Mailing lists / fpc-devel / 2012-08
3 messages · 3 people · –
Reply by email Subscribe to this list Your subscription
Read-only archive. Use j and k to move between messages.
I have long understood that each stable release of freepascal is made to be compilable with the previous release but at a quick glance at the website I couldn't seem to spot any official statement of this. Is it indeed "supported" to compile a stable release with the previous stable release and if so is there a statement of this anywhere.
I ask because i've been helping someone in debian who is trying to convince another compiler (for a totally different language) to adopt a similar policy.
peter green wroteI have long understood that each stable release of freepascal is made to be compilable with the previous release but at a quick glance at the website I couldn't seem to spot any official statement of this. Is it indeed "supported" to compile a stable release with the previous stable release and if so is there a statement of this anywhere.
This automatically follows from the fact that the current development versions are always only guaranteed to be compilable by the previous release (which is ://www.freepascal.org/develop.var ). A new release is nothing but a snapshot of a development version, other than a different version number (all other changes are also part of the fixes/trunk branch).
Joans
I have long understood that each stable release of freepascal is made to be compilable…
In our previous episode, peter green said:
I have long understood that each stable release of freepascal is made to be compilable with the previous release but at a quick glance at the website I couldn't seem to spot any official statement of this.
Mostly correct, except for new targets/architectures. There the order is "starting compiler last release" -> "current version on supported target/arch" -> "crosscompile current version to new target/arch"
Is it indeed "supported" to compile a stable release with the previous stable release and if so is there a statement of this anywhere.
Yes.
I ask because i've been helping someone in debian who is trying to convince another compiler (for a totally different language) to adopt a similar policy.
There is some at
http://www.stack.nl/~marcov/buildfaq/#toc-Section-2.2
I guess the reasons why are a bit thin though in the buildfaq.
The main reason is binary version management, a random version only has to be compilable by itself and the last release before it. A random version being any distinct SVN version (fixes, trunk or a new release tag).