Re: [PATCH] gnulib-tool.py: Follow gnulib-tool changes, part 50.

[Bruno Haible]

Hi Collin,

> Assuming that there are no other issues with this patch

There are no issues :)

> can you check
> the commit message and ChangeLog entry before pushing? I noticed that
> the ChangeLog entry disagrees with the git log and gnulib-tool.py.TODO
> for one of the commits. I'm new to the ChangeLog stuff so I will let
> you decide which is more useful to reference.

When there is a mistake in a ChangeLog and 'git log' entry, we can only
change the ChangeLog a posteriori, not the 'git log' entry.

> Here is the commit it was changed:
> 
> $ git  diff eed8d6e1b4924b0a7e93905871094909ba0118a3 
> eed8d6e1b4924b0a7e93905871094909ba0118a3^

This is a reverse diff. I prefer to look at

$ git diff eed8d6e1b4924b0a7e93905871094909ba0118a3^ 
eed8d6e1b4924b0a7e93905871094909ba0118a3

This makes it clear that the text in the ChangeLog should be used.

> I'm still learning how gnulib-tool works so I'm not sure how to get
> 'tests=lib/alloca.c' in "$tmp"/new-files to test the second
> conditional [1]. Feel free to let me know if there is an easy way to
> do it.

I think you did it correctly. In the function rewrite_new_files, it can
happen that a file starts with 'tests=lib/'. In GLImport.py line 967,
you can see that this file argument is stored as second element of a tuple
that later gets added to filetable['new'].

Bruno