could not find 
> ftconfig.h.  With freetype correctly installed via homebrew you can do the 
> following: 
> >> 
> >> cd /usr/local/include 
> >> ls -l | grep freetype2 
> >> 
> >> and you should see that that folder is symlinked to 
> /usr/local/Cellar/freetype/2.10.2/include/freetype2 which is the folder 
> that actually contains all the freetype header files. 
> >> 
> >> Not sure how much you want to invest in fixing this, but my suggestion 
> would be to head to /usr/local and try to find ftheader.h, which is the 
> header file that defines where ftconfig.h is located.  If you find any 
> ftheader.h file that is not originating from 
> local/Cellar/freetype/2.10.2/include/freetype2/freetype/config/ftheader.h 
> then you should delete it. 
> >> 
> >> 
> >> 
> >> On Wednesday, July 1, 2020 at 2:53:05 AM UTC-4 dim...@gmail.com wrote: 
> >>> 
> >>> 
> >>> 
> >>> On Wed, 1 Jul 2020, 07:41 Anne Schilling,  
> wrote: 
>  
>  
>  
>  On Tuesday, June 30, 2020 at 12:22:47 AM UTC-7, Dima Pasechnik wrote: 
> > 
> > 
> > 
> > On Tue, 30 Jun 2020, 07:46 Anne Schilling,  
> wrote: 
> >> 
> >> Here is the new config.log file. It still does not build: 
> >> 
> >> https://www.math.ucdavis.edu/~anne/config.log 
> > 
> > 
> > do you get the same error while building pillow (missing header)? If 
> so, it looks as if you removed the old freetype dylib files, but left 
> headers there. 
>  
>  
>  Thank you, but what exactly am I supposed to delete? I delete the 
> folder /usr/local/Cellar/freetype 
> >>> 
> >>> 
> >>> you had to delete the installation of freetype in /usr/local/{lib, 
> include, lib/pkgconfig}. 
> >>> 
> >>> and after that, (re)install freetype in Homebrew. 
> >>> 
> >>> Deleting /usr/local/Cellar/freetype 
> >>> probably was not a good idea - that is where, I guess, you have a good 
> installation of freetype (which you should uninstall using the brew 
> commands, not just deleting from shell) 
> >>> 
> >>> But I don't know whether your Homebrew is not too old. Can you do 
> >>> 
> >>> brew install freetype 
> >>> 
> >>> (or is it freetype2) 
> >>> 
> >>> paying attention to the version it mentions? 
> >>> 
> >>> 
> >>> in any case, 
> >>> 
> >>> ./configure --with-freetype=system 
> >>> 
> >>> should force sage to build freetype on its own. 
> >>> 
> >>> 
>  
> > Also, did you source .homebrew-build-env 
> > before doing ./configure && make ? 
> > As you use libraries from Homebrew, this might be crucial. 
>  
>  
>  I did that and it still does not work. See 
>  
>  https://www.math.ucdavis.edu/~anne/config.log 
>  
>  Is it possible that this might be an issue 
>  
>  https://github.com/Homebrew/legacy-homebrew/issues/43471 
>  
>  
>  -- 
>  You received this message because you are subscribed to the Google 
> Groups "sage-devel" group. 
>  To unsubscribe from this

Re: [sage-devel] build errors on void linux

[Michael Orlitzky]

On 2020-07-01 09:23, Nicolo' Piazzalunga wrote:
> Thanks.
> I'm puzzled by 'does not support check for system package'/'no suitable 
> system package' in config.log

We haven't written the script to detect a system copy of ECL yet. We're
in the middle of upgrading to a new major version of it, and only after
that's done will it make sense to detect a system copy (which will have
to be brand new). See,

  * https://trac.sagemath.org/ticket/22191
  * https://trac.sagemath.org/ticket/29617


> E.g. why is it not finding boost, isl, and why is it building ecl (just 
> to list a few), all of which are installed as system packages?

I think Dima answered your question for these two, but in general, you
can check config.log to see what decision the ./configure script made
and why. If you need to dig further, the detection scripts are in

  build/pkgs//spkg-configure.m4

in the sage source tree.

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-devel/42743afb-3bd9-3981-d5f3-d11fa7173816%40orlitzky.com.

Re: [sage-devel] build errors on void linux

[Dima Pasechnik]

On Wed, Jul 1, 2020 at 2:23 PM Nicolo' Piazzalunga
 wrote:
>
> Thanks.
> I'm puzzled by 'does not support check for system package'/'no suitable
> system package' in config.log
> E.g. why is it not finding boost, isl, and why is it building ecl (just
> to list a few), all of which are installed as system packages?

boost and isl should be possible to use from the system (make sure the
headers are installed - on some Linux
distros they are separate packages)

What happened is that apparently you didn't have boost installed when
you ran ./configure 1st time,
which prompted Sage to build its own copy, and now we see

## -- ##
## Checking whether SageMath should install SPKG boost_cropped... ##
## -- ##
configure:12971: boost_cropped has already been installed by SageMath
## -- ##
## Checking whether SageMath should install SPKG boost... ##
## -- ##
configure:13439: checking whether any of boost_cropped is installed as
or will be installed as SPKG
configure:13443: result: yes; install boost as well
configure:13500: no suitable system package found for SPKG boost

(boost the Sage package depends on boost_cropped)



>
> On 7/1/20 3:15 PM, Michael Orlitzky wrote:
> > On 2020-07-01 08:36, Dima Pasechnik wrote:
> >>
> >> In building giac, you get even more strange error, a missing gammaf 
> >> function.
> >> (something to do with musl?)
> >
> > Giac upstream probably needs to
> >
> >#define _GNU_SOURCE
> >
> > to get the gammaf() function with musl, since it's not standardized.
> >
>
> --
> You received this message because you are subscribed to the Google Groups 
> "sage-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to sage-devel+unsubscr...@googlegroups.com.
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/sage-devel/6eb2b70d-6946-8d8c-6c66-219fd75634db%40gmail.com.

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-devel/CAAWYfq3EKuestdSMWMc%2BQXPYVHnGm8HnSe4FPyj3Kw_FPqA7EQ%40mail.gmail.com.

Re: [sage-devel] build errors on void linux

[Nicolo' Piazzalunga]

Thanks.
I'm puzzled by 'does not support check for system package'/'no suitable 
system package' in config.log
E.g. why is it not finding boost, isl, and why is it building ecl (just 
to list a few), all of which are installed as system packages?


On 7/1/20 3:15 PM, Michael Orlitzky wrote:

On 2020-07-01 08:36, Dima Pasechnik wrote:


In building giac, you get even more strange error, a missing gammaf function.
(something to do with musl?)


Giac upstream probably needs to

   #define _GNU_SOURCE

to get the gammaf() function with musl, since it's not standardized.



--
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-devel/6eb2b70d-6946-8d8c-6c66-219fd75634db%40gmail.com.

Re: [sage-devel] build errors on void linux

[Michael Orlitzky]

On 2020-07-01 08:36, Dima Pasechnik wrote:
> 
> In building giac, you get even more strange error, a missing gammaf function.
> (something to do with musl?)

Giac upstream probably needs to

  #define _GNU_SOURCE

to get the gammaf() function with musl, since it's not standardized.

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-devel/033ad8c8-6b88-9234-454f-a1bd26d7a28f%40orlitzky.com.

Re: [sage-devel] trouble building sage on mac

[Dima Pasechnik]

my working conjecture was that Anne had two conflicting installations
of freetype, for I cannot imagine Homebrew serving version 2.3 of
freetype.


On Wed, Jul 1, 2020 at 1:36 PM Zachary Scherr  wrote:
>
> Also, you might also consider doing something like the following:
>
> > cd /usr
> > find -L . -name 'ftheader.h'
>
> you should see the following locations:
>
> ./local/include/freetype2/freetype/config/ftheader.h
> ./local/var/homebrew/linked/freetype/include/freetype2/freetype/config/ftheader.h
> ./local/opt/freetype2/include/freetype2/freetype/config/ftheader.h
> ./local/opt/freetype/include/freetype2/freetype/config/ftheader.h
> ./local/Cellar/freetype/2.10.2/include/freetype2/freetype/config/ftheader.h
>
> notice though that 
> "/usr/local/Cellar/freetype/2.10.2/include/freetype2/freetype/config/ftheader.h"
>  is the real file and the others should all be symlinks of this one.
>
> On Wednesday, July 1, 2020 at 8:26:58 AM UTC-4 Zachary Scherr wrote:
>>
>> Hi All,
>>
>>I just thought I would chime in on the homebrew side of things.  I have 
>> freetype installed via homebrew on Catalina 10.15.5 and when I execute
>>
>> > pkg-config --modversion freetype2
>>
>> I get back 23.2.17.  Anne seemed to get back version 9.8.3, which means that 
>> something is screwy with your freetype.
>>
>> Some things you could try:
>>
>> 1). If you have already installed freetype via homebrew then you can try
>>  brew unlink freetype && brew link freetype
>>
>> 2). It's possible that "brew doctor" will pick up on the inconsistencies and 
>> will give recommendations as to how to fix it.
>>
>> Your log file specifically mentioned that the system could not find 
>> ftconfig.h.  With freetype correctly installed via homebrew you can do the 
>> following:
>>
>> cd /usr/local/include
>> ls -l | grep freetype2
>>
>> and you should see that that folder is symlinked to 
>> /usr/local/Cellar/freetype/2.10.2/include/freetype2 which is the folder that 
>> actually contains all the freetype header files.
>>
>> Not sure how much you want to invest in fixing this, but my suggestion would 
>> be to head to /usr/local and try to find ftheader.h, which is the header 
>> file that defines where ftconfig.h is located.  If you find any ftheader.h 
>> file that is not originating from 
>> local/Cellar/freetype/2.10.2/include/freetype2/freetype/config/ftheader.h 
>> then you should delete it.
>>
>>
>>
>> On Wednesday, July 1, 2020 at 2:53:05 AM UTC-4 dim...@gmail.com wrote:
>>>
>>>
>>>
>>> On Wed, 1 Jul 2020, 07:41 Anne Schilling,  wrote:



 On Tuesday, June 30, 2020 at 12:22:47 AM UTC-7, Dima Pasechnik wrote:
>
>
>
> On Tue, 30 Jun 2020, 07:46 Anne Schilling,  wrote:
>>
>> Here is the new config.log file. It still does not build:
>>
>> https://www.math.ucdavis.edu/~anne/config.log
>
>
> do you get the same error while building pillow (missing header)? If so, 
> it looks as if you removed the old freetype dylib files, but left headers 
> there.


 Thank you, but what exactly am I supposed to delete? I delete the folder 
 /usr/local/Cellar/freetype
>>>
>>>
>>> you had to delete the installation of freetype in /usr/local/{lib, include, 
>>> lib/pkgconfig}.
>>>
>>> and after that, (re)install freetype in Homebrew.
>>>
>>> Deleting /usr/local/Cellar/freetype
>>> probably was not a good idea - that is where, I guess, you have a good 
>>> installation of freetype (which you should uninstall using the brew 
>>> commands, not just deleting from shell)
>>>
>>> But I don't know whether your Homebrew is not too old. Can you do
>>>
>>> brew install freetype
>>>
>>> (or is it freetype2)
>>>
>>> paying attention to the version it mentions?
>>>
>>>
>>> in any case,
>>>
>>> ./configure --with-freetype=system
>>>
>>> should force sage to build freetype on its own.
>>>
>>>

> Also, did you source .homebrew-build-env
> before doing ./configure && make ?
> As you use libraries from Homebrew, this might be crucial.


 I did that and it still does not work. See

 https://www.math.ucdavis.edu/~anne/config.log

 Is it possible that this might be an issue

 https://github.com/Homebrew/legacy-homebrew/issues/43471


 --
 You received this message because you are subscribed to the Google Groups 
 "sage-devel" group.
 To unsubscribe from this group and stop receiving emails from it, send an 
 email to sage-devel+...@googlegroups.com.

 To view this discussion on the web visit 
 https://groups.google.com/d/msgid/sage-devel/6ec9e78d-fd63-4122-91b2-d4600836aee8o%40googlegroups.com.
>
> --
> You received this message because you are subscribed to the Google Groups 
> "sage-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to sage-devel+unsubscr...@googlegroups.com.
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/sage-

Re: [sage-devel] build errors on void linux

[Dima Pasechnik]

Something is very unusual in your setup; e,g. in building R you get

Error in iconv(x, "UTF-8", outputEncoding, sub = "byte", mark = FALSE) :
  unsupported conversion from 'UTF-8' to 'ASCII//TRANSLIT'

In building giac, you get even more strange error, a missing gammaf function.
(something to do with musl?)

suitesparse errors out with
gcc: fatal error: cannot read spec file 'libgomp.spec': No such file
or directory
(try to install OpenMP, I guess)

no idea what happens in the ecl build.



On Wed, Jul 1, 2020 at 10:55 AM Nicolo'  wrote:
>
> Hi, can anyone help?
>
> I'm trying to build Sage on Void Linux.
> My system uses musl and libressl (see also attached list of installed 
> packages).
> My goal is to build in a chroot to get a binary package, but for the moment 
> I'm just building locally with
>
> ./configure
> MAKE='make -j12' make
>
> I get errors building packages, some of which are actually installed on my 
> system.
> It seems I cannot attach full install log, which is 27MB, but it ends with
>
> real 7m53.793s
> user 31m38.472s
> sys 1m31.524s
> ***
> Error building Sage.
>
> The following package(s) may have failed to build (not necessarily
> during this run of 'make all-start'):
>
> * package: giac-1.5.0.63-p0
>   last build time: Jun 30 20:57
>   log file:/home/nicolo/build/sage-9.1/logs/pkgs/giac-1.5.0.63-p0.log
>   build directory: 
> /home/nicolo/build/sage-9.1/local/var/tmp/sage/build/giac-1.5.0.63-p0
>
> * package: suitesparse-5.6.0
>   last build time: Jun 30 20:57
>   log file:/home/nicolo/build/sage-9.1/logs/pkgs/suitesparse-5.6.0.log
>   build directory: 
> /home/nicolo/build/sage-9.1/local/var/tmp/sage/build/suitesparse-5.6.0
>
> * package: ecl-16.1.2.p5
>   last build time: Jun 30 20:58
>   log file:/home/nicolo/build/sage-9.1/logs/pkgs/ecl-16.1.2.p5.log
>   build directory: 
> /home/nicolo/build/sage-9.1/local/var/tmp/sage/build/ecl-16.1.2.p5
>
> * package: r-3.6.2.p0
>   last build time: Jun 30 20:59
>   log file:/home/nicolo/build/sage-9.1/logs/pkgs/r-3.6.2.p0.log
>   build directory: 
> /home/nicolo/build/sage-9.1/local/var/tmp/sage/build/r-3.6.2.p0
>
> --
> You received this message because you are subscribed to the Google Groups 
> "sage-devel" group.
> To unsubscribe from this group and stop receiving emails from it, send an 
> email to sage-devel+unsubscr...@googlegroups.com.
> To view this discussion on the web visit 
> https://groups.google.com/d/msgid/sage-devel/a82769f3-5685-4276-8ef6-2782202071f3o%40googlegroups.com.

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-devel/CAAWYfq0rR67L1p-NOYOuzGzxMng6tsENqNeFJfn_WcEwYuSpEw%40mail.gmail.com.

Re: [sage-devel] trouble building sage on mac

[Zachary Scherr]

Also, you might also consider doing something like the following:

> cd /usr
> find -L . -name 'ftheader.h'

you should see the following locations:

./local/include/freetype2/freetype/config/ftheader.h
./local/var/homebrew/linked/freetype/include/freetype2/freetype/config/ftheader.h
./local/opt/freetype2/include/freetype2/freetype/config/ftheader.h
./local/opt/freetype/include/freetype2/freetype/config/ftheader.h
./local/Cellar/freetype/2.10.2/include/freetype2/freetype/config/ftheader.h

notice though that 
"/usr/local/Cellar/freetype/2.10.2/include/freetype2/freetype/config/ftheader.h"
 
is the real file and the others should all be symlinks of this one.

On Wednesday, July 1, 2020 at 8:26:58 AM UTC-4 Zachary Scherr wrote:

> Hi All,
>
>I just thought I would chime in on the homebrew side of things.  I have 
> freetype installed via homebrew on Catalina 10.15.5 and when I execute
>
> > pkg-config --modversion freetype2
>
> I get back 23.2.17.  Anne seemed to get back version 9.8.3, which means 
> that something is screwy with your freetype.
>
> Some things you could try:
>
> 1). If you have already installed freetype via homebrew then you can try
>  brew unlink freetype && brew link freetype
>
> 2). It's possible that "brew doctor" will pick up on the inconsistencies 
> and will give recommendations as to how to fix it.
>
> Your log file specifically mentioned that the system could not find 
> ftconfig.h.  With freetype correctly installed via homebrew you can do the 
> following:
>
> cd /usr/local/include
> ls -l | grep freetype2
>
> and you should see that that folder is symlinked to 
> /usr/local/Cellar/freetype/2.10.2/include/freetype2 which is the folder 
> that actually contains all the freetype header files.
>
> Not sure how much you want to invest in fixing this, but my suggestion 
> would be to head to /usr/local and try to find ftheader.h, which is the 
> header file that defines where ftconfig.h is located.  If you find any 
> ftheader.h file that is not originating from 
> local/Cellar/freetype/2.10.2/include/freetype2/freetype/config/ftheader.h 
> then you should delete it.
>
>
>
> On Wednesday, July 1, 2020 at 2:53:05 AM UTC-4 dim...@gmail.com wrote:
>
>>
>>
>> On Wed, 1 Jul 2020, 07:41 Anne Schilling,  wrote:
>>
>>>
>>>
>>> On Tuesday, June 30, 2020 at 12:22:47 AM UTC-7, Dima Pasechnik wrote:



 On Tue, 30 Jun 2020, 07:46 Anne Schilling,  
 wrote:

> Here is the new config.log file. It still does not build:
>
> https://www.math.ucdavis.edu/~anne/config.log 
> 
>

 do you get the same error while building pillow (missing header)? If 
 so, it looks as if you removed the old freetype dylib files, but left 
 headers there.

>>>
>>> Thank you, but what exactly am I supposed to delete? I delete the folder 
>>> /usr/local/Cellar/freetype
>>>
>>
>> you had to delete the installation of freetype in /usr/local/{lib, 
>> include, lib/pkgconfig}. 
>>
>> and after that, (re)install freetype in Homebrew.
>>
>> Deleting /usr/local/Cellar/freetype 
>> probably was not a good idea - that is where, I guess, you have a good 
>> installation of freetype (which you should uninstall using the brew 
>> commands, not just deleting from shell)
>>
>> But I don't know whether your Homebrew is not too old. Can you do
>>
>> brew install freetype 
>>
>> (or is it freetype2) 
>>
>> paying attention to the version it mentions?
>>
>>
>> in any case, 
>>
>> ./configure --with-freetype=system 
>>
>> should force sage to build freetype on its own.
>>
>>
>>
>>> Also, did you source .homebrew-build-env
 before doing ./configure && make ?
 As you use libraries from Homebrew, this might be crucial.

>>>
>>> I did that and it still does not work. See
>>>
>>> https://www.math.ucdavis.edu/~anne/config.log 
>>> 
>>>
>>> Is it possible that this might be an issue
>>>
>>> https://github.com/Homebrew/legacy-homebrew/issues/43471
>>>
>>>
>>> -- 
>>> You received this message because you are subscribed to the Google 
>>> Groups "sage-devel" group.
>>> To unsubscribe from this group and stop receiving emails from it, send 
>>> an email to sage-devel+...@googlegroups.com.
>>>
>> To view this discussion on the web visit 
>>> https://groups.google.com/d/msgid/sage-devel/6ec9e78d-fd63-4122-91b2-d4600836aee8o%40googlegroups.com
>>>  
>>> 
>>> .
>>>
>>

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@goog

Re: [sage-devel] trouble building sage on mac

[Zachary Scherr]

Hi All,

   I just thought I would chime in on the homebrew side of things.  I have 
freetype installed via homebrew on Catalina 10.15.5 and when I execute

> pkg-config --modversion freetype2

I get back 23.2.17.  Anne seemed to get back version 9.8.3, which means 
that something is screwy with your freetype.

Some things you could try:

1). If you have already installed freetype via homebrew then you can try
 brew unlink freetype && brew link freetype

2). It's possible that "brew doctor" will pick up on the inconsistencies 
and will give recommendations as to how to fix it.

Your log file specifically mentioned that the system could not find 
ftconfig.h.  With freetype correctly installed via homebrew you can do the 
following:

cd /usr/local/include
ls -l | grep freetype2

and you should see that that folder is symlinked to 
/usr/local/Cellar/freetype/2.10.2/include/freetype2 which is the folder 
that actually contains all the freetype header files.

Not sure how much you want to invest in fixing this, but my suggestion 
would be to head to /usr/local and try to find ftheader.h, which is the 
header file that defines where ftconfig.h is located.  If you find any 
ftheader.h file that is not originating from 
local/Cellar/freetype/2.10.2/include/freetype2/freetype/config/ftheader.h 
then you should delete it.



On Wednesday, July 1, 2020 at 2:53:05 AM UTC-4 dim...@gmail.com wrote:

>
>
> On Wed, 1 Jul 2020, 07:41 Anne Schilling,  wrote:
>
>>
>>
>> On Tuesday, June 30, 2020 at 12:22:47 AM UTC-7, Dima Pasechnik wrote:
>>>
>>>
>>>
>>> On Tue, 30 Jun 2020, 07:46 Anne Schilling,  wrote:
>>>
 Here is the new config.log file. It still does not build:

 https://www.math.ucdavis.edu/~anne/config.log 
 

>>>
>>> do you get the same error while building pillow (missing header)? If so, 
>>> it looks as if you removed the old freetype dylib files, but left headers 
>>> there.
>>>
>>
>> Thank you, but what exactly am I supposed to delete? I delete the folder 
>> /usr/local/Cellar/freetype
>>
>
> you had to delete the installation of freetype in /usr/local/{lib, 
> include, lib/pkgconfig}. 
>
> and after that, (re)install freetype in Homebrew.
>
> Deleting /usr/local/Cellar/freetype 
> probably was not a good idea - that is where, I guess, you have a good 
> installation of freetype (which you should uninstall using the brew 
> commands, not just deleting from shell)
>
> But I don't know whether your Homebrew is not too old. Can you do
>
> brew install freetype 
>
> (or is it freetype2) 
>
> paying attention to the version it mentions?
>
>
> in any case, 
>
> ./configure --with-freetype=system 
>
> should force sage to build freetype on its own.
>
>
>
>> Also, did you source .homebrew-build-env
>>> before doing ./configure && make ?
>>> As you use libraries from Homebrew, this might be crucial.
>>>
>>
>> I did that and it still does not work. See
>>
>> https://www.math.ucdavis.edu/~anne/config.log 
>> 
>>
>> Is it possible that this might be an issue
>>
>> https://github.com/Homebrew/legacy-homebrew/issues/43471
>>
>>
>> -- 
>> You received this message because you are subscribed to the Google Groups 
>> "sage-devel" group.
>> To unsubscribe from this group and stop receiving emails from it, send an 
>> email to sage-devel+...@googlegroups.com.
>>
> To view this discussion on the web visit 
>> https://groups.google.com/d/msgid/sage-devel/6ec9e78d-fd63-4122-91b2-d4600836aee8o%40googlegroups.com
>>  
>> 
>> .
>>
>

-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-devel/8ffdfa07-94c0-45f2-81e0-d0cc4d710315n%40googlegroups.com.

[sage-devel] [Bug] Possible Bug in FrobeniusMap for ntl finite fields of characteristic 2

[Maxime Bombar]

Hi,

I think there is a bug with the Frobenius Maps of finite fields of
characteristic 2 implemented using NTL :

Here is what happened with finite fields implemented using the Givaro
library (finite fields of characteristic 2 and degree <= 15). This is
what we expect.

```
sage: k = GF(2^15)
sage: k.frobenius_endomorphism() == k.frobenius_endomorphism()
True
sage: type(k)

```

Here is what happened for finite field of odd characteristic
(implemented using the PARI library) :

```
sage: k = GF(3^16)
sage: k.frobenius_endomorphism() == k.frobenius_endomorphism()
True
sage: type(k)

```


However, it raises an exception for finite field implemented using NTL
(I can give a full traceback, but this is easy to reproduce) :

```
sage: k = GF(2^16)
sage: type(k)

sage: k.frobenius_endomorphism() == k.frobenius_endomorphism()
TypeError: None fails to convert into the map's domain Finite Field in
z16 of size 2^16, but a `pushforward` method is not properly implemented
```

The same behaviour happens if we force the implementation of a finite
field with NTL :


```
sage: k = GF(2^15, impl="ntl")
sage: type(k)

sage: k.frobenius_endomorphism() == k.frobenius_endomorphism()
TypeError: None fails to convert into the map's domain Finite Field in
z15 of size 2^15, but a `pushforward` method is not properly implemented
```

I checked this on each version of sagemath since 8.6 using CoCalc on the
sage website.

I'm trying to help here, and if I find something I will be pleased to
propose a fix, so if someone familiar with the codebase is willing to
give me a hint, I would be glad.


Many thanks,

-- 
Maxime Bombar






-- 
You received this message because you are subscribed to the Google Groups 
"sage-devel" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to sage-devel+unsubscr...@googlegroups.com.
To view this discussion on the web visit 
https://groups.google.com/d/msgid/sage-devel/3a3e2eda-230c-fd42-3401-c53d3f497f53%40inria.fr.


signature.asc
Description: OpenPGP digital signature