cURL / Mailing Lists / curl-library / Single Mail


RE: High number of conversion warnings compiling for 64bit Windows using MinGW

From: Steve Holme <>
Date: Tue, 6 Jan 2015 00:15:24 +0000

Hiya Marc,

On Mon, 5 Jan 2015, Marc Hörsken wrote:

> > Yes - to fix 'em ;-)
> okay, so I will just go ahead and try to fix some of them one after
> another. ;-)

If we can - I think it would be worth it.

If we are doing things like "int..." or "long len = strlen(string);" then we should be using the right type.

If these sort of warnings are as a result of the values being passed into curl functions that accept "long" parameters, for example via curl_easy_setopt(), then we can use the warnless functions ;-)

> > In all seriousness, I've not looked at these particular issues yet -
> > I've been trying to minimise the warnings in the main curl and
> > libcurl source code and made some good progress over the last
> > few months ;-)
> Thank you very much, I was mostly focused on fixing issues
> detected by the clang or VS static code analyzers. ;-)

I've been focused on reducing the warnings in the auto builds myself, in particularly those from Dan F's builds, and some that I can't remember now, as well as the recent spike in warnings from smb.c. Next on my list was to try and compile the test server (under Visual Studio if I can) and look at those, then the libcurl unit test and finally the examples but any assistance would be great as I don't know how much work it will be to get the test server into the Visual Studio project files will be yet.

> > However, most of the issues you listed seem to be example source
> > test server (sws etc...) and libcurl unit tests - is that correct?
> Yes, it looks like that.

It may be a bit late in the day to try and address them before the release, but any reduction is good in my book, so I will leave that judgement with you and for anyone else to comment on ;-)

Kind Regards


List admin:
Received on 2015-01-06