eclib version checking

[John Cremona]

In build/pkgs/eclib/spkg-configure.m4 tere is code to check whether an already installed version of eclib is new enough.  It use to check that the version agreed exactly, but now it uses [ge] ("greater or equal") to do this.

BUT I have version 20240408 installed which has greater than the minimal expected 20231212 and yet configure does not accept it:

## Checking whether SageMath should install SPKG eclib... ##
## ------------------------------------------------------ ##
configure:25381: checking whether any of ntl pari flint is installed as or will be installed as SPKG
configure:25391: result: no
configure:25397: checking for eclib = 20231212
configure:25404: $PKG_CONFIG --exists --print-errors "eclib = 20231212"
Requested 'eclib = 20231212' but version of eclib is 20240408
configure:25407: $? = 1
configure:25421: $PKG_CONFIG --exists --print-errors "eclib = 20231212"
Requested 'eclib = 20231212' but version of eclib is 20240408
configure:25424: $? = 1
configure:25438: result: no
Requested 'eclib = 20231212' but version of eclib is 20240408
configure:25578: no suitable system package found for SPKG eclib
## ---------------------------------------------------- ##

Is there something wrong in the m4 code?

John

[Michael Orlitzky]

On Wed, 2024-06-26 at 15:26 +0100, John Cremona wrote:
> In build/pkgs/eclib/spkg-configure.m4 tere is code to check whether an
> already installed version of eclib is new enough. It use to check that the
> version agreed exactly, but now it uses [ge] ("greater or equal") to do
> this.
>

Did you update spkg-configure.m4 yourself? The one I have (develop
branch) is still using equality:

m4_pushdef([SAGE_ECLIB_VER],["20231212"])
PKG_CHECK_MODULES([ECLIB], [eclib = SAGE_ECLIB_VER],
...

In any case, after changing it, you have to re-run ./bootstrap.

[Dima Pasechnik]

there are two tests - and you get the failure on the 2st one, which checks for equality of versions, in
PKG_CHECK_MODULES.
You need to change '=' there to '>='

(and ./bootstrap)

[John Cremona]

Dime correctly diagnosed the problem, and I just made a trivial PR to fix it: https://github.com/sagemath/sage/pull/38284

John