|
necrotic posted:same as xml or a dumb soap-as-json. Encoding types and trusting them by default is dumb as poo poo anywhere. xml doesn't specify, as part of the standard spec, a way to embed arbitrary serialised objects in any xml file. you can't deserialise a random object by accident if you use xml: you know that the only things a file contains are the specific things your schema is expecting. yaml does define embedding arbitrary objects, and there have been plenty of rce vulnerabilities that resulted directly from that unspeakably dumb design flaw
|
# ? Aug 24, 2017 08:20 |
|
|
# ? May 25, 2024 01:45 |
|
JewKiller 3000 posted:actually, xml > csv > json > yaml deuteroshaggar was right
|
# ? Aug 24, 2017 08:21 |
|
lol if your ini files don’t live in S:
|
# ? Aug 24, 2017 08:22 |
|
fart simpson posted:old man yells at cloud? I'm basically the guy in that cartoon saying "you kids and your cloudy computers and googly maps" to a coworker responding "steve you're like 3 years older than me" edit: it's me Powerful Two-Hander fucked around with this message at 10:02 on Aug 24, 2017 |
# ? Aug 24, 2017 09:14 |
|
why are mysql error messages so crap. naming the column "to" isnt allowed ok fine but frikkin tell me that when i gently caress it up thanks
|
# ? Aug 24, 2017 10:58 |
|
i don't know what magical land of unicorns and roses y'all live in that has schemas to validate xml against, but i can assure you that in the real world the xml documents you receive will be just as arbitrary as some random yaml file
|
# ? Aug 24, 2017 11:01 |
|
i quite like protocol buffers because they're basically impossible to create by hand, so there's always a definition for them somewhere that outlines the boundaries of the api. it's not always a comprehensive description of what is actually expected and returned in various cases, but it is at least a workable baseline.
|
# ? Aug 24, 2017 11:16 |
|
redleader posted:i don't know what magical land of unicorns and roses y'all live in that has schemas to validate xml against, but i can assure you that in the real world the xml documents you receive will be just as arbitrary as some random yaml file I once sent someone a schema that was just <some wrapper poo poo> <xs:any>
|
# ? Aug 24, 2017 11:42 |
|
don't get me wrong, i *like* xml (and its bastard children xpath and xslt) i just don't think most people care about ensuring their input is well-defined - see e.g. json, the proliferation of dynamic langs, ... it is nice being able to wave a mutually-agreed schema in someone's face when their poo poo is malformed
|
# ? Aug 24, 2017 12:06 |
|
it's crunch time on a fantastically critical but poorly-managed project. as is traditional, i'm rubber stamping some truly awful code through in reviews (when we actually do reviews), and writing some utter poo poo that mostly works in many cases - unless you hit this or that edge condition, or an error happens this is complicated by another guy, who keeps dropping half or mostly broken code on me, apparently under the impression that no syntax errors = good to go
|
# ? Aug 24, 2017 12:11 |
|
JewKiller 3000 posted:actually, xml > csv > json > yaml CSV is pretty much the worst data format possible. It's even more laxly specified than URI querystrings. This: code:
is technically one of:
which one it is depends on the parser implementation, since even the shittiest CSV RFC allows this ambiguity by design.
|
# ? Aug 24, 2017 13:44 |
|
tab separated value is sort of ok
|
# ? Aug 24, 2017 13:58 |
|
Sapozhnik posted:tab separated value is sort of ok 0x1e is the only correct separator value
|
# ? Aug 24, 2017 14:01 |
|
Powerful Two-Hander posted:I once sent someone a schema that was just thats basically what .net does for a lot of its poo poo. i was tasked for writing a client that pulled xml from some location and it was all serialized DataSet objects and the guy who wrote it originally got made at me because "its just a STANDARD xml dataset object"
|
# ? Aug 24, 2017 14:02 |
|
that was the only time i told someone to "suck my dick" in my professional career
|
# ? Aug 24, 2017 14:05 |
|
CRIP EATIN BREAD posted:thats basically what .net does for a lot of its poo poo. one of the EHRs I integrate w/ is like that and it sucks. Their "API" is basically a service that returns serialized datasets which means its totally up to whoever wrote the sql as to what column names and types will be. there is pretty much no naming consistency across all the different methods. its horrible.
|
# ? Aug 24, 2017 14:43 |
|
redleader posted:it's crunch time on a fantastically critical but poorly-managed project. as is traditional, i'm rubber stamping some truly awful code through in reviews (when we actually do reviews), and writing some utter poo poo that mostly works in many cases - unless you hit this or that edge condition, or an error happens Spoiler alert: you will regret this lol
|
# ? Aug 24, 2017 14:43 |
|
i'll also regret being out of work because we didn't ship a product before the money people lost patience life is full of regrets
|
# ? Aug 24, 2017 15:00 |
|
i think it was nbsd who said that a hobby becomes a job the moment somebody else imposes their goals and/or their timelines on you
|
# ? Aug 24, 2017 15:01 |
|
Shaggar posted:one of the EHRs I integrate w/ is like that and it sucks. Their "API" is basically a service that returns serialized datasets which means its totally up to whoever wrote the sql as to what column names and types will be. there is pretty much no naming consistency across all the different methods. its horrible. i integrate with an api that returns a string of encoded xml wrapped inside of real xml. i wonder about people sometimes.
|
# ? Aug 24, 2017 15:24 |
|
Soricidus posted:xml doesn't specify, as part of the standard spec, a way to embed arbitrary serialised objects in any xml file. you can't deserialise a random object by accident if you use xml: you know that the only things a file contains are the specific things your schema is expecting. huh. i'd never poked at the spec but always assumed the dumb type poo poo was a layer on top (like with xml).
|
# ? Aug 24, 2017 16:12 |
vscode is now x64, and continues to own surprisingly much. visual studio is slowly growing on me too, even though i miss intellij
|
|
# ? Aug 24, 2017 17:02 |
|
Shaggar posted:swashbuckle works great. also web service integration works great in vs so idk what you're on about just spent this morning adding swashbuckle to my api and documenting it can confirm
|
# ? Aug 24, 2017 18:18 |
|
cinci zoo sniper posted:vscode is now x64, and continues to own surprisingly much. visual studio is slowly growing on me too, even though i miss intellij the one thing i miss from an ide is suggested fixes for problems. otherwise it's good though
|
# ? Aug 24, 2017 18:30 |
carry on then posted:the one thing i miss from an ide is suggested fixes for problems. otherwise it's good though i don't use it for actual deving, but its nice to janitor some textable files up
|
|
# ? Aug 24, 2017 18:39 |
|
jony neuemonic posted:i integrate with an api that returns a string of encoded xml wrapped inside of real xml. i wonder about people sometimes. an offshore team's idea of a message format is html scraped off a page, wrapped in cdata and bundled into an xml blast that just has a node named "data" also they can't send you just messages for your system because they don't know what consumer a message is for (because it's in the scraped html they can't parse properly themselves and isn't consistent between consumers because every page is some bespoke poo poo with its own mess of spelling errors) so you are expected to read every message off the bus and parse the whole thing to check if it's relevant i told them that their solution was terrible and refused to use it
|
# ? Aug 24, 2017 19:49 |
|
InfrastructureWeek posted:Spoiler alert: you will regret this lol oh, i know this very well, especially since there's a decent chance i'll end up maintaining the whole mess. but, Sapozhnik posted:i'll also regret being out of work because we didn't ship a product before the money people lost patience
|
# ? Aug 24, 2017 21:02 |
|
Powerful Two-Hander posted:an offshore team's idea of a message format is html scraped off a page, wrapped in cdata and bundled into an xml blast that just has a node named "data" also they can't send you just messages for your system because they don't know what consumer a message is for (because it's in the scraped html they can't parse properly themselves and isn't consistent between consumers because every page is some bespoke poo poo with its own mess of spelling errors) so you are expected to read every message off the bus and parse the whole thing to check if it's relevant you win.
|
# ? Aug 24, 2017 21:13 |
|
cinci zoo sniper posted:vscode is now x64, and continues to own surprisingly much. visual studio is slowly growing on me too, even though i miss intellij i just wish visual studio wasn't so retarded when it comes to handling folders. like it has no concept at all what a "folder" is and just uses its own dumb filter thing
|
# ? Aug 24, 2017 22:45 |
|
HoboMan posted:just spent this morning adding swashbuckle to my api and documenting it its great cause it pulls the documentation from the existing comments in ur code.
|
# ? Aug 24, 2017 22:58 |
|
anyone here use ROS? it's supposed to be good for controlling robot stuff i guess. i used a package where one of the services serialized a dense voxel volume and sent it out as an int array, which would occasionally smack into ROS's hardcoded 1 GB message size limit.
|
# ? Aug 24, 2017 23:15 |
|
Shaggar posted:its great cause it pulls the documentation from the existing comments in ur code. oh man I am going to set this up so that someone has to deal with my comments of "todo: wtf" when they go looking for docs
|
# ? Aug 24, 2017 23:28 |
|
Ok so say I've done ten checkins to my branch on TFS/TFVC, to accomplish one or two overall changes or action items or whatever you wanna call them I merge from main to my branch to get any changes, fix conflicts, and do one more checkin, then I merge from my branch to main and checkin is there any way for my checkin comment on that last merge to be anything other than "merge from dev-jrn"? like a trail to the changesets that contributed to the merge? i don't know wtf i'm doing in this UI, too used to git log Ciaphas fucked around with this message at 23:31 on Aug 24, 2017 |
# ? Aug 24, 2017 23:28 |
|
redleader posted:i don't know what magical land of unicorns and roses y'all live in that has schemas to validate xml against, but i can assure you that in the real world the xml documents you receive will be just as arbitrary as some random yaml file what's best are XML files that are obviously composed by string manipulation that wind up using invalid syntax for things like tag names and then get deployed to production that way with like a three month turnaround for any changes
|
# ? Aug 24, 2017 23:46 |
|
Powerful Two-Hander posted:an offshore team's idea of a message format is html scraped off a page, wrapped in cdata and bundled into an xml blast that just has a node named "data" also they can't send you just messages for your system because they don't know what consumer a message is for (because it's in the scraped html they can't parse properly themselves and isn't consistent between consumers because every page is some bespoke poo poo with its own mess of spelling errors) so you are expected to read every message off the bus and parse the whole thing to check if it's relevant make them write a regex to parse the HTML
|
# ? Aug 24, 2017 23:50 |
|
Powerful Two-Hander posted:oh man I am going to set this up so that someone has to deal with my comments of "todo: wtf" when they go looking for docs does this not just happen automatically in your IDE?
|
# ? Aug 24, 2017 23:51 |
|
Powerful Two-Hander posted:an offshore team's idea of a message format is html scraped off a page, wrapped in cdata and bundled into an xml blast that just has a node named "data" also they can't send you just messages for your system because they don't know what consumer a message is for (because it's in the scraped html they can't parse properly themselves and isn't consistent between consumers because every page is some bespoke poo poo with its own mess of spelling errors) so you are expected to read every message off the bus and parse the whole thing to check if it's relevant ......what
|
# ? Aug 24, 2017 23:52 |
|
Glorgnole posted:anyone here use ROS? it's supposed to be good for controlling robot stuff i guess. haven't used it myself but I know a lot of people who like it but idk it's an os so it's probably a pos
|
# ? Aug 25, 2017 00:00 |
|
Bloody posted:haven't used it myself but I know a lot of people who like it but idk it's an os so it's probably a pos tinkered with it a bit with turtlebots and hooboy
|
# ? Aug 25, 2017 00:05 |
|
|
# ? May 25, 2024 01:45 |
|
lancemantis posted:......what I'm completely serious. the application it's from is basically a form for requesting access to other apps but rather than have, idk, a generic data model in which to fit collections of possible app roles they have html stored in a database that the vomit onto the page per application and then barf the whole lot back at the backend on submit and do.... Something but mostly just seem to store the entire thing as text. I think that when you view a request it's actually just dumping that html blob back into the ui each time. this means there's zero consistency in the data that's actually stored for anything it's insane eschaton posted:make them write a regex to parse the HTML posted it before but I have seen someone write their own xml parser using regex in c# because..... gently caress I have no idea
|
# ? Aug 25, 2017 00:06 |