tag:blogger.com,1999:blog-7094652.post4289862185865729339..comments2024-03-23T14:36:09.980+00:00Comments on Neil Mitchell's Blog (Haskell etc): Existential SerialisationNeil Mitchellhttp://www.blogger.com/profile/13084722756124486154noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-7094652.post-78983189224379207902017-09-20T21:11:26.265+01:002017-09-20T21:11:26.265+01:00nomeata: Thanks, an excellent observation - I'...nomeata: Thanks, an excellent observation - I've altered the text to include that as a disadvantage. I guess it depends how many changes you can make in practice - it's certainly going to make it difficult for the Shake use case.Neil Mitchellhttps://www.blogger.com/profile/13084722756124486154noreply@blogger.comtag:blogger.com,1999:blog-7094652.post-61778049742277896462017-09-18T14:20:06.747+01:002017-09-18T14:20:06.747+01:00Also keep in mind that the static pointer will not...Also keep in mind that the static pointer will not necessary be valid once you re-compile your binary. Maybe it is fine to have to throw away the shake data base when the build system changes, and maybe that is even a good idea to ensure consistency, but it can also be very annoying when developing a Shake program.nomeatahttps://www.blogger.com/profile/05986291481431063004noreply@blogger.com