This is a good question and it is good that you are reaching out for advice. I would like to begin by saying this: every decision you make when it comes to film development is based on how important your photos are to you. There are many ways to cut corners, but invariably they all lead to inferior quality.
First, dividing any powder formula is a big no-no. There is no way to guarantee that the powder will pour out in equal proportions. Don't do it.
Second, don't use Diafine. Popular with many amateurs, the formula makes many claims, including speed increasing for low light, superior tonality ... test the developer against D-11, and you will see that the claims are false. While many photographers swear by it, and I include the formula in
The Darkroom Cookbook, there is a reason that photographers such as Ansel Adams, Wynn Bullock, Ralph Gibson, Imogene Cunningham, Mary Ellen Mark,
et al, never used it. At least not a second time.
Third, why mix a gallon of any developer? Purchase a 1 liter bottle of premixed Ilford DD-X, D-76/ID-11, Rodinal, Kodak HC-110, or other formula and dilute it as needed for use. You can increase the shelf-life of any developer if you decant it out of the plastic bottle into an amber or brown 1-liter bottle and store it in a cool, dark place such as under your kitchen or darkroom sink.
Finally, no developer or fixer lasts forever. If you don't use it before it goes bad,
c'est la vie. Don't try to "poor boy" it when it comes to the quality of your work.
BTW, the one developer that appears to last forever is Rodinal. I have a bottle that is more than 20 years old, and it is still developing film. (Why do I have a 20-year-old bottle of Rodinal? Because it got misplaced in a storage box on my move from Colorado to Oregon in 2006 and when I found it recently, I developed a roll of film, and the negatives came out as good, some might say better, than those developed in fresh Rodinal. I intend to develop another roll every few years until either the developer goes bad or I do; whichever comes first.)