Re: can include really compile in any order ? let's see

2015-10-20 Thread Michael Stahl
On 15.10.2015 21:45, Norbert Thiebaud wrote: > to test that I wrote a little scritp to do a 'naive' compile a dummy > c++ program that include > each and every 'public' include of the project one at the time... btw, "make iwyudummy.generate && make iwyudummy" should do that already (though i forge

Re: can include really compile in any order ? let's see

2015-10-16 Thread Norbert Thiebaud
On Fri, Oct 16, 2015 at 5:20 AM, Miklos Vajna wrote: > Hi Norbert, > > On Thu, Oct 15, 2015 at 02:45:52PM -0500, Norbert Thiebaud > wrote: > > In any case, thanks for the experiment, now we have numbers that also > say that the current situation is sub-optimal. ;-) Actually I made a mistake in

Re: can include really compile in any order ? let's see

2015-10-16 Thread Norbert Thiebaud
On Fri, Oct 16, 2015 at 7:54 AM, Stephan Bergmann wrote: > On 10/15/2015 09:45 PM, Norbert Thiebaud wrote: > > > That sounds odd; or is that just follow-up errors, and the real error being > an unknown or incomplete base class? yes you are right... these are follow-up errrors. they dominate the o

Re: can include really compile in any order ? let's see

2015-10-16 Thread Stephan Bergmann
On 10/15/2015 09:45 PM, Norbert Thiebaud wrote: The bulk of the errors are O[U]String[Buffer|Hash] related The OUString thing can be mitigated by adding some stuff before the include namespace rtl { class OUString; } using rtl::OUString; namespace rtl { class OUStringBuffer; } using rtl::OUStri

Re: can include really compile in any order ? let's see

2015-10-16 Thread Stephan Bergmann
On 10/15/2015 09:45 PM, Norbert Thiebaud wrote: The bulk of the errors are O[U]String[Buffer|Hash] related The OUString thing can be mitigated by adding some stuff before the include namespace rtl { class OUString; } using rtl::OUString; namespace rtl { class OUStringBuffer; } using rtl::OUStri

Re: can include really compile in any order ? let's see

2015-10-16 Thread Miklos Vajna
Hi Norbert, On Thu, Oct 15, 2015 at 02:45:52PM -0500, Norbert Thiebaud wrote: > So, following the ESC call, I wondered how true is the stated goal > that include should be self contained.. iow that you can include any > of them in whaever order and it should 'work' I think that idea comes from

can include really compile in any order ? let's see

2015-10-15 Thread Norbert Thiebaud
So, following the ESC call, I wondered how true is the stated goal that include should be self contained.. iow that you can include any of them in whaever order and it should 'work' to test that I wrote a little scritp to do a 'naive' compile a dummy c++ program that include each and every 'public