[Top][All Lists]

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[bug#31733] [PATCH] gnu: encfs: Update to 1.9.5.

From: Kei Kebreau
Subject: [bug#31733] [PATCH] gnu: encfs: Update to 1.9.5.
Date: Sat, 16 Jun 2018 11:23:33 -0400
User-agent: Gnus/5.13 (Gnus v5.13) Emacs/26.1 (gnu/linux)

Pierre Neidhardt <address@hidden> writes:

>>> Any particular reason you decided to use '(zero? (system* ...))' rather
>>> than '(invoke ...)' for this phase? I tried it with 'invoke' and it
>>> worked fine.
> I did not know about `invoke' :)
> I just found its definition in guix/build/utils.scm.

Well now you know! :)

>> Also, do note that the snippet needs to be updated, as bundled libraries
>> (now including googletest) have moved to the "vendor" subdirectory. Once
>> you remove the "vendor" subdirectory, you may have to add 'googletest'
>> from '(gnu packages check)' to the native-inputs.
> So the vender directory has 3 dependencies:
> - tinyxml which is already used externally, so it's safe to remove the
>   embedded version.
> - easyloggingpp which is not packaged in Guix I believe.  We need to
>   package it first before we can remove it.

Observation agrees with both of these.

> - googletest: this one does not have an option in CMakeList.txt to be
>   used externally.  So either we report upstream or we patch the
>   CMakeList.txt.

I see that while CMakeLists.txt doesn't have an external option to
control this, it *does* look for googletest in the vendor
directory. Perhaps we can add our own googletest as a native input and
remove the bundled googletest and CMakeLists.txt's references to it in a

> All things considered, why not sticking to the vendored libraries?

We try to unbundle third-party software in our packages in order to ease
the burden of security maintenance and avoid duplication of work in
general. There are probably other reasons that I can't remember right
now, too.

> What do you think?

This sounds actionable to me!

Attachment: signature.asc
Description: PGP signature

reply via email to

[Prev in Thread] Current Thread [Next in Thread]