Andre Poenitz <[EMAIL PROTECTED]> writes:
| We do not do anything fancy with templates that surpasses the requirements
| the STL has. So "being able to compile code using the STL" seems to be a
| sensible requirement for compilers in order to support them.
It is a lot of compilers that are not able to compile the C++ standard
library.
| Before you throw purists' arguments: There is currently no such thing as
| a standard conforming compiler. Not a single one. So all compilers are
| bad.
bad != bad
| Since we do not want do to the translation .C -> .o with paper and
| pencil we need to draw a line *somewhere*. You draw a very tight line
| around your current system configuration and I am saying that this area
| is to small for a widespread acceptance of LyX.
Huh? The line is far away from my system configuration.
Gcc 2.95.2 and Gcc 2.96 with libstdc++-v3
| > Let's say keeping 2.8.x for some time would be nice.
|
| If 2.8.x compiles STL, that's a very good compromise. I went the route
| 2.7 -> egcs -> 2.95, so I can't say for sure whether 2.8 suffices.
2.8 is just barely sufficient.
| A compiler's support for namespace should have no influence on whether
| It compiles LyX or not. We do not _need_ namespaces.
and we don't require them.
| Maybe it would be
| nice to have them in the future so we should be prepared for that time,
| but it's no requirement right now.
|
| Especially if it makes LyX less attractive.
I really don't understand what you are bitching about, we go to great
lengths to support as many compilers as feasible without sacrificing
too much.
| Andre'
|
| PS:
|
| > Lars> RFCs can also be wrong.
|
| Please try to explain to a mathematician (i.e. me in this case)
| how a _definition_ can be wrong. Granted, it might be useless,
| it might be awkward, but it can't be wrong. It's a definition.
is it?
I can define my way out of anything.
besides news has been able to cope with 8bit chars in headers for
ages, and IMO not allowing this in mail is a bug in the rfc
| Usually I am not especially interested in reading emails from people
| who stomp on good manners, but this case is somewhat special, since I
| respect the worker for the work's sake.
I am usually not especially interested in communicating with people
that hinder me from using my real name.
| And it is indeed strange: On the one side you despise compilers for not
| following the rules, on the other you are actively breaking rules.
| Does not make much sense to me.
Think about "why break the rules".
compilers -> have not catched up to the standard (rules)
mail headers -> excludes 8bit chars for no good reason
| PPS: It is Friday for several hundred millions of people...
Yes, now it is. :-)
Lgb
