|
that sounds like code that academics who don't give a poo poo about code write globals into parameterless functions and unnecessary, repeated deserialization are two specific calling cards i've seen
|
#
?
Mar 28, 2019 03:49
|
|
|
|
| # ? Apr 25, 2024 22:18 |
|
|
Doom Mathematic posted:People do the best work they can this is YOSPOS you don’t need to keep up appearances you can say what you really know deep down: people are lazy idiots
|
|
#
?
Mar 28, 2019 03:51
|
|
|
raminasi posted:repeated deserialization "this time will be different"
|
|
#
?
Mar 28, 2019 04:32
|
|
|
pokeyman posted:"this time will be different" what if the deserialization depends on global state, now you need it in the inner loop
|
|
#
?
Mar 28, 2019 04:56
|
|
|
raminasi posted:that sounds like code that academics who don't give a poo poo about code write the repeated deserialisation was entirely my fault. I called a getButt() function that I assumed would be a trivial getter but lol nope it parses the butt out of a string every time. (I also wrote that but a long time ago) the other bits, I don’t know the programmer personally but yeah it wouldn’t surprise me if they had an academic background
|
|
#
?
Mar 28, 2019 10:52
|
|
|
c tp s: just blew the mssql stack with a non-recursive query
|
|
#
?
Mar 28, 2019 11:02
|
|
|
eschaton posted:new front page column? seems more like a discarded weekend web idea tbh if only i werent so anal about tiny details that nobody in their right mind would notice
|
|
#
?
Mar 28, 2019 12:45
|
|
|
c tp s: finished auditing a database. my favourite part in this one was, i think, either the fact that the reworked “super good” entity B model has literally no way to link records to entity A that causes them, or the one table with metric ton of records that have modification_date < creation_date- but not all, so it’s clearly not a mapping error
|
|
#
?
Mar 28, 2019 15:57
|
|
|
or the one lonely record with NULL creation_date and non-null modification_date. i think our etl guy will have a stroke when i send the full report in
|
|
#
?
Mar 28, 2019 15:59
|
|
|
CPColin posted:I wrote a BBCode parser for Experts Exchange's special unicorn of a BBCode flavor and, thanks to a long request notification service that I also wrote, discovered that mismatched italics tags exhibited O(n²) or O(n³) behavior as it repeatedly scanned the input back and forth, desperate to find a matching [/i] tag. That was a fun one to hotfix. counterpoint: SA is still alive, so that’s a pretty low bar
|
|
#
?
Mar 28, 2019 16:39
|
|
|
NihilCredo posted:c tp s: just blew the mssql stack with a non-recursive query 
|
|
#
?
Mar 28, 2019 16:55
|
|
|
exotic unicode whitespace: there's always more of it and users will always find a way to insert it into your database today i discovered the existence of U+E0020 TAG SPACE, which apparently breaks stuff in safari i had no idea what a TAG SPACE was, but some googling reveals a presumably well-intentioned late 90's attempt to add in-band language metadata, but i have no idea how the user managed to input it. copy-paste from some word processor or other is our current best guess.
|
|
#
?
Mar 28, 2019 17:22
|
|
|
I couldn't use the proper indigenous name for one of our on-campus housing projects, yakʔitʸutʸu, because, while the web application was displaying the non-ASCII characters properly, something in the pipeline was translating them into question marks before they hit the database, where they then triggered validation errors because "yak?it?ut?u" was not a valid option.
|
|
#
?
Mar 28, 2019 17:33
|
|
|
ctps: updated to intellij 2019.1 and now my code won't build. everything works just fine if i run mvn manually in a terminal but intellij is all "internal java compiler error"
|
|
#
?
Mar 28, 2019 22:21
|
|
|
That sounds great I'm gonna update right now and relax for the rest of the day! lol the update window popped up over my browser twice while I was typing this and one of my spacebar hits threatened to cancel it
|
|
#
?
Mar 28, 2019 22:28
|
|
|
CPColin posted:I couldn't use the proper indigenous name for one of our on-campus housing projects, yakʔitʸutʸu, because, while the web application was displaying the non-ASCII characters properly, something in the pipeline was translating them into question marks before they hit the database, where they then triggered validation errors because "yak?it?ut?u" was not a valid option. Ahh stream encoding from unknown sources. "The right way to handle that is to simply reject everything that isn't ASCII, right?" (sarcasm here can not be overestimated)
|
|
#
?
Mar 28, 2019 23:00
|
|
|
CPColin posted:I couldn't use the proper indigenous name for one of our on-campus housing projects, yakʔitʸutʸu, because, while the web application was displaying the non-ASCII characters properly, something in the pipeline was translating them into question marks before they hit the database, where they then triggered validation errors because "yak?it?ut?u" was not a valid option. correct me if i'm wrong, but i have a terrible feeling "not a valid option" has a drastically different meaning from "not a valid input"
|
|
#
?
Mar 28, 2019 23:35
|
|
|
like, i'm imagining that not only did the app mangle the name, but it broke apart the input at the ?'s and treated the result as multiple parameters
|
|
#
?
Mar 28, 2019 23:41
|
|
|
Nahh, there was a <select> element showing the various student housing communities we have and a form validator that returned true if the submitted value matched one of the ones we expected. Since "yak?it?ut?u" didn't match "yakʔitʸutʸu", it was returning false, but only after the bad value had made it into the database and the user (me, testing, fortunately) was trying to proceed to the next page of the flow.
|
|
#
?
Mar 28, 2019 23:56
|
|
|
CPColin posted:Nahh, there was a <select> element showing the various student housing communities we have and a form validator that returned true if the submitted value matched one of the ones we expected. Since "yak?it?ut?u" didn't match "yakʔitʸutʸu", it was returning false, but only after the bad value had made it into the database and the user (me, testing, fortunately) was trying to proceed to the next page of the flow. solution: use the value as a glob and succeed if that gives a unique answer
|
|
#
?
