What versions are bug fixes implemented in ?

[440bx]

Hello,

On Mantis I see that most bug fixes state that it is fixed in v3.3.1.

I was wondering if the bugs are also fixed in v3.2 or will v3.2 be released with bugs that have only been fixed in v3.3.1 ?

Here is a "random" thought for the developers.  I think it would be practical (and desirable too) to "feature freeze" one version of FPC and only apply bug fixes (and possibly optimizations) to it.  That way, there would be one version whose goal is simply to be as bug-free as possible and, if optimizing is included, as high performance as possible.

Thank you for your help.

[rvk]

Here is a "random" thought for the developers.  I think it would be practical (and desirable too) to "feature freeze" one version of FPC and only apply bug fixes (and possibly optimizations) to it.  That way, there would be one version whose goal is simply to be as bug-free as possible and, if optimizing is included, as high performance as possible.

Isn't that what's already being done with the fixes branche?

There already is a Fixes 3.2.
Some fixes to 3.3.1 will also be applied to Fixes 3.2 if they are important enough.

[440bx]

Isn't that what's already being done with the fixes branche?

I don't know the answer to that, that's why I'm asking.  I'd like one (or more) of the FPC developers to make it clear.

There already is a Fixes 3.2.
Some fixes to 3.3.1 will also be applied to Fixes 3.2 if they are important enough.

Aren't all bugs "important" ?  your use of the word "some", gives the impression that there will be known bugs that won't be fixed in v3.2.  I'd like one of the developers to confirm that.

[Bart]

When fixes 3.2 is tagged, there is a feature freeze for the 3.2 branch.
All new features and improvements go into trunk (3.3.1).
Only bugfixes get merged into 3.2.

It's the same for Lazarus.

Bart

[440bx]

Only bugfixes get merged into 3.2.

Thank you Bart.  I read that to mean that all reported bugs are fixed in v3.2 and v3.3.1 until it is decided to "freeze and release" v3.2.  Can you confirm that I've understood this correctly ?

Thanks!.

[Bart]

At some point it is decide that trunk has matured enough (after the last release) to warrant a next release.
At this point the new release is tagged: 3.2 in this case.
Nothing is release yet.
Trunk then gets 3.3.1 (was 3.1.1).

As of that moment there is a feature freeze on the 3.2 (3.2 fixes branch).
All bugs are fixed in trunk (any bug that is in 3.2 by definition also is in trunk at the moment 3.2 was tagged).

Now trunk will develop further and have bugs on features that are not in 3.2 (fixes).
Patches are merged to 3.2 if the issue is present in 3.2 as well (unless there is some reason not to).

Clear enough?

Bart

[440bx]

Clear enough?

Bart

yes! thank you.

I can't wait for v3.2.

[marcov]

Note that the fields of the bugtracker are not really an indication of what is going to be merged or not.

Like Bart says, the merge decision is more about features vs fixes, and the intrusiveness of the fixes (IOW the likeliness that something else breaks).

Isolated fixes (like your windows header fixes) are much more likely to be merged than whole rearrangements of the compiler.

[440bx]

Isolated fixes (like your windows header fixes) are much more likely to be merged than whole rearrangements of the compiler.

Ok... I guess that means the possibility exists that a bug that was found in v3.0.4 is fixed only in v3.3.1 if the fix required modifications that were considered too "extensive" to be "back ported" into v3.2.   Is this right ?

[Thaddy]

Ok... I guess that means the possibility exists that a bug that was found in v3.0.4 is fixed only in v3.3.1 if the fix required modifications that were considered too "extensive" to be "back ported" into v3.2.   Is this right ?

You are ignorant.

[440bx]

Ok... I guess that means the possibility exists that a bug that was found in v3.0.4 is fixed only in v3.3.1 if the fix required modifications that were considered too "extensive" to be "back ported" into v3.2.   Is this right ?

You are ignorant.

I consider myself very lucky, I'm not like you.  see... it could be a lot worse.

[marcov]

Ok... I guess that means the possibility exists that a bug that was found in v3.0.4 is fixed only in v3.3.1 if the fix required modifications that were considered too "extensive" to be "back ported" into v3.2.   Is this right ?

"Certainly". Though to precise it will be fixed only in the version that 3.3.1 becomes when the next release branch is branched off.  Like 3.4.0 or 4.0.0.