[Proj] Code style in Proj

Kristian Evers kreve at sdfe.dk
Mon Apr 30 10:44:45 PDT 2018 

I briefly searched the C89 draft for occurrences of “compound statement” and “block” and couldn’t find a anywhere
where it says that stand-alone blocks are not allowed. I think [0] is the most relevant section for this discussion.

/Kristian

[0] https://port70.net/~nsz/c/c89/c89-draft.html#3.6.2

On 30 Apr 2018, at 16:22, Thomas Knudsen <knudsen.thomas at gmail.com<mailto:knudsen.thomas at gmail.com>> wrote:

Yes, I understand the intention, and have used the same kind of constructs myself earlier. I stopped using it roughly 20 years ago, since my interpretation of the standard text led me to conclude that the construct was not complying.

Since your tests indicate that my interpretation was wrong, I have googled around a bit to see what other people think, and it actually appears that I was wrong. The alternative exegesis of the gospel is given by a number of commentators on [0], so I suppose I was wrong - and will unfortunately now have to reread the standard to be sure :-)

/Thomas

[0] https://stackoverflow.com/questions/7859424/why-was-mixing-declarations-and-code-forbidden-up-until-c99


2018-04-30 15:57 GMT+02:00 Kristian Evers <kreve at sdfe.dk<mailto:kreve at sdfe.dk>>:
Thomas,

Actually this was part of Kurts original proposal. The else is not missing here as you can see from the original code [0].
The code passed my compiler just fine with all the C89 specific flags turned on (as mentioned in an earlier email in
this thread). So here it is just a block introduced to create a scope where declarations can be made more local to
where they are used. If that is prohibited according to the standard I don’t know. I don’t have other ways to test
standard compliance than to compile the code with C89 flags turned on. For a more thorough test we can create a
pull request with the proposed changes and see how the various compilers on Travis and AppVeyor fares.

/Kristian

[0] https://github.com/OSGeo/proj.4/blob/d0fefa4104d9b655d59e400cda616f0b4d407071/src/PJ_aeqd.c#L110-L126


On 30 Apr 2018, at 15:44, Thomas Knudsen <knudsen.thomas at gmail.com<mailto:knudsen.thomas at gmail.com>> wrote:

I think line 116 in Kristian's revised version misses an "else": Gnu C, and probably many other compilers, accepts "naked" compound statements, but I do not think it is in accordance with the C89 standard

2018-04-30 15:00 GMT+02:00 Kristian Evers <kreve at sdfe.dk<mailto:kreve at sdfe.dk>>:
Kurt,

I think it looks good. I tried compiling the code and again I had the problem with declarations like

XY xy = {0.0, 0.0};

I’ve fixed those in a fork of your gist. See revision 2 in [0] for a proposed solution that I think
is in line with your ideas

I like how you’ve separated includes from the standard library on the PROJ internal includes. On
a related note (for those who don’t already know) I have recently introduced the proj_math.h
header which is intended as a PROJ replacement of math.h. proj_math.h’s purpose in life is to
make sure that we can use C99 math functions on systems where they are not available.
proj_math.h includes math.h and defines functions such as isnan() and hypot(). This way they
can be used as you would on a C99-compatible system as long as proj_math.h is included.
See revision 3 in [0].

Would you be willing to write up a simple style guide for PROJ based on the ideas you have presented in
this thread?

/Kristian

[0] https://gist.github.com/kbevers/17ec703942646e401b28a0975b28771d/revisions

On 24 Apr 2018, at 12:05, Mateusz Loskot <mateusz at loskot.net<mailto:mateusz at loskot.net>> wrote:

On 24 April 2018 at 11:45, Thomas Knudsen <knudsen.thomas at gmail.com<mailto:knudsen.thomas at gmail.com>> wrote:
In C too, but since C99.

AFAIK this only applies to the "for (int i = 0;  i < n; i++)" syntax.

while (...) {
   int i = 123, j = 345;
   k = i + j;
}

has been correct (and imho, preferable) syntax since K&R C

AFAIK, it's about compound statement.
C89 requires declaration at the top of blocks, while C99 lifts that:

$ cat test.c
int main()
{
   int i = 0;
   while (i < 10)
   {
       int j = 345;
       i = i + j;
   }

   {
       int z = 0;
       z = 1;
   }
   int x = 1;
}

$ gcc -std=c89 -pedantic test.c
test.c: In function ‘main’:
test.c:14:5: warning: ISO C90 forbids mixed declarations and code
[-Wdeclaration-after-statement]
    int x = 1;
    ^~~

$ gcc -std=c99 -pedantic test.c
$


Best regards,
--
Mateusz Loskot, http://mateusz.loskot.net<http://mateusz.loskot.net/>
_______________________________________________
Proj mailing list
Proj at lists.maptools.org<mailto:Proj at lists.maptools.org>
http://lists.maptools.org/mailman/listinfo/proj


_______________________________________________
Proj mailing list
Proj at lists.maptools.org<mailto:Proj at lists.maptools.org>
http://lists.maptools.org/mailman/listinfo/proj

_______________________________________________
Proj mailing list
Proj at lists.maptools.org<mailto:Proj at lists.maptools.org>
http://lists.maptools.org/mailman/listinfo/proj


_______________________________________________
Proj mailing list
Proj at lists.maptools.org<mailto:Proj at lists.maptools.org>
http://lists.maptools.org/mailman/listinfo/proj

_______________________________________________
Proj mailing list
Proj at lists.maptools.org<mailto:Proj at lists.maptools.org>
http://lists.maptools.org/mailman/listinfo/proj

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.osgeo.org/pipermail/proj/attachments/20180430/c8380ca9/attachment.html>

More information about the Proj mailing list