Fixpoint

2022-08-01

#jwrd Logs for Aug 2022

Filed under: #jwrd logs, Logs — Jacob Welsh @ 21:46
Day changed to 2022-08-01
[21:46] jfw: welcome back jwm!
[21:47] jfw: caai: sounds good.
[21:47] jfw: we'll want to make sure you have a newer Gales release to work from once we get started; there's minimal change in the overall process though.
Day changed to 2022-08-02
[16:30] dorion: I pushed the draft of the project plan for the tradebot.
[22:22] jfw: reading.
[22:43] jfw: why does backtesting logic differ between the two strategies?
[22:46] jfw: got a feeling the time estimates aren't allowing for the writing 'overhead' of the status reports & documentation
[22:46] jfw: or more generally the customer or even internal communications
[22:50] jfw: overall I like the document and think they'll appreciate the way it breaks it all down. there's some sentence and paragraph level issues especially in the intro that I could probably massage.
[22:53] jfw: ability to determine the cause of errors or unintended actions or inactions is important indeed as I mentioned; but how do you conclude that it's ~the~ ~main~ challenge?
[22:56] jfw: in other crm updates, it's "denny" isn't it? "deny" makes it just a tad confusing, lol
[23:05] dorion: <jfw> why does backtesting logic differ between the two strategies? -- hmm, perhaps an unfounded assumption.
[23:06] dorion: true about quick and dirty estimates not accounting for documentation, status reports and internal communication. I'll add.
[23:07] dorion: (not going to be able to add tonight though, about to go to a couple networking events)
[23:08] jfw: the backtesting piece I'm picturing as a generic test harness, plug in the data and the strategy function, it runs it for each candle and reports the result.
[23:09] jfw: perhaps that's a nice part of looking at it in functional terms like that; the bot doesn't "do" anything, it just answers for any given input with its intended action.
[23:09] jfw: the strategy part, at least.
[23:11] jfw: not sure if "strategy" is the best term, I recall running afoul of diana_coman for using it in that casual sense
[23:12] jfw: probably comes right from my "game theory" exposure though.
[23:13] jfw: off to make some dinner, will see if you add anything further then probably take a crack at some edits.
[23:15] dorion: "how do you conclude that it's ~the~ ~main~ challenge?" -- seems like at least the most important thing for us to nail down, apart perhaps from implementing the "strategy" as they specify, because it where adjustments need to be made. what, from your perspective, are the main challenges ?
[23:15] dorion: thanks for clarifying your picture of the test harness and agreed about functional perspective.
[23:17] dorion: re "deny" I thought I'd seen him write it as deny, but looking closer, you're right. wil change it back.
[23:18] jfw: I'd expect the main challenges as usual would be the parts we're weakest at, and thus least able to correctly anticipate.
Day changed to 2022-08-03
[00:02] jfw: I suppose part of it is that I was thinking of it as a CYA kind of thing - bot loses money, they come to us, we'll need to defend its moves. thus, more just something for us to be prepared for rather than some core aspect of the project.
[00:05] jfw: in a more positive light, it would be part of the larger trader's feedback process where they reflect on what happened and attempt to improve
[00:08] jfw: but in that sense, it's ultimately up to them to determine the causes of the trading successes & failures, while our role is more confined to ensuring the bot operates as intended.
[00:14] jfw: and in that regard, it's what motivated the idea of including tables to keep a full log of the bot's actions along with source code and configuration changes.
[04:20] jfw: dorion: oops, this was the public chan.
[16:18] dorion: dang, second time in a couple weeks I wrote to the wrong chat window. sorry.
Day changed to 2022-08-06
[14:25] caai: jfw: update: i have been reviewing the first advanced module training video, which has led me to read about multiple new terms that came up, a few being; gcc, gnu vs musl, c standard library, compiler vs assembler
[18:21] jfw: caai: so getting into some under-the-hood components basically; or for another analogy, moving from the wrench to the furnaces & factory machines by which the wrench is produced. closed-toe shoes & eye protection recommended!
