On 02/09/10 16:12, Hicham Mouline wrote: [snip]
I have another problem for now.
PP processing is starting to be slow, I need help to optimize it. Attached in the PP code.
Also, in the above, all the Ss need to be generated only once. However, in multi_array
, the N depends on the translation unit. I will have like 100 translation units, each with a different N. If I were to include the PP code to generate the Ss in each of the 100 trs, it would repeat the preprocessing uselessly. I need to run the first part of the PP just once, then the 2nd part of the PP that generates the multi_arrays for each of the 100 tr units.
Is this possible?
What about precompiled headers? Microsoft provides them: http://msdn.microsoft.com/en-us/library/szfdksca(VS.71).aspx as well as gcc: http://gcc.gnu.org/onlinedocs/gcc/Precompiled-Headers.html HTH. -Larry