[SoaS] Issue with telepathy-salut

[Chihurumnaya Ibiam]

On Tue, Dec 22, 2020 at 8:53 PM James Cameron <quozl at laptop.org> wrote:

> Summary: you are conflating git patches with diff patches.
>

Agreed, makes sense now.


> On Tue, Dec 22, 2020 at 12:55:49PM +0100, Chihurumnaya Ibiam wrote:
> > telepathy-salut-0.8.1.tar.gz does contain lib/ext/wocky as a
> > directory with files in it and it has the github repo as it's
> > upstream and that doesn't contain lib/ext /wocky as a directory but
> > as a wocky submodule. I don't see a reason why the git submodule is
> > affecting it at the moment
>
> It has affected it because of how you prepared the patch using git
> instead of using diff.
>

Yeah I normally prepare such patches with git as most of the time the
upstream on github is still active.
Now I know different.


> > Yeah I haven't created an PRs as I'm trying to get it to build
> > locally first, you can find the patch here.
>
> The patch is in error because;
>
> (a) it is generated using git(1), and contains no content except
> submodule hashes, yet you are asking to apply using patch(1),
>
> (b) unlike patch0, it does not have complete file paths.
>
> You should generate the patch again, using diff -ru, between two
> clones that have been prepared to the corresponding states.
>

Yeah I will.

>
> > The commit upstream shows that there was a change but the change was
> > to a submodule, the source files from telepathyfreedesktop.org
> > contains lib/ext/wocky but the source on github doesn't and that
> > explains the error from build.log
>
> You should not use the tar.gz from github, because it
>
> (a) does not equal the distributed software; therefore unsupported,
>
> (b) does not contain autoconf prepared source; therefore broken, and
>
> (c) does not have submodules expanded; therefore incomplete.
>
> > I used rpmbuild to build just the %prep section and ls shows there
> > files in lib /ext/wocky but still throws same error when applying
> > the patch and I don't get why right now but I think it's because the
> > patch is pointing to a submodule commit.
>
> No, it is because the patch is not intended for processing by
> patch(1).
>

Yeah, makes sense now.


> The patch(1) command has no clue (or glue) for handling git patches in
> that way.
>

I understand now.


> Hope that helps!
>

Yeah, thanks a lot.

>
> --
> James Cameron
> https://quozl.linux.org.au/



-- 

Ibiam Chihurumnaya
ibiamchihurumnaya at gmail.com
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.sugarlabs.org/archive/soas/attachments/20201223/1f881239/attachment.htm>