Ideas from debian

[asarhaddon]

Hello.
Here are some ideas from the Debian packaging. It expect the quote or desktop fixes to be consensual. On the other hand, the changes in the build system are only drafts in order to start discussion.

[shlomif]

Hi @asarhaddon . I reviewed the commits and while some changes seem like steps forward, I dislike:

• move away from rinutils
• move away from shlomif_common
• some other scriptology changes
• splitting the ritchie/Jobs/etc. fortune into 3 as the quotes seem related.

[asarhaddon]

Hello.

move away from rinutils

I think GCC_UNUSED increases the complexity for no benefit here.
For LAST and COUNT, I think that managing a dependency for two macros is overkill, but I understand that the compromise is a question of taste.

move away from shlomif_common

I have removed this part, except the VERSION stuff that seems more simple and robust.

some other scriptology changes

What you are referring to is not obvious for me. My motivations were:

• depending on sh is more robust than depending on perl modules (some parts of the current cmake stuff already depends on sh and redirections)
• mixing the generator and the preprocessor only simplifies the overall process
• an external tool like docbook2xml is easyer to deal with than a $language module.
• when possible, I try and avoid variables with distinct meanings/values hiding each other.

splitting the ritchie/Jobs/etc. fortune into 3 as the quotes seem related.

I do not understand. I can merge the 4 Pratchett quotes if that is what you mean (or forward the question to the author :-).

[asarhaddon]

I have moved more trivial commits down in the sequence in order to make a partial merge fast-forward. The introduction of components cannot be applied before the replacement of INSTALL_RENAME_MAN. What should change further ?

[shlomif]

@asarhaddon : hi. I find dealing with a 14-commits-pullreq overwhelming. Can you prepare a single-purpose pullreq?

[asarhaddon]

Each commit is now in its own merge request, with one exception.
#125 splits the install-data and install-exec targets. It has non-trivial conflicts with #130 and #141. I am keeping them duplicated in #125 for now, and will rebase #125 properly once you have reviewed them.