Hello, the attacched patch fixes a compilation problem of uBLAS with GCC 3.4. The fix looks legit to me, but I cannot assure this won't break other compilers due to bugs. Maybe you can monitor closely the regression fixes after having applied this. OK to commit to HEAD? -- Giovanni Bajo begin 666 triangular_gcc34.txt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`` ` end