This is the mail archive of the mailing list for the GDB project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH 22/348] Fix -Wsahdow warnings

On Tue, Nov 22, 2011 at 10:27 PM, Mike Frysinger <> wrote:
> On Tuesday 22 November 2011 09:04:07 Andrey Smirnov wrote:
>> On Tue, Nov 22, 2011 at 8:35 PM, Marek Polacek <> wrote:
>> > On 11/22/2011 02:33 PM, Andrey Smirnov wrote:
>> >>> Is the typo intentional?
>> >>
>> >> Sorry about that. No it isn't. Please ignore that particular patch I'll
>> >> send corrected version soon(I'm in the middle of a git rebase
>> >> --interactive, so I can't edit that particular commit just yet).
>> >
>> > That typo is in other patches too.
>> OK, once again, sorry about that, I'll recheck all the patches related
>> to bcache.c, for now please ignore those.
> please condense down your patches if you resend. ?there's way too many little
> tiny ones that really should be squashed into a single changeset.

Initially, there were 17 patches, which, upon suggestion from Tom
Tromey, I split so that every patch contain only changes to one
particular function or some other small unit of the source code. I
tend to agree with Tom that my initial decision to make only 17
patches made it rather hard to review each, because every one of them
contained many small but disparate changes.

Squashing or splitting commits is not really a problem and I can do
this, but if you want me to do so, than please point out the patches
I should squash together.

Right now I have 258 patches to which I have yet to write a ChangeLog
entry and 100 patches whose ChangeLog I have to change to, as you
pointed out, conform to GNU policy. There are also patches that I
screwed up with typos and patches that will eventually will have to be

Squashing all commits based on file level will still leave you with
something like 150 patches, but some of them would be quite large and
harder to review than they are now. Doing so on API level would
require me to go through all the patches and relative source code,
figure out to what API every change belong(which I'll probably do wrong
because I do not yet have a very good grasp on GDB's internals) and
split and squash all commits accordingly.

So given the aforementioned amount of work, can't we ignore that the
patch count is over 9000?

Andrey Smirnov

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]