TJT
Mr. Poopybutthole
- 40,889
- 102,596
Multiple people are generally involved in the reviewing process. Dudes like this just want it to work, shit it out and move on.
IT/Software career thread: Invert binary trees for dollars.
- Thread starter Traak
- Start date
Multiple people are generally involved in the reviewing process. Dudes like this just want it to work, shit it out and move on.
- 16,366
- 7,363
I find myself programming "in C", as in variables towards the top of functions, more often than not to avoid scope vagueness.
Noodleface
A Mod Real Quick
- 37,961
- 14,508
I just refuse to approve code if I don't think it's up to snuff. I've done it many times. If the boss thinks it's ok then so be it, I let him approve it.
We also require every code review is approved by two reviewers and our git hooks won't let a check-in occur without a completed code review. There's ways around it of course, but everyone would see it.
I would just say "I don't approve" and if he's got an issue he can raise it to someone else. Having a single reviewer sounds terrible for anything even remotely complicated.
Something to keep in mind is I've been bitten by approving a code review and it being incorrect later on - like a field bug or whatever. Generally you look like an idiot because your name is attached to their code.
We also require every code review is approved by two reviewers and our git hooks won't let a check-in occur without a completed code review. There's ways around it of course, but everyone would see it.
I would just say "I don't approve" and if he's got an issue he can raise it to someone else. Having a single reviewer sounds terrible for anything even remotely complicated.
Something to keep in mind is I've been bitten by approving a code review and it being incorrect later on - like a field bug or whatever. Generally you look like an idiot because your name is attached to their code.
Vinen
God is dead
- 2,782
- 486
I suspect this next bonus round is going to blow based on both our companies earnings reports. Even for people who normally get 120% (I usually get 120%-130%) or so its not going to be that.
Mist
Eeyore Enthusiast
- 30,362
- 22,121
Being in telecom is funny right now.
Customers are sending a steady stream of tickets such as:
Provision 600 VPN phones.
Install softphones on every laptop in our company.
How to enable cellphone twinning for 1000+ people in bulk.
Audit license and trunk capacity for remote workers.
Etc.
Like every one of our customers is preparing to empty out their buildings and go into full Business Continuity mode.
Customers are sending a steady stream of tickets such as:
Provision 600 VPN phones.
Install softphones on every laptop in our company.
How to enable cellphone twinning for 1000+ people in bulk.
Audit license and trunk capacity for remote workers.
Etc.
Like every one of our customers is preparing to empty out their buildings and go into full Business Continuity mode.
Noodleface
A Mod Real Quick
- 37,961
- 14,508
Noodleface
A Mod Real Quick
- 37,961
- 14,508
Well I mean besides being CT... I kid. I went there over the summer to Mohegan sun.
My code is working but we keep finding corner cases. I'll own up to those, but to be fair we didn't do integration until today (code freeze was Friday)
My code is working but we keep finding corner cases. I'll own up to those, but to be fair we didn't do integration until today (code freeze was Friday)
Noodleface
A Mod Real Quick
- 37,961
- 14,508
Rabkorik
Silver Knight of the Realm
- 158
- 39
Speaking of Python, word got out that I have been doing some script work using Python (not my job responsibility) and now I'm being asked to write a script that adds 'Natual Language Searching' to one of our document repositories. I've started looking at the NLTK module but am unsure if that is the best place to start - any advice on how to begin or if it's even really feasible?
Asshat wormie
2023 Asshat Award Winner
- 16,820
- 30,963
I don't know what exactly you are trying to do but nltk is more of a teaching module. Consider spaCy instead.
Noodleface
A Mod Real Quick
- 37,961
- 14,508
Not too worried. I don't consider the bonus part of my salary, we end up just saving it or in the case of this year were paying off my car.
I'm still surprised they haven't closed us yet
Vinen
God is dead
- 2,782
- 486
Noodleface
A Mod Real Quick
- 37,961
- 14,508
Kharzette
Watcher of Overs
- 4,906
- 3,549
Are any of you amazing at complex git stuff? For a long time I've wanted to restore a very troubled repository of mine.
It started out life as a perforce repo, however all of that history is lost. In 2008 when it was imported into SVN, that's where the history begins.
Several years of SVN go by, and eventually sometime around 2014 it is converted into mercurial.
There are several goals I'd like to accomplish by bringing this into git:
Cull binaries: There are alot of meshes and textures and general crap that all needs to go for good. I think git-filter-repo will work well for this.
Cull sidetracks: Lots of experimental projects that never went anywhere
Fix moves/renames: A LOT of the time when a file was moved (say pulled out of a tool into a lib), the conversion destroyed the link between them changing the move to a delete + add new file.
Fix profane rants: I want to adjust some of the checkin notes to make it more suitable for githublandia.
Separate into submodules: Library code should eventually land in a library submodule. Same for tools and tests.
The tricky bit with that is alot of code started out in a game or a tool, then migrated out into a library. If I use filter repo to chop games, will the history only go back to the time when a file was moved out of a game project?
I've started doing this by advancing one commit at a time, and hand fixing up bad renames / moves. I set the date to the OG repository dates, however I later noticed git has 2 dates. Author and commit, no idea why. Author is correct but commit is not, but I think there may be automated ways to set all commit dates to author dates when I'm done.
It started out life as a perforce repo, however all of that history is lost. In 2008 when it was imported into SVN, that's where the history begins.
Several years of SVN go by, and eventually sometime around 2014 it is converted into mercurial.
There are several goals I'd like to accomplish by bringing this into git:
Cull binaries: There are alot of meshes and textures and general crap that all needs to go for good. I think git-filter-repo will work well for this.
Cull sidetracks: Lots of experimental projects that never went anywhere
Fix moves/renames: A LOT of the time when a file was moved (say pulled out of a tool into a lib), the conversion destroyed the link between them changing the move to a delete + add new file.
Fix profane rants: I want to adjust some of the checkin notes to make it more suitable for githublandia.
Separate into submodules: Library code should eventually land in a library submodule. Same for tools and tests.
The tricky bit with that is alot of code started out in a game or a tool, then migrated out into a library. If I use filter repo to chop games, will the history only go back to the time when a file was moved out of a game project?
I've started doing this by advancing one commit at a time, and hand fixing up bad renames / moves. I set the date to the OG repository dates, however I later noticed git has 2 dates. Author and commit, no idea why. Author is correct but commit is not, but I think there may be automated ways to set all commit dates to author dates when I'm done.
Noodleface
A Mod Real Quick
- 37,961
- 14,508
Had my yearly review today. All good things. They're challenging me to take more of a leadership role so I can move into tech lead eventually.
In my peer review comments every person said I'm reserved and they can't read me at all. That's the autism I think
In my peer review comments every person said I'm reserved and they can't read me at all. That's the autism I think