RE: [ros-diffs] [sedwards] 19675: XML.cpp -> xml.cpp, XML.h -> xml.h.

[Casper Hornstrup]

If the rule is always lowercase filenames why not vote on it and
make it official and save ourselves some trouble in the future?

Casper

[Steven Edwards]

On 11/27/05, Casper Hornstrup <ch(a)csh-consult.dk> wrote:
>   If the rule is always lowercase filenames why not vote
> on it and
>  make it official and save ourselves some trouble in the future? 
I don't see why we need a vote for something that is logical and that
we have maintained everywhere else in the source tree. Its simply a
defacto standard. I mean you don't want to support MSVC but you want
to support gcc. Some of us that use gcc use it under Linux because of
the fact gcc sucks on windows. Be it a make problem, a gcc bug or
Windows sucking, gcc sucks on Windows and so I and others use Linux to
cross-compile. Mixed case issues and path seperator issues have always
been a problem, this is nothing new. With or without a vote if someone
can still commits a patch with it that will slip through even by
mistake.

Thanks
Steven

--
Steven Edwards - ReactOS and Wine developer

"There is one thing stronger than all the armies in the world, and
that is an idea whose time has come." - Victor Hugo

[Casper Hornstrup]

>   -----Original Message-----
>  From: ros-dev-bounces(a)reactos.org [mailto:[e-mail]] On Behalf Of
> Steven Edwards
>  Sent: 27. november 2005 22:13
>  To: ReactOS Development List
>  Subject: Re: [ros-dev] RE: [ros-diffs] [sedwards] 19675: XML.cpp -> xml.cpp,XML.h
> -> xml.h.
>  
>  On 11/27/05, Casper Hornstrup <ch(a)csh-consult.dk> wrote:
> >   If the rule is always lowercase filenames why not
> > vote on it and
> >  make it official and save ourselves some trouble in the future? 
>  
>  I don't see why we need a vote for something that is logical and that
>  we have maintained everywhere else in the source tree. Its simply a
>  defacto standard. I mean you don't want to support MSVC but you want
>  to support gcc. Some of us that use gcc use it under Linux because of
>  the fact gcc sucks on windows. Be it a make problem, a gcc bug or
>  Windows sucking, gcc sucks on Windows and so I and others use Linux to
>  cross-compile. Mixed case issues and path seperator issues have always
>  been a problem, this is nothing new. With or without a vote if someone
>  can still commits a patch with it that will slip through even by
>  mistake.
>  
>  Thanks
>  Steven 
It is not logical. See how many .rc files have upper-case first letter
for instance.

I won't actively test my changes on both compilers. I WILL help find
solutions that work on both compilers. There is a big difference.
Supporting MSVC was the reason I made rbuild support multiple backends.
I knew that if it didn't, then we would have had lots of MSVC project
files in the tree today like there currently is for explorer and
rbuild.

When it is documented in wiki, it is more easily remembered. I don't
remember these rules (logical or not) and many others don't either
apparently. Having an explanation of how to do something which can be
easily linked to, makes it easier to agree on what to do the day such
a decision is needed.

Casper

[Steven Edwards]

Hi,

On 11/27/05, Casper Hornstrup <ch(a)csh-consult.dk> wrote:
>   I won't actively test my changes on both
> compilers. I WILL help find
>  solutions that work on both compilers. There is a big difference.
>  Supporting MSVC was the reason I made rbuild support multiple backends.
>  I knew that if it didn't, then we would have had lots of MSVC project
>  files in the tree today like there currently is for explorer and
>  rbuild. 
Sorry I did not mean to sound rude I was just frustrated because I
don't think we need to get in to voting over every little issue like
this. I appreciate the fact you added support for backends and
understand that you can't test on every platform known to man. My
point was mainly that I would not care as much about the Linux build
if the msvc build worked and I am just frustrated because I think the
faster we get it done, the better of our project will be all around in
terms of new developers and the quality of our code. I expect it to be
as fast if not faster than gcc on Linux and (for me at least) mixed
case would not be a issue. As for everyone else that uses Linux to
cross-compile I still think its a good rule to have.

>   When it is documented in wiki, it is more easily
> remembered. I don't
>  remember these rules (logical or not) and many others don't either
>  apparently. Having an explanation of how to do something which can be
>  easily linked to, makes it easier to agree on what to do the day such
>  a decision is needed. 
Well we never had a firm rule "Always import filenames like this and
make sure paths are seperated like this" but then again Art, James and
myself have only been fixing problems like this as they come up for
years :) In the case of these rbuild changes its always worked for us
but annoyed the crap out of me because it did not conform to the rest
of the sources and if someone tried to edit the include statements and
or makefiles to make it conform it would have ended up with a broken
build for us.

I doubt a rule would do any good. I am for making sure we can host SIN
on Linux build boxes to catch problems like this. I think I can also
speak for everyone when I say we all do appreicate the work you have
done on rbuild, SIN, CIS and hosting. Once again sorry my comments in
the prior email were mostly off topic and rude.

--
Steven Edwards - ReactOS and Wine developer

"There is one thing stronger than all the armies in the world, and
that is an idea whose time has come." - Victor Hugo