In the early 2000's I worked on some Fortran-77 code that we were parallelizing (As a babe, this was my first experience with Fortran-77). We were having strange bugs because it seemed like some MPI calls were just not happening and others were. There were no compiler errors or warnings. Everything was supposedly fine. However, certain properties were propagating across grid elements just fine and others weren't. e.g. A particle might cross a vertical boundary and be just fine in the next CPU's element, but if it crossed a horizontal boundary it would lose it's horizontal momentum but keep everything else.
Eventually we figured out that the lines that weren't working were just over 80 characters long and the ones that were working were under 80 characters long. The compiler we were using was adhering to the Fortran-77 spec of a maximum of 80 characters per line (because that's what would fit on a punch-card), but it was just ripping those lines out and ignoring them without raising any kind of warning. If a line of code was over 80 characters long, it was just silently ignored. All we had to do was split the lines over 80 characters up and that particular bug was squished.
In my humble opinion, if a language can trip you up because of how many characters fit on a punch-card, it qualifies as "old school".
Eventually we figured out that the lines that weren't working were just over 80 characters long and the ones that were working were under 80 characters long. The compiler we were using was adhering to the Fortran-77 spec of a maximum of 80 characters per line (because that's what would fit on a punch-card), but it was just ripping those lines out and ignoring them without raising any kind of warning. If a line of code was over 80 characters long, it was just silently ignored. All we had to do was split the lines over 80 characters up and that particular bug was squished.
In my humble opinion, if a language can trip you up because of how many characters fit on a punch-card, it qualifies as "old school".