Hi, I write many source blocks that take a fairly long time to execute (several minutes). Sometimes I find myself wanting to do cosmetic changes to the code, that I know (for sure) do not change the output, after the first evaluation. Unfortunately, after having made these changes the cache obviously does not match the code anymore and I have to re-evaluate to update the cache. I know it goes against the strict 'reproducible research' idea but still I would like to be able to re-compute the cache without actually evaluating the block as it is not efficient to re-evaluate for the sole purpose of updating the cache.
What do you think? -- /Henning