[18:25] jfw: the training though will just involve pushing the buttons to make the machines go, not how to build new things with them, fascinating as that is to someone like me.
[18:27] jfw: caai: how goes down in Panama? were you affected by those "protests"?
Day changed to 2022-08-07
[15:38] caai: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4774 - i am correctly using the newer version now: gales-linux-22. should i leave the previous version in the /var/build directory or delete it?
[15:38] sourcerer: 2022-08-01 21:47:58 (#jwrd) jfw: we'll want to make sure you have a newer Gales release to work from once we get started; there's minimal change in the overall process though.
[15:43] caai: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4804 - yes, i just want to gain a basic understanding because it is rather fascinating; but no building new things for the time being.
[15:43] sourcerer: 2022-08-06 18:25:01 (#jwrd) jfw: the training though will just involve pushing the buttons to make the machines go, not how to build new things with them, fascinating as that is to someone like me.
[15:46] caai: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4805 - all is well in Panama. i have not been affected much by the protests; i was just not able to find certain fruits or vegetables for around a week.
[15:46] sourcerer: 2022-08-06 18:27:25 (#jwrd) jfw: caai: how goes down in Panama? were you affected by those "protests"?
[15:50] caai: It seems to me that the sentiment of the pueblo es correct, but the protests are misdirected. The government isn't in control of petroleum prices, and stopping importation of produce from the interior only hurts fellow Panamanians, not government officials.
[15:50] caai: The time to protest was 2 years ago. The thing to protest was restricting businesses from operating and misuse of billion-dollar loans that were given to the government for covid relief; in other words, demanding transparency about where governments funds were being spent. The place to protest was outside the asamblea and the home/properties purchased by the diputados with the pueblo's money.
[15:52] caai: It seems to be that 'the pueblo' wants some 'free things', since the government gets 'free things', which is a road to socialism. Although, I agree that the corruption is ridiculous and needs to stop, I don't think this is the way forward. Properly investing in better infrastructure, better schools, tuition to study abroad, etc. is the way forward.
[15:55] caai: due to the contamination concerns, do you suggest i do the build on the online thinkpad or the asus (gentoo) machine?
[17:51] jfw: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4806 - up to you; certainly can't hurt to keep old code around at least in archival form though.
[17:51] sourcerer: 2022-08-07 15:38:10 (#jwrd) caai: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4774 - i am correctly using the newer version now: gales-linux-22. should i leave the previous version in the /var/build directory or delete it?
[17:57] jfw: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4812 - just within my own panama experience, it was already established that 'el pueblo' (which seems to be used as a sentimental group term, something like 'the american people' over here) doesn't give a hoot about producers, figuring the state is the source of all resources, by that 'control de precios de la canasta basica de
[17:57] sourcerer: 2022-08-07 15:50:19 (#jwrd) caai: It seems to me that the sentiment of the pueblo es correct, but the protests are misdirected. The government isn't in control of petroleum prices, and stopping importation of produce from the interior only hurts fellow Panamanians, not government officials.
[17:57] jfw: alimentos' that Varela campaigned on IIRC.
[17:58] jfw: glad to hear this one was no more than a minor disruption though.
[18:02] jfw: I didn't get the sense that this was really some grassroots thing though, rather orchestrated by the red labor unions - the good folks who don't believe in garbage collection. perhaps it could be argued there's little distinction though since el pueblo puts up with it.
[18:07] jfw: as far as better infrastructure, at least one missing ingredient there is clarification & strengthening of property rights, in general and specifically as seen in land, so that private developers can actually take on large scale projects without expecting to be robbed within the decade.
[18:09] jfw: as far as free tuition to study abroad, to me that just reads as 'party abroad'.
[18:11] jfw: ie, money is rarely the real problem when it comes to lack of education, at least these days.
[18:18] jfw: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4815 - I wouldn't worry too much about contamination: there hasn't been seen such a thing as the immaculately conceived computer since the dawn of the microprocessor, at least outside of classroom toys; meanwhile, the gestation process is designed about as well as it can be to minimize such contamination (eg, cross compiling functions
[18:18] sourcerer: 2022-08-07 15:55:16 (#jwrd) caai: due to the contamination concerns, do you suggest i do the build on the online thinkpad or the asus (gentoo) machine?
[18:18] jfw: quite like the separation of bloodstreams between mother and fetus)
[18:19] jfw: however, I'd advise using the Gales machine for the sake of building strength in that environment.
[18:27] jfw: on what 'strengthening property rights' might look like, for one thing, if I owned a turnpike or railroad etc. and some "protestors" tried blockading and burning shit on my land, I'd be at liberty to expel them and shoot dead every last one that refused to comply.
Day changed to 2022-08-08
[14:47] dorion: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4830 -- incidentally, I was just reading http://trilema.com/2014/lets-pretend/#footnote_0_55902
[14:47] sourcerer: 2022-08-07 18:27:04 (#jwrd) jfw: on what 'strengthening property rights' might look like, for one thing, if I owned a turnpike or railroad etc. and some "protestors" tried blockading and burning shit on my land, I'd be at liberty to expel them and shoot dead every last one that refused to comply.
[21:59] caai: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4826 - i brought this point up because in the training you mentioned that if one targets the same machine he is running on it may pull in existing binaries. but then you went on to say that for said reason one should use a cross compiler. i assume this is the ISO C90 Complier (gcc 2.95+)?
[21:59] sourcerer: 2022-08-07 18:18:50 (#jwrd) jfw: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4815 - I wouldn't worry too much about contamination: there hasn't been seen such a thing as the immaculately conceived computer since the dawn of the microprocessor, at least outside of classroom toys; meanwhile, the gestation process is designed about as well as it can be to minimize such contamination (eg, cross compiling functions
Day changed to 2022-08-09
[02:44] jfw: caai: yes. c90 refers to the 1990 version of the C language standard, which (among other languages & variants) is implemented by gcc. gcc however can be configured as either a native or a cross compiler. the "bootstrap" process goes through a couple different stages, building up from that minimal requirement to a full native compiler & system binaries for any supported CPU architecture.
[02:46] jfw: gcc 2.95 is quite old and we haven't tested it, but that's what our gcc (4.7) cites as the minimum needed to build *it*. once we have 4.7, we're working from a known toolset.
[02:56] jfw: the point I was making which you mentioned, is that one might be tempted to "optimize" the process by saying hey, I'm on an Intel box and bootstrapping an OS for Intel boxes, so why bother with the extra complexities of cross compiling? Besides the loss of the ability to even target other architectures, the risk is that something within the large volume of automated configuration & build logic
[02:56] jfw: will then feel itself entitled to make similar "optimizations" and import artifacts from the host system (binaries, libraries or even header files) into the new one being built, and it might work well enough to go unnoticed.
[04:12] jfw: dorion: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4831 - good link and yeah, it was a thing, eg the notorious Pinkertons in their glory days. as to the article, I mostly recalled just the intro, though the topic of gold surely came up many times of which I'm mainly recalling http://trilema.com/2014/how-many-years-you-got/
[04:12] sourcerer: 2022-08-08 14:47:13 (#jwrd) dorion: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4830 -- incidentally, I was just reading http://trilema.com/2014/lets-pretend/#footnote_0_55902
[04:15] jfw: (especially gold as contrasted with bitcoin, of course)
Day changed to 2022-08-10
[14:08] caai: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4838 - i understand, well, enough to move forward at least :)
[14:08] sourcerer: 2022-08-09 02:56:55 (#jwrd) jfw: will then feel itself entitled to make similar "optimizations" and import artifacts from the host system (binaries, libraries or even header files) into the new one being built, and it might work well enough to go unnoticed.
[14:19] caai: i am on Step 1. Define a clean build environment. i see on my machine that '/var/build/gales' has the resultant folders/files from the previous build. would you suggest that i remove the contