[prev in list] [next in list] [prev in thread] [next in thread] 

List:       koffice-devel
Subject:    Re: koffice/kchart/shape
From:       Cyrille Berger <cberger () cberger ! net>
Date:       2009-08-21 9:29:14
Message-ID: 200908211129.14736.cberger () cberger ! net
[Download RAW message or body]

[Attachment #2 (multipart/alternative)]


On Friday 21 August 2009, Jaroslaw S wrote:
> That would be a solution.
> Or another solution would be to allow for both approaches.
> I still encourage to use Q_UNUSED() and adding includes in advance.
> One reason is that we add library interdependencies in advance this way.
Not sure if I see a point to that. Anyway, we can just have the policy of 
using Q_UNUSED with #include or /* variable */, it's just that the first one is 
bond to cause regular breakage on windows, so we might want to recommand the 
second one.

> Is including headers so costly these days? I hope not, especially if
> we're talking about adding them to .cpp files.
Parsing a header is reasonnably cheap, but it might trigger a recompilation 
when the header change (which in case of a qt/kde header isn't much of a 
problem, since anyway it's unlikely only header changed, but for koffice header 
it is different). That the file is only included in the cpp, just mean that 
there isn't a cascading effect like what you have with unused header included 
in hpp.

> Moreover, not using the C-like /* variable commenting */ can be
> perhaps more friendly for Doxygen.
We are speaking cpp file, and the KDE/KOffice policy is doxygen belongs to 
headers.

-- 
Cyrille Berger

[Attachment #5 (text/html)]

<!DOCTYPE HTML PUBLIC "-//W3C//DTD HTML 4.0//EN" \
"http://www.w3.org/TR/REC-html40/strict.dtd"><html><head><meta name="qrichtext" \
content="1" /><style type="text/css">p, li { white-space: pre-wrap; \
}</style></head><body style=" font-family:'DejaVu Sans'; font-size:9pt; \
font-weight:400; font-style:normal;">On Friday 21 August 2009, Jaroslaw S wrote:<br> \
&gt; That would be a solution.<br> &gt; Or another solution would be to allow for \
both approaches.<br> &gt; I still encourage to use Q_UNUSED() and adding includes in \
advance.<br> &gt; One reason is that we add library interdependencies in advance this \
way.<br> Not sure if I see a point to that. Anyway, we can just have the policy of \
using Q_UNUSED with #include or /* variable */, it's just that the first one is bond \
to cause regular breakage on windows, so we might want to recommand the second \
one.<br> <p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; \
margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; \
-qt-user-state:0;"><br></p>&gt; Is including headers so costly these days? I hope \
not, especially if<br> &gt; we're talking about adding them to .cpp files.<br>
Parsing a header is reasonnably cheap, but it might trigger a recompilation when the \
header change (which in case of a qt/kde header isn't much of a problem, since anyway \
it's unlikely only header changed, but for koffice header it is different). That the \
file is only included in the cpp, just mean that there isn't a cascading effect like \
what you have with unused header included in hpp.<br> <p \
style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; margin-left:0px; \
margin-right:0px; -qt-block-indent:0; text-indent:0px; \
-qt-user-state:0;"><br></p>&gt; Moreover, not using the C-like /* variable commenting \
*/ can be<br> &gt; perhaps more friendly for Doxygen.<br>
We are speaking cpp file, and the KDE/KOffice policy is doxygen belongs to \
headers.<br> <p style="-qt-paragraph-type:empty; margin-top:0px; margin-bottom:0px; \
margin-left:0px; margin-right:0px; -qt-block-indent:0; text-indent:0px; \
-qt-user-state:0;"><br></p>-- <br> Cyrille Berger</p></body></html>



_______________________________________________
koffice-devel mailing list
koffice-devel@kde.org
https://mail.kde.org/mailman/listinfo/koffice-devel


[prev in list] [next in list] [prev in thread] [next in thread] 

Configure | About | News | Add a list | Sponsored by KoreLogic