I was looking at some of the PRs for DIPs and was wondering how
to go ahead and use them in my code. A few of the forum regulars
have mentioned applying their own patches before, and so I would
like to ask those of you that do, what your preferred method is,
or if this is an all-around bad idea and I should just wait for
the PRs to be merged?

On Thu, 25 Dec 2014 20:15:10 +0000
Vlad Levenfeld via Digitalmars-d <digitalmars-d@puremagic.com> wrote:

> I was looking at some of the PRs for DIPs and was wondering how to go ahead and use them in my code. A few of the forum regulars have mentioned applying their own patches before, and so I would like to ask those of you that do, what your preferred method is, or if this is an all-around bad idea and I should just wait for the PRs to be merged?

as for me, i have a build script for that. my directory structure looks like this:

dmd_git
 dmd      # cloned git repo of dmd
 druntime # clone git of runtime
 phobos   # you got the idea
 tools    # and this too
 BUILD    # update and build scripts
 PATCHES  # here they all!
  dmd
  druntime
  phobos
  tools

build script in BUILD/ does `git reset --hard ; git clean -dxf` for all repos, and then applies *.patch files found in corresponding subdirs of PATHCES to corresponding components. it also does alphanumeric sorting of patch files, so i can control the order of applying.

to temporarily disable some patch i can rename it to something like "abc.patch.no" (any extension, actually).

also i'm adding some comments and dependencies to patch files. patch utility ignores that.

my build script accept some args too:
% ./build.sh -h
--noreset  -- don't do 'git reset --hard'
--nopatch  -- don't do patching
--phobos   -- build phobos
--runtime  -- build runtime
--dmd      -- build dmd
--debug    -- enable debug build
--all      -- build all (default)
--cdgc     -- build with CDGC

this is not the best way to manage patches, it simply emerges from my simplistic build scripts as i was adding features over time.

i also has 'install.sh' script, which installs fresh DMD into my local binary dir, writes proper dmd.conf, builds tools and copies them to the place where DMD binary is.

and 'update.sh' script, which updates all components and writes update logs. update logs are just a snippets of 'git log' -- from the new HEAD till the old HEAD, so i can easily read what's new here while building fresh DMD in another terminal.

On Thursday, 25 December 2014 at 20:15:11 UTC, Vlad Levenfeld wrote:
> I was looking at some of the PRs for DIPs and was wondering how
> to go ahead and use them in my code. A few of the forum regulars
> have mentioned applying their own patches before, and so I would
> like to ask those of you that do, what your preferred method is,
> or if this is an all-around bad idea and I should just wait for
> the PRs to be merged?

There is https://github.com/CyberShadow/Digger which will automate the process for you.

Thanks, guys! Custom build scripts are beyond me at the moment (I
just use posix-make-install script from dmd website, at most I
will throw in -g to check out what dmd does behind the scenes) so
I'll give Digger a shot.