m4ri, pynormaliz, etc.

[Aaron Lauve]

I'm trying to update to sage 10.4.rc1

last local-working copy of sage was 10.4.rc0

[lauve@localhost sage %]  git branch

* develop

[lauve@localhost sage %]  git pull

[lauve@localhost sage %]  make distclean && make

This aborts with an error building m4ri-20200125 package (error log attached).

...

Some possibly relevant background:

Wanting to compute the h* vector of a polytope, I learned that I needed to install pynormaliz.

Running sage -i pynormaliz threw an error (e_antic log attached). 

Before going any further, I figured I should upgrade from 10.4.rc0 to 10.4.rc1.

That seemed to cause a different problem. Where I stand now, as mentioned above, is an error installing the m4ri-20200125 package.

Any help would be appreciated.

[Aaron Lauve]

Sorry, I just assumed 10.4.rc1 was the most recent, from a quick glance through the posts to sage-level.

As the config log shows (attached), I am in fact trying to build 10.5.beta0

> (From Dima)

> Why 10.4.rc1, and not 10.4? 
> 
> It might be more natural to try the latest beta, 10.5.beta0, too.

> 

> Anyhow, the problem is with libpng, as you can see in your m4ri log.

> 

> Please post your top-level config.log so that we can see where your png should come from.
> 

[Aaron Lauve]

Hmm... 

Looking through the install-from-source-code advice, it looks like I may have forgotten this step:

     .homebrew-build-env

Is this necessary before I run 

     make distclean && make

if I'm installing on a Mac and am using homebrew to manage (many many) packages?

[Dima Pasechnik]

On Fri, Aug 2, 2024 at 9:22 PM Aaron Lauve <la...@math.luc.edu> wrote:
>
> Hmm...
> Looking through the install-from-source-code advice, it looks like I may have forgotten this step:
>
> .homebrew-build-env
>
> Is this necessary before I run
>
> make distclean && make
>
> if I'm installing on a Mac and am using homebrew to manage (many many) packages?

yes, doing

. .homebrew-build-env

before "make"

is essential for telling Sage where to look for Homebrew packages.
That's why you get this png trouble.
(note that "." is a shell command...)

HTH
Dima

>
> On Friday, August 2, 2024 at 3:09:38 PM UTC-5 Aaron Lauve wrote:
>>
>> Sorry, I just assumed 10.4.rc1 was the most recent, from a quick glance through the posts to sage-level.
>> As the config log shows (attached), I am in fact trying to build 10.5.beta0
>>
>>
>>
>>
>> > (From Dima)
>> > Why 10.4.rc1, and not 10.4?
>> >
>> > It might be more natural to try the latest beta, 10.5.beta0, too.
>> >
>> > Anyhow, the problem is with libpng, as you can see in your m4ri log.
>> >
>> > Please post your top-level config.log so that we can see where your png should come from.
>> >
>> On Friday, August 2, 2024 at 12:52:00 PM UTC-5 Aaron Lauve wrote:
>>
>> I'm trying to update to sage 10.4.rc1
>> last local-working copy of sage was 10.4.rc0
>>
>> [lauve@localhost sage %] git branch
>> * develop
>> [lauve@localhost sage %] git pull
>> [lauve@localhost sage %] make distclean && make
>>
>> This aborts with an error building m4ri-20200125 package (error log attached).
>>
>> ...
>> Some possibly relevant background:
>>
>> Wanting to compute the h* vector of a polytope, I learned that I needed to install pynormaliz.
>>
>> Running sage -i pynormaliz threw an error (e_antic log attached).
>> Before going any further, I figured I should upgrade from 10.4.rc0 to 10.4.rc1.
>>
>> That seemed to cause a different problem. Where I stand now, as mentioned above, is an error installing the m4ri-20200125 package.
>>
>> Any help would be appreciated.
>>
>>

> --
> 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/5e58b898-73d8-4518-8ab2-770e3e4d6b0en%40googlegroups.com.

[Matthias Koeppe]

Yes, this step is important.

[Aaron Lauve]

Duh. Sorry for the trouble. Everything compiled correctly.

For future reference, is it also necessary if I've only made small changes to packages?...

where it would ordinarily suffice to just run `sage -br`?

[Matthias Koeppe]

On Friday, August 2, 2024 at 6:59:13 PM UTC-7 Aaron Lauve wrote:

is it also necessary if I've only made small changes to packages?...

where it would ordinarily suffice to just run `sage -br`?

Yes