"Andre' Poenitz" <[EMAIL PROTECTED]> writes:
| Since it's public now, I might as well comment on it a bit ;-)
But your patch is _bad_
gcc 2.7.x has support for foreach and a lot of the other things in
algorithm (perhaps that is only with x >= 3).
I am not going to step back into the C++ stone age.
| Generally I think it would be a good idea if LyX could be compiled
| with gcc-2.7.2 since this seems to be some kind of de-facto
| standard.
Only generally... In linux it is a de-facto standard to compile C
progs and/or the _kernel_ with gcc-2.7.x, _but_ to compile C++ with egcs.
| The current development version does not compile, since gcc-2.7.2 does
| not support a couple of features:
|
| 1. Name spaces (or at least std::) are broken. The only code that uses it
| is the encoding stuff. It compiles quite well after a
|
| perl -pi -e "s/std:://g"
|
| in this dir. I don't have a clue if removing this breaks something (I
| doubt it), but there should be a way using #defines do hide this from
| gcc-2.7
This might be the only valid point in your patch.
| 2. The <algorithm> stuff used in DepTable does not work. I don't think
| it is really necessary there anyway and a simple loop will probably
| suffice.
Bogus, I have used this _a lot_ with gcc-2.7.x.
| 3. Problems with operator declarations in wrap_string. There is a
| constructor of a wrap_string out of " wrap_string<S>::value_type const *"
| anyway, so the "offending" operators could be removed. This could mean
| that an additional intermediate object is created but as I told JM
| already I am fairly sure that this won't occur when using -O2.
I developed wrapstring on Gcc 2.7.x
| 4. --with-included-string probably has to be used when.
|
| So I don't think it's too much effort to stay gcc-2.7 compatible.
I do.
You (we) just has to face it sometime: Gcc-2.7.x is not nice when it
comes to C++, more standard conforming compilers should be used. AND
several of the things you change in your patch *do* work with
gcc-2.7.x.
Lgb
