Fallback
-
We have insanely complex code running and well, our test coverage isn't great.
Sometimes, important issues are raised, which need to be solved within a few days. Obviously, it happens that the cure is worse than the disease, and that we have to fallback the code.This is OK (kinda).
What's not is that after code is loaded in production, it is put in the HEAD of our CVS (sigh). And if a fallback is needed, here are the kind of mail that we see:
Hello,
Currently, there is no procedure for head fallback.
The only way is to promote the previous code level for failing codes.All this because there were new files in this release and they don't know how to remove them...
CVS is a POS, but hell, the company has been using it for years and have faced this kind of situation quite regularly.... Get a clue people !
-
Rollback?
-
Holy shit... the goto statements are advancing... FALL BACK MEN!!
-
-
-
-
@morbiuswilters said:
@eViLegion said:
Holy shit... the goto statements are advancing... FALL BACK MEN!!
longjmp! longjmp!
Segmentation fault
-
@Ben L. said:
@morbiuswilters said:
@eViLegion said:
Holy shit... the goto statements are advancing... FALL BACK MEN!!
longjmp! longjmp!
Segmentation fault
I have a segfault handler which tells you to STFU.
-
-
@da Doctah said:
Yeah, but can it do this?
Man, Uri Geller has really let himself go..
-
In my previous job I met a guy who had been coding telecom equipment in assembly. He made large blobs of NOP instructions scattered around in between code. Why? Because he thought compile-time variables and labels were impacting performance -- so he calculated and hard coded the byte offset of each jump. Ergo, no updates could shuffle around the old code, so it had to be meticulously placed in a suitable pool of NOPs in the existing binary.
He was particularly proud of that, since he planned for the future and everything.