Mar 29, 2019 00:13
|
|
|
Soricidus posted:ctps: updated to intellij 2019.1 and now my code won't build. everything works just fine if i run mvn manually in a terminal but intellij is all "internal java compiler error" I've had this before, doing a clean reinstall fixed it for me.
|
|
#
?
Mar 29, 2019 01:11
|
|
|
Soricidus posted:solution: use the value as a glob and succeed if that gives a unique answer pull request accepted
|
|
#
?
Mar 29, 2019 03:21
|
|
|
Krankenstyle posted:pull request accepted
|
|
#
?
Mar 29, 2019 03:58
|
|
|
Fatty Crabcakes posted:You are just like ur mom wontfix, works as intended
|
|
#
?
Mar 29, 2019 04:11
|
|
|
Soricidus posted:ctps: updated to intellij 2019.1 and now my code won't build. everything works just fine if i run mvn manually in a terminal but intellij is all "internal java compiler error" IntelliJ 2019.1 is broken on our poo poo too, something about a gradle hack that IntelliJ has technically never supported but started enforcing this version. which blows because 2019.1 has some cool poo poo in it. but on the other hand this is getting some folks to get around to making the build less hacky
|
|
#
?
Mar 29, 2019 04:42
|
|
|
I love me some IntelliJ but it's a pretty glitchy program. after every update the first project I open will open a window but then just sit there. restarting IntelliJ solves that... 
|
|
#
?
Mar 29, 2019 08:03
|
|
|
Soricidus posted:lol fixing those two things literally cuts the runtime in half again on typical inputs. well that was a productive hour you're like a little baby, watch this: see, i was caching user results from the database in a hashmap, like this code:well, our development environment doesn't have many users, so when we read the erp ids from another endpoint and start iterating over them, the call hits the database almost every time. i profiled the code, found the above offending line of code, read the documentation of computeIfAbsent and now the lambda returns a static USER_NOT_FOUND constant object instead if a user is not found in the database. runtime went from 1 hour 40 minutes to 20 seconds. thankfully this wasn't a problem in our production environment because all the erp ids exist in the production database. this is some of the code produced during february month of crunch
|
|
#
?
Mar 29, 2019 10:00
|
|
|
I “fixed” intellij by repeatedly mashing the build project button and eventually it succeeded still the least bad ide tho, I don’t miss eclipse in the slightest
|
|
#
?
Mar 29, 2019 10:01
|
|
|
Wheany posted:runtime went from 1 hour 40 minutes to 20 seconds. 
|
|
#
?
Mar 29, 2019 10:21
|
|
|
Soricidus posted:I “fixed” intellij by repeatedly mashing the build project button and eventually it succeeded You know what I'd love? Sublime to emulate the plugin APIs of intellij and vscode so it can use their plugins
|
|
#
?
Mar 29, 2019 10:35
|
|
|
i started loving around with a toy autoencoder in my spare time at work because im writing tests and its boring af, and i think i'm actually being more productive now than i was before when i was only working on tests
|
|
#
?
Mar 29, 2019 17:28
|
|
|
https://www.theregister.co.uk/2019/03/28/hcsec_huawei_oversight_board_savaging_annual_report/ lomarf
|
|
#
?
Mar 29, 2019 17:49
|
|
|
cinci zoo sniper posted:https://www.theregister.co.uk/2019/03/28/hcsec_huawei_oversight_board_savaging_annual_report/ lomarf This is a fun part: quote:In the first version of the software, there were 70 full copies of 4 different OpenSSL versions, ranging from 0.9.8 to 1.0.2k (including one from a vendor SDK) with partial copies of 14 versions, ranging from 0.9.7d to 1.0.2k, those partial copies numbering 304. Fragments of 10 versions, ranging from 0.9.6 to 1.0.2k, were also found across the codebase, with these normally being small sets of files that had been copied to import some particular functionality.
|
|
#
?
Mar 29, 2019 17:57
|
|
|
Private Speech posted:This is a fun part: what about code:
|
|
#
?
Mar 29, 2019 18:11
|
|
|
Private Speech posted:This is a fun part: wow. sounds like we should recruit from huawei to get some fresh blood in the thread
|
|
#
?
Mar 29, 2019 18:23
|
|
|
Private Speech posted:This is a fun part: thats amazing cinci zoo sniper posted:what about 
|
|
#
?
Mar 29, 2019 18:29
|
|
|
Private Speech posted:This is a fun part: (Not to say that they weren't put up to it by the US government or that any other telecom companies are likely to be better.)
|
|
#
?
Mar 29, 2019 18:56
|
|
|
I bet the chinese backdoors are there and the uk government is just pretending it didn’t find them so it can exploit them too
|
|
#
?
Mar 29, 2019 20:16
|
|
|
Soricidus posted:I bet the chinese backdoors are there and the uk government is just pretending it didn’t find them so it can exploit them too Well for one thing it does say that they're not sure if the source code they have is actually all that gets compiled because they don't have everything they need to build it. On the other hand I'm sure they've kept some vulnerabilities back for themselves anyway.
|
|
#
?
Mar 29, 2019 21:10
|
|
|
|
| # ? Apr 25, 2024 22:18 |
|
|
runtime is now down to about 15% of what it was before I started fixing performance. possibly I should stop profiling now? but yesterday I thought 200ms was as fast as I could get and now it’s around 140ms and I can’t stop wondering whether 100ms is achievable, help
|
|
#
?
Mar 29, 2019 22:28
|
|