Robert's script bisects over binaries, while git bisects over source.
Since our tree takes so long to build, git bisect hasn't been too
helpful in my experience.

On Fri, Jun 5, 2009 at 10:56 PM, Jeremy Moskovich<[email protected]> wrote:
> That's great and I don't mean to belittle the awesomeness of this in any way
> but:
> git has built-in support for this along with many other advantages over SVN,
> I use git along with the gitx gui on OSX and it's greatly increased my
> productivity.
> It would make me much happier if more people on the team tried using it,
> there are instructions on the wiki and I doubt many people will go back to
> using svn after trying it.
> (credit for git integration with our infrastructure goes to Evan and a bunch
> of other awesome people).
> Best regards,
> Jeremy
>
> On Fri, Jun 5, 2009 at 9:48 PM, Robert Sesek <[email protected]> wrote:
>>
>> Tonight I wrote a Python script called chrome-bisect.py that performs
>> a binary search on the Mac continuous build archive. It takes a
>> starting known-good and bad revision and then downloads, unzips, and
>> opens the archived builds for you. You can then test the running build
>> and report back to chrome-bisect if the build is good or bad. The
>> binary search will then continue until it finds the version in which
>> the regression was introduced.
>>
>> You can download it here:
>> http://www.bluestatic.org/git/?p=Toolchain.git;a=blob_plain;f=chrome-bisect.py;hb=HEAD
>>
>> The script will not work on Windows (it should work on Linux, for the
>> most part) because it uses `unzip` and `open` commands via os.system
>> (). Patches are welcome to make it more cross-platform. You can clone
>> the git repository it lives in here:
>> http://www.bluestatic.org/git/Toolchain.git
>>
>> - Robert Sesek
>>
>>
>
>
> >
>

--~--~---------~--~----~------------~-------~--~----~
Chromium Developers mailing list: [email protected] 
View archives, change email options, or unsubscribe: 
    http://groups.google.com/group/chromium-dev
-~----------~----~----~----~------~----~------~--~---

Reply via email to