Hi Corinna,

On Thu, 18 Dec 2025, Corinna Vinschen wrote:

> On Dec 18 09:08, Johannes Schindelin wrote:
> > As I have debugged Cygwin bugs quite often in the past, I can tell you
> > that it is quite frustrating to end up finding commits that leave
> > everything unclear, whose diffs are too large to find any obvious
> > bugs.
> 
> You'll come across them especially back in CVS times.  CVS didn't
> support patchsets along the lines of "one patch per problem" and the
> blessed way of CVS ChangeLog entries didn't support the notion of
> explaining at great length why a patch was necessary but rather only
> what a patch is doing where.  Whoever had the idea how a CVS ChangeLog
> entry is supposed to look like, got it wrong, and thousands of
> developers followed suit, including me.  Example:
> 
>   2015-04-22  Corinna Vinschen  <[email protected]>
> 
>       * fhandler_tty.cc (fhandler_pty_slave::fch_close_handles): Don't close
>       handles not opened via fhandler_pty_slave::fch_open_handles.
> 
> I have no idea why this was necessary at the time and we have the entire
> 20 years of history since 1995 documented this way.
> 
> And, we have a couple of years using git, in which we had to wrap our
> heads around the new-fangled way to document patches in git commit
> messages.  And even if some of our commit messages might still be
> lacking, they are much better than some other git entries in projects,
> which don't seem to have learned how useful git commit messages can
> be, if they try to describe the problem they are solving and the way
> how to solve it.  Example from FreeBSD in 2025:
> 
>   cxgbe: Stop using bus_space_tag/handle directly
> 
>   Reviewed by:    np, imp
>   Sponsored by:   [...]
>   Differential Revision:  https://reviews.freebsd.org/D53030
> 
> That's it.  They are getting better, there are actually some git commit
> messages explaining the problem they are trying to solve and how, but
> it's not standard throughout.

And let me state how grateful I am about them getting better!

> What I'm trying to say is this:
> 
> It doesn't make sense to reflect on the past.
> 
> We can do better, but for that, let's please stay professional and just
> review the patch.  Tell the patch contributor what you think is wrong
> and wait for v2. Rinse and repeat with v3, v4, ... until you're
> satisfied.
> 
> Johannes, I think we all understood what you're trying to say, but this
> is *not* how to say it.

Yes, you're right, my behavior was unacceptable. I apologize!

> Review the patch, criticize the patch (including it's commit message)
> and stick to the patch.
> 
> If you have good arguments or can point out bugs or problems, I and, I'm
> sure, Takashi and others as well, will change their patch accordingly,
> or even drop the patch if it's incorrect.  But: don't tell people *how*
> to spend their *voluntary* time.  Stick to the patch, it's what counts.
> 
> As a side-note, I lost track of the pty patches which were already GTG
> and the patches which are not.  Takashi, can you please push what's
> already not in question and make a clean restart with a new thread on
> this list?  I'm really confused, and maybe I'm not alone...

I could not pay attention to anything on the list for the past 6h because
my laptop lost its ability to connect to the internet, and I worked
frantically to fix it ever since.

My take was that Takashi's
https://inbox.sourceware.org/cygwin-patches/[email protected]/
is in a pretty decent shape. If it was up to me, I'd take it as-is (in the
interest of moving on, and taking time off over the holidays). That's what
I did in https://github.com/msys2/msys2-runtime/pull/322 and in
https://github.com/git-for-windows/msys2-runtime/pull/122: I replaced my
patches with Takashi's v4.

Ciao,
Johannes

Reply via email to