On 06/27/2016 04:14 PM, Georg Baum wrote:
> Guillaume Munch wrote:
>
>> Dear list,
>>
>>
>> Here is a patch that removes all uses of std::bind and boost::bind in
>> src/. I think this is something that we want in the long term, because
>> it makes the changed code much more readable and maintainable.
> You mean std::bind, or did I miss any remaining boost::bind?
>
>> I am conflicted about it, be cause I would like to be eventually able to
>> read and maintain things that have to do with concurrency, but I would
>> not like to introduce hard-to-debug crashes. It no longer crashes for
>> me, but I did not understand the cause of the previous crash.
>>
>> Any advice? Also, since this touches a lot of sensitive code it would
>> not hurt if someone offered to proof-read it.
> For me personally the patch is too big to proof-read it. I'd prefer to
> change simple things like
>
> - BufferStorage::iterator it = bstore.begin();
> - BufferStorage::iterator end = bstore.end();
> - for (; it != end; ++it) {
> - Buffer * buf = *it;
> + for(Buffer * buf : bstore)
>
> which are not related to removal of std::bind first.
Yes, that's a good idea.
> Apart from that I am not too familiar with lambdas yet, so for me personally
> the new code is more difficult to read than the old one ATM, and I am
> probably not the only one. Your explanation to Richard is nice, but it will
> soon be buried in email archives, and people reading LyX sources do not see
> it.
>
> I am not opposed to using lambdas (I hope I can learn soemthing from you
> here), but I believe that our developer documentation should contain a short
> motivation why they are used in LyX (you basically explained already to
> Richard how they make code more readable), and a pointer to a more detailed
> introduction to lambdas. Then it should also be possible for novices or old-
> timers like me to understand the new hot stuff.
It wouldn't hurt to put some comments in the code, too.
Richard
