Re: [Monotone-devel] The everlasting release

[Stephen Leake]

Thomas Keller <address@hidden> writes:

> Review the
> complete chapter 5 and 7 

This is now done (really :).

There are a few issues:

unix/terminal.cc terminal_width returns 0 on fail; win32 returns 80

    the manual used to say "returns 72", now it says "80". Not a big
    deal, but is there a reason for returning 0 on Unix?

Perhaps we should describe Lua functions defined in std_hooks.lua, such
as "temp_file", "execute", "hook_functions"? These are not available with
--no-builtin-rcfiles, but they are useful.

We should probably delete "hook_wrapper" and the section "Meta Hooks"
from the manual; there is no reason for the user to override it.

> for example, the things about our "apparent" IDNA support read
> dubious, 

I don't know enough about internationalization to review this.

> there is a hook mentioned "display_decoded_idna" which is nowhere else
> referenced any longer, 

Deleted.

> the hash integrity discussion might need an update to accompany more
> recent development, 

I don't know enough about hashing/cryptography to review this.

> move the cvs phrasebook stuff into the wiki 

I can work on this.

> and add accompanying sections for svn, git, hg..., 

I don't know those well enough.

> and finally I still think the huge pcre pattern documentation is a bit
> overwhelming for the little usage of patterns we have for the user
> actually.)

pcresyntax.texi and pcrepattern.texi are derived from upstream sources.
It's very simple to just include them. It's also very easy for the
reader to ignore them; they are just entries in the table of contents.

(aside: Igoring the pcre section is easier with info than with HTML,
since the HTML is one big file, with no internal navigation links. It
would help if we deleted the "--no-headers" option, but that interacts
very badly with the texinfo.css style sheet. I'd prefer deleting the
style sheet)

I don't think it's worth the manpower to try to simplify the pcre
section.

On the other hand, it would be good to add a section defining a "glob".

> As a fall-out of the aforementioned docathon I started to clean up our
> source tree a little and moved stuff around (work is kept in
> nvm.source-tree-cleanup). Thanks again to Richard which helped me with
> some automake issues here and there this slowly gets in a usable shape.
> Whats missing is the documentation and more testing on other platforms
> (Win32 most noticably) before this hits the trunk. 

I think this is ready.

> In the meantime Tim also worked quite hard on nvm.visualc, which we
> might also want to get a little earlier into trunk for better testing.
> (Please see Tim's earlier mail about details here.)

I can't help here.

-- 
-- Stephe