Re: [Tinycc-devel] Error when compiling with tcc

[Somchai Smythe]

I'm sure this has probably come up before, but I really don't
understand the policy of no release, not ever, no matter what.

So why does tcc _NEVER_ release new versions any more?  Is it
uncertainty whether the post-release code contain pieces that aren't
properly licensed or that are patent infringing or something like
that?  Or is there something where maintainers feel that until you
reach C17 compliance, you won't release?

On the mailing list, every time someone reports an error using the
last official release, they are told to use the mob branch, because it
has fixes.  Years of fixes.  Years of fixes the maintainers refuse to
package and release.

So why not make a new tcc release based on the mob branch once every 6
months or so?  That way people aren't years behind, and distributions
that only package releases will not package the last release version
which even it's own developers have deemed too broken to use.

And I don't mean some snapshot, I mean an actual release.  It seems
silly to have a compiler that many people use where running the
release version is _ACTIVELY_ discouraged by the developers of the
compiler itself.   And how do you convince a supervisor to consider a
compiler even it's own developers have decided is forever-beta?

I'm not wanting to start a flame war, I really just want to know why
the maintainer(s) won't ever release a new version.

On 1/16/20, Michael Matz <address@hidden> wrote:
> Hello,
>
> On Tue, 14 Jan 2020, Someone who expects to use a released version wrote:
>
>> Hey all,
>> works with every other compiler I've tried, including GCC, clang, MSVC,
>> MinGW, etc.)
>>
>> When I attempt to compile, I get
>>
>> tcc: x86_64-gen.c:421: load: Assertion `((ft & 0x000f) == 0) || ((ft &
>> 0x000f) == 12) || ((ft & 0x000f) == 4) || ((ft & 0x000f) == 5) || ((ft &
>> 0x000f) == 6)' failed.
>
> You use a TCC compiler from before 2017-Jul-9, i.e. you miss 2.5 years of
> bugfixes.  Please reproduce with the current mob branch ...
>
>> I'm not sure exactly what this means, it seems to be something failing
>> inside the compiler. Any insights would be appreciated.
>>
>> I can provide more details/link to the project if anyone wants to have a
>> look at it.
>
> ... and if it's then still a problem, then yes, more details will be
> required.
>
>
> Ciao,
> Michael.