Lots of good suggestions here. Please break them down into specific fixes.

Thanks. Which commit you want to separate?

Each commit should be simple and focused on a specific issue. Just scanning the comments, it looks like most can be separate commits with those about compilation warnings grouped together and those about ^Z grouped together.

Look at it from my point of view. I really appreciate the help. The more you can make the changes in bite sized, related pieces the better for you and for me!

I separeted some commit. And I noticed usage of CharUpper is not wrong. So removed commits about CharUpper from this PR.

If you want to squash commits for the groups, I'll do it.

Thanks; I'll go over the changes again.

Hi @mattn, I took a bunch of the fixes from your PR that were not specific to mingw.

Can you now resubmit your PR with just the remaining changes -- which presumably are just those for mingw?

Thanks.

Sorry, GitHub merge tool break encoding of source code. I'll fix in later.

Hmm, you've better to use cherry-pick to avoid conflicting. Also the commit you did lost my contributions.

Now this is remaining diffs.

Hi @mattn,

I need you to take a closer look at some things:

1. Why might PARTIALKEY be defined?
2. Why might GUID_DEFINED be defined?
3. the changes to treectl.c are no longer needed since I changed the type of i.
4. the change to wfdrives.c and wfinit.c to include shlobj.h are redundant.
5. the change to wfdirrd.c is wrong since I changed the type of the second parameter to AppendToPath.

Thanks.

1, Why might PARTIALKEY be defined?

Recent compiler set have HELP_PARTIALKEY in winuser.h

Why might GUID_DEFINED be defined?

DEFINE_GUID is defined in another implementation in guidef.h on mingw compiler. So need to avoid re-define.

the change to wfdrives.c and wfinit.c to include shlobj.h are redundant.

No. IsNetDrive is not defined without shlobj.h

Now reverted some changes in this PR to match master branch. But the changes you did is changes I did. I say again. You had to use git cherry-pick.

Thanks; all your changes are now merged.

Please revert the change b51bf08. That is commit I did not you.

Correct. I cherry-picked your changes and committed some of them.

Let me know specifically what changes I committed that do not meet your needs.

Why I don't listed in here https://github.com/Microsoft/winfile/graphs/contributors ? Because you didn't cherry-pick correctly.

445ea8e and a06f6f1, b51bf08 are commits I did.

I didn't cherry-pick them. I used a merge --squash and reformatted the commit comment. I guess that didn't get you the correct attribution. Sorry about that.

Well, in any case your changes are part of the project. I appreciate your help!

@craigwi, you should revert b51bf08; otherwise you stole code @mattn wrote.

Commit history must be like this or this.
You can cherry-pick them by:

git checkout master
git remote add tmpremote git@github.com:wataash/winfile.git
git fetch tmpremote rewrite-author
git cherry-pick master..rewrite-author
git push
git remote remove tmpremote

I feel caught in a bind because I want both high quality commits to a very complex code base and definitely wanting to provide attribution to all who contribute. Help me solve that.

@craigwi Sorry for confusing. I do not want to say opinions any more because I will be seem I'm talking about childishly. I value the respect for contributors in OSS. For example, I'll keep author name in commit even if I must make squash the commit. If GitHub is gone and I must move my repositories into another VCS, I'll move commit history into the VCS to remain the information of contributions. I feel that this is a difference in respect to contributions. At least if you instructed me, I might make same commit as your commit. If you feel that I don't have the ability to do it, please close this PR.

@craigwi Which commits do you want to squash? If you want, I'll separate this PR to several PRs.

As I feebly tried to say above, I really, really, really want you all to get the credit for changes. I've been working on WinFile for 10 years and personally pushed to get this project out so you could all benefit. I have enough credit to my name here.

In the end, I will:

• make sure the author attribution is correct, no matter how the commit happens
• work more closely with willing authors to shape PR before integration

Given that I already screwed up the commit history, I made one last change so that the most recent commit is authored by Matthew and that he gets the credit on the contributors board.

Again, I want you folks to get the credit and want us to have a great, high quality result.

Sincerely,
Craig.

@craigwi Thank you for the faithful and wonderful your work. About twenty years ago, I often used WinFile on Windows 3.1. And it worked very well and charms. I believe that your work was great. Thank you. We are not doubting you. Just simple thing, We hope to share the pleasure of being merge of commits from contributors. So let's make a constructive discuss.

@craigwi I'll separete this PR to several PRs. If you like single commit, please let us known, I'll squash several commits in the PR before you will merge.

For about separaing PRs, How about this?

• fix syntax of res.rc
• remove compilation warnings
• add Makefile.mingw
• fix build errors related re-define macros.

@craigwi
Thank you for your kind cooperation :)

Please keep in mind that git has the concept of "author" and "committer". The committer is just a proxy for a commit, while the author should be kept as the original one.

If the commits in PRs have so good quality that they don't need to be squashed, I recommend merging them by "Create a merge commit" or "Rebase and merge" on GitHub UI. But please don't use "Squash and merge". I wrote the details in here.

If you need to squash commits, I recommend the following procedure:

git checkout master
git branch master-backup
git fetch origin pull/42/head:tmp
git reset --hard tmp
git rebase -i master-backup
git branch -d master-backup
git branch -D tmp

I wrote more helps in here.

While this procedure, we often get confused -- what is the branch status? and where am I? (where is HEAD?) So I strongly recommend using a commit-tree visualizing tool like Sourcetree (or, of course, git log --graph if you think it's enough).

If you mistakenly rewrite the author, rewrite it back again by git commit --author='Author Name <author@example.com>' --amend.

Now this PR is separated to #93 and #94.