Andre Poenitz <[EMAIL PROTECTED]> writes:

| You can't simply go out and tell people: Reinstall your system just 
| to use my program *and* expect that many people will do that.

The point then is "If you want to compile"

| Using bleeding edge techniques actually prevent an immediate
| widespread use of software. This is no help for any project.

It is not bleeding edge, it is an antiquated compiler.

| This holy quest for sstream and namepaces buys us currently nothing
| except a lot of trouble.

Ever tried using these features?
I buys a lot.

| People that might have done great work on
| LyX otherwise are fixing compatibility issues. That's a complete waste
| of resources. Patches that were intended to clean up and to make
| program flow clearer get cluttered by #ifdefs. That is stupid, too.

sstream is way better than stringstream and we want to use them if possible.
iostreams is way better than c-style io.

we don't require namespaces.
we don't require exceptions.
we don't require rtti

| It does not even buy us anything for the future, since we have to go
| through all the code again and remove all of these monstrosities.

grep -n HAVE_SSTREAM is not that hard.

| LyX is still a community project and I strongly feel that no single
| person's private feelings on what is cool and what is not shall
| dominate the community's ratio. No matter how much work gets done by
| this individuum.

So you want us to use gcc 2.7.x still? Don't take advantage of the C++
standard library, templates and such?

| 
| Oh yes. I mean it. And I am sober.

Not sober enough.
The C++ standard is now three years old (i.e. it is three years since
it was finalized), for gcc there has been, since 2.8.x, at least two
releases: egcs and gcc 2.95.2.

| A decent compromise a step away from the bleeding edge usually 
| leads to broader acceptance. 

Wake up! It is not bleeding edge.

| PS: Since we are at it: Unquoted eight-bit-characters in email
| headers are not RFC conform and thus unpolite. They hinder
| communication. To put it mildly.

RFCs can also be wrong. And you of all people should understand that
very well. Also most (all?) mailsevers of today are able to handle
8bit chars in headers. If not they have manually been configured to
not allow 8bit chars or they have not been updated in ages (and thats
lazy sysadmins). I just won't buy it. (nice switch from on topic to
person btw.)

| PPS: I think it's already Friday somewhere in the far east.

Probably not.

        Lgb

Reply via email to