On Mon, 4 Jul 2022 at 21:49, Daniel Sahlberg <[email protected]>
wrote:

> Den mån 4 juli 2022 kl 17:09 skrev Bert Huijben <[email protected]>:
>
>> That line will add the ./configure time dependencies (or gen-make.py for
>> Windows). The dependencies on our own libraries are added by the ‘libs =’
>> line right above.
>>
>
> The libs at issue (libsvn_delta and libsvn_subr) are already in libs. The
> problem seems to be that the linker is looking for these libs in the
> install location ($PREFIX/lib) and they are not installed there when it
> tries to link libsvn_ra_serf.
>
> libsvn_delta and libsvn_subr seems to be installed by install-fsmod-lib.
>
>  If you don’t get these, then it is most likely because the files don’t
>> exist/aren’t build at the right time, and in that case you need to use a
>> different make target, like you already suggested.
>>
>
> I tried to change the dependency to $(SVN_RA_LIB_INSTALL_DEPS) (as with
> libsvn_ra a few lines above), however I get a warning about a circular
> dependency. I could also depend directly on install-fsmod but I'm not sure
> this is better.
>
> Adding more and more unneeded dependencies will make things far harder to
>> diagnose in the future, so we should remove strange inner dependencies for
>> just a specific environment, or use the proper systems if we ever want to
>> be able to find bugs in this space.
>>
>
> Agreed. Thanks for taking your time to explain!
>
> Can you take the time to check the reproduction receipt to see if there is
> anything obvious missing? I'm doing this in a fairly clean Ubuntu 21.10 (in
> particular, with no SVN packages installed - I removed them before making
> the tests and I use the freshly built client whenever need to access the
> repo).
>
> Would it be better to factor out libsvn_delta and libsvn_subr to their own
> install target and make everything else depend on this new target?
> (Grasping for straws here!)
>
> Hi Daniel,

I am also confused by this change: libsvn_ra_serf is unrelated
to SVN_FS_LIB_INSTALL_DEPS. Maybe we have the problem, but the fix doesn't
look right.

I would suggest that we either investigate the root cause of the problem or
revert the change.

-- 
Ivan Zhakov

Reply via email to