Saltar al contenido ForoHackers

legacy_cobol_whisperer

Se unió el 9 diciembre 2023

Mensajes en el foro

  1. Respondió en On-call from home: keeping the pager from eating the house

    The handoff note is the modern form of the operator's shift log, which we kept on paper in the machine room. Same content, same reason: the person who comes in cold should not have to guess what happ…

  2. Inició Replacing a legacy system without a big bang: the strangler pattern

    Every "rewrite from scratch" proposal I have seen has the same shape: eighteen months of work, a cutover weekend, and a list of behaviours nobody knew the old system had. The old system carried years…

  3. Respondió en New here: Windows admin, mostly patching

    The "backwards-compatibility promise from a very long time ago" is the main storyline of my career too. If you ever write up one of those, I would read it.

  4. Respondió en Qué pregunto yo en una entrevista a quien viene a mantener código ajeno

    Number 2 is excellent and I steal it shamelessly. I add a variant: I hand over a nightly job log with one failed step and ask "what do you check first?". The people who say "restart it" and the peopl…

  5. Respondió en REDEFINES: la unión de COBOL (y la trampa de sus 8 bytes)

    For the English readers: this is exactly the same idea as a tagged union, minus the tag. In my experience the real production bugs were the redefined areas where the "tag" lived in a different field,…

  6. Respondió en Constant-time comparison: why memcmp is the wrong tool for secrets

    On the mainframe side, we compared PINs and passwords with a routine that always walked the full field. Nobody called it "constant time" then. It was just "the check that does not get clever".

  7. Respondió en How ECC memory actually corrects one bit and notices two (SECDED)

    On mainframes this was normal decades ago. I remember the console logging "corrected error, DIMM 3" for weeks, and the engineer swapping the board during the quiet window without anyone noticing. Cor…

  8. Respondió en Characterization tests: how to touch code nobody understands

    Inject it, duck. Pass the date in as a parameter, default to the real clock only at the outermost layer. Then the test runs "on" the same day forever.

  9. Respondió en Characterization tests: how to touch code nobody understands

    Yes, and that is a social problem more than a technical one. The rule on my team was: the golden update goes in its own commit, with a message explaining what changed and why.

  10. Inició Characterization tests: how to touch code nobody understands

    You inherit a program that works, nobody knows exactly how, and you have to change it. Writing "correct" tests is impossible because you do not know what correct is. Write **characterization tests** …

  11. Respondió en "Exactly-once delivery" is at-least-once plus a dedup you must design

    On the mainframe we called it a restart key. Same idea, older name: every input record carries a unique key and the update step refuses a key it has seen. Every decade rediscovers a good idea and giv…

  12. Respondió en A 20-line TLA+ spec finds the lost update before you write the code

    "Show me what happens if the machine loses power between these two lines" is the same exercise on paper. The spec just does it for you faster and without feelings.

  13. Respondió en NOT IN with a NULL silently returns zero rows

    COBOL has no NULL in the language itself, so we had sentinel values: spaces in a field, `99999999` in a date. Embedded SQL brought NULL in through the back door via indicator variables. The first tim…

  14. Respondió en Two-phase commit: all the commitment issues, none of the closure

    On the mainframe side, two-phase commit has been in production for decades between the transaction monitor and the database. It works because both live in the same machine room under the same operati…

  15. Respondió en I wanted to rename a host. Day four.

    On the mainframe we kept a naming document: every system name, every alias, who depends on it. Boring, and exactly what you want on day four.

  16. Respondió en A restore drill that takes thirty minutes (and what it finds)

    On the mainframe the restore drill was a scheduled event with a name: "disaster recovery test", twice a year. The first one always failed. That was the point. The third one was quiet. Be patient with…

  17. Respondió en Before you design the protocol: write down the failure model

    On mainframes we wrote the failure model into the operating procedure: "if the machine loses power between these two steps, here is the recovery". It took the form of a runbook rather than a theorem,…

  18. Respondió en Restartable batch jobs: how to survive a failure at 3 a.m.

    It is, and that is why I always ask in design reviews: "show me what happens if the machine loses power between these two lines." People go quiet, then draw a better diagram.

  19. Inició Restartable batch jobs: how to survive a failure at 3 a.m.

    Mainframe batch has a lot to teach modern pipelines. The central idea: a job will fail halfway, so design for the restart, not for the success. Pattern I keep reusing: 1. **A restart table** with j…