So I've started work on converting everything over to smart pointers. You can check it out on my github here: https://github.com/rm5248/log4cxx-testing/tree/smart_pointers
Some things to note so far:
Overall, everything is going somewhat smoothly. Also note that the changes I am doing probably break API compatability, but I'm not sure how bad it will be at the moment. Given that ObjectPtr overrode operator-> and shared_ptr does as well, it seems that most of this will be transparent to other code(this is why I was able to pretty quickly change the LOG4CXX_PTR_DEF macro and have things almost work).
On Sun, Nov 20, 2016 at 9:39 AM, Thorsten Schöning <[hidden email]> wrote:
Guten Tag Robert Middleton,
It depends on how you want to merge it in. If you don't care about the history, just do a 'git diff' to whatever the first commit was, which will essentially squash everything into one diff that you can apply via patch.
The only thing that it doesn't do is run checks to make sure that either the compiler supports C++11, or if boost is installed; that's the dirty part of it. Ideally it should check to see what is installed and fail if nothing is found in the configure phase.
Success! log4cxx now compiles, but it almost certainly doesn't work yet. I need to do some testing on it.
Things that need to be changed:
Flags that the Linux kernel uses(note C not C++, but most of these are still applicable). This is not an exhaustive list and is a subset of -Wall, since -Wall can produce a lot of output that is not always helpful.
This is from the main Makefile.
If anybody wants to try it out, the commit that it works on(without style changes) is this one(I accidentally committed binaries, so this commit simply removes them): https://github.com/rm5248/log4cxx-testing/commit/10f8fdd99bde2410ee92fefb46a414a5bf12669a
Otherwise, the commit with all of the formatting changes is here: https://github.com/rm5248/log4cxx-testing/commit/6c5e3f3ba7075fbe7d82001e90510e4ed9c19029
On Sun, Nov 20, 2016 at 12:10 PM, Robert Middleton <[hidden email]> wrote:
I'm not sure what the status is of log4cxx at the moment, but progress has been made on my branch since the last time I posted. I have fixed all of the problems that the unit tests found once I converted everything over to smart pointers. Fortunately, there are a large number of unit tests that cover all aspects of the code, so the large majority of issues should have been fixed as of this point. This still requires testing on actual user programs though.
On Wed, Nov 23, 2016 at 6:43 PM, Robert Middleton <[hidden email]> wrote:
I can make a new fork and do the PR. There are still a few things
that I need to check though; I think I figured out why the original
code had the special log4cxx casting macros. Apparently the
dynamic_cast doesn't always work across shared library boundaries, so
that could cause a problem. Some of the code is also a little
inconsistent at the moment and should be more standardized(I changed
the casting about halfway through).
I would like to do this after an actual release though, as since it
touches pretty much everything I wouldn't want to break code for
anybody who is using a current SVN version, since there are a lot of
fixes for current bugs in SVN(well, git now).
On Tue, Apr 11, 2017 at 8:24 AM, Thorsten Schöning
<[hidden email]> wrote:
> Guten Tag Robert Middleton,
> am Freitag, 20. Januar 2017 um 03:41 schrieben Sie:
>> git link:
> Now that we have a GIT repo, how do we deal with your branch? Are you
> able to fork the new GitHub mirror and apply your changes easily again
> in the fork? From my understanding you are only afterwards able to
> create a PR on GitHub.
> Alternatively I need to add your repo as a remote in my repo and
> merge directly I guess?
> I think a somewhat "official" PR from a fork of you against the new
> mirror would be cleanest solution.
> Mit freundlichen Grüßen,
> Thorsten Schöning
> Thorsten Schöning E-Mail: [hidden email]
> AM-SoFT IT-Systeme http://www.AM-SoFT.de/
> Telefon...........05151- 9468- 55
> Fax...............05151- 9468- 88
> Mobil..............0178-8 9468- 04
> AM-SoFT GmbH IT-Systeme, Brandenburger Str. 7c, 31789 Hameln
> AG Hannover HRB 207 694 - Geschäftsführer: Andreas Muchow
|Free forum by Nabble||Edit this page|