Hello everybody,
On Jan 24, 2008 9:55 AM, Johannes Schindelin
<[EMAIL PROTECTED]> wrote:
> Your commit messages are not easily readable for me, since they
> extend well over the 76 characters that I can read with the 4-space
> indent of "git log". Please fix.
Dscho, thanks to your step-by-step guide, the two patches will follow.
On Jan 24, 2008 9:55 AM, Johannes Schindelin
<[EMAIL PROTECTED]> wrote:
> Also, you mention that gcc behaves more like VS2005 when two options
> are used; why not add them to the Makefile in the same commit?
It did not work out that well. Even though the warning were "similar"
they were different enough that I'd rather keep the Makefile as is. If
you want me to produce the most compatible source code, I can probably
do that. But given that so far I'm the only one who uses VS2005, I'll
manage that one warning about putenv.
On Jan 24, 2008 9:55 AM, Johannes Schindelin
<[EMAIL PROTECTED]> wrote:
> Also, reading the patch, I see that we had "char * name" before; I
> do not like that coding style. I'd rather use the same coding style
> as in git.git, i.e. "char *name".
Hopefully I caught all [three] cases in the menu.c; will fix the rest
on meet-and-greet basis.
On Jan 24, 2008 9:55 AM, Johannes Schindelin
<[EMAIL PROTECTED]> wrote:
> On Thu, 24 Jan 2008, Kirill wrote:
> > - ability to install for the current user only.
>
> I see that you have at least one hard coded path in there:
>
> +"PathToMsys"="c:\\download\\msysgit"
>
> That's not good. Please imitate what you find in the Makefile, and
> provide an "install-user" target as well as an "install-user.reg.in"
> instead of "install-user.reg".
Fixed in the following patches. However, how about removing those two
keys from .reg all together? We can rewrite DllRegisterServer to
- create [HKEY_CLASSES_ROOT\CLSID\
{ca586c80-7c84-4b88-8537-726724df6929}\InProcServer32] with the result
of GetModuleFileName;
- create [HKEY_CURRENT_USER\SOFTWARE\Git-Cheetah], following the same
path as qgit to find git.exe.
On Jan 24, 2008 12:56 PM, Johannes Schindelin
<[EMAIL PROTECTED]> wrote:
> Well, since I do not consider "push to mob" to be something like
> "publish", it is not too late at all to amend that commit. Here's
> how:
It worked beautifully on my local master branch. Now should I hope
that you'll do the magic with the remote mob branch [after the patches
review]?
Thanks!
--
Kirill.
