minutes of ESC call ...

[Norbert Thiebaud]

On Mon, Nov 11, 2013 at 5:46 AM, Stephan Bergmann <sbergman at redhat.com> wrote:
> On 11/11/2013 12:38 PM, Norbert Thiebaud wrote:
>>
>> On Mon, Nov 11, 2013 at 5:24 AM, Stephan Bergmann <sbergman at redhat.com>
>> wrote:
>>>
>>> Can you point me to such a case in our code base where replacing "..."
>>> with
>>> <...> would break the build?
>>
>>
>> basctl/source/inc/basidesh.hxx:#include "../basicide/objdlg.hxx"
>>
>> $ cat basctl/source/basicide/objdlg.hxx | grep include
>> #include "bastype2.hxx"
>> #include "bastypes.hxx"
>> ...
>>
>> git grep basidesh\.hxx
>> [...]
>> basctl/source/basicide/objdlg.cxx:#include "basidesh.hxx"
>> basctl/source/dlged/dlged.cxx:#include "basidesh.hxx"
>> [...]
>
>
> OK, that looks like something that begs for clean-up anyway (moving
> objdlg.hxx and bastype2.hxx from basctl/source/basicide/ to
> basctl/source/inc/).

sure, sure... but that is just one of many... took me a whole minute
to find it blind...
I guess my point is that you _could_ do a script to replace stuff that
are in the form
include "\([a-zA-Z]*/[^"]*\)"] to include <\1>
and that is very,very likely to just work

but scripting the replacement of the other "" include... will be less
successful, and require quite a bit of manual tweak and clean-up...
which should really not be manually swept into a mass
pseudo-script-generated commit.

Norbert