At some point I have realised, that just letting cJSON be the horror of a codebase it was before I started maintaining it would have been the best thing to do, but that's what you call hindsight.
At least now I know.
One reason I burned out was because I felt horrified by how many people are using cJSON and felt a responsibility to make it better and safer for all of them, even though I myself haven't ever been really using it. The combination of this mounting pressure plus the the impossible fight against semantic versioning and memory unsafety gave me the rest.
By improving it, but being unable to completely "fix" it, I probably led more people to use it, making the situation even worse than it was before.
At some point I have realised, that just letting cJSON be the horror of a codebase it was before I started maintaining it would have been the best thing to do, but that's what you call hindsight.
At least now I know.
One reason I burned out was because I felt horrified by how many people are using cJSON and felt a responsibility to make it better and safer for all of them, even though I myself haven't ever been really using it. The combination of this mounting pressure plus the the impossible fight against semantic versioning and memory unsafety gave me the rest.
By improving it, but being unable to completely "fix" it, I probably led more people to use it, making the situation even worse than it was before.