On 2/29/20 9:10 PM, Scott Kostyshak wrote:
> On Sat, Feb 29, 2020 at 02:56:19PM -0500, Richard Kimberly Heck wrote:
>> On 2/29/20 12:38 PM, Scott Kostyshak wrote:
>>> On Sat, Feb 29, 2020 at 05:45:29AM +0100, Richard Kimberly Heck wrote:
>>>> commit 4460df329226ab5e72adab0ff764b7496022bbbd
>>>> Author: Richard Kimberly Heck <rikih...@lyx.org>
>>>> Date: Fri Feb 28 23:52:06 2020 -0500
>>>>
>>>> Fix header.
>>>> ---
>>>> src/support/Systemcall.h | 2 +-
>>>> 1 files changed, 1 insertions(+), 1 deletions(-)
>>>>
>>>> diff --git a/src/support/Systemcall.h b/src/support/Systemcall.h
>>>> index d3d2189..537a701 100644
>>>> --- a/src/support/Systemcall.h
>>>> +++ b/src/support/Systemcall.h
>>>> @@ -15,7 +15,7 @@
>>>> #ifndef SYSTEMCALL_H
>>>> #define SYSTEMCALL_H
>>>>
>>>> -#include <string>
>>>> +#include "strfwd.h"
>>>>
>>>> namespace lyx {
>>>> namespace support {
>>>> --
>>>> lyx-cvs mailing list
>>>> lyx-...@lists.lyx.org
>>>> http://lists.lyx.org/mailman/listinfo/lyx-cvs
>>> What was broken? Did it produce a warning?
>> We generally try not to include <string> in header files.
> Why not? I read through git log strfwd.h to try to answer this question
> myself but couldn't figure it out. I also tried reading through the file
> itself but could not understand why we prefer including this header over
> <string> if the purpose is to use std::string (as seems the case in
> src/support/Systemcall.h)
The string.h file here is 18K, i.e., big for a header file. So including
it could add 18K to the size of what we have to compile whenever
Systemcall.h is included. Probably in lots of those cases, string.h has
been included as well (and, of course, it will have to be included one
way or another in Systemcall.cpp). But my understanding, anyway, is that
the intent was to reduce the size of includes where possible. André was
really big on this for a while.
Riki
--
lyx-devel mailing list
lyx-devel@lists.lyx.org
http://lists.lyx.org/mailman/listinfo/lyx-devel
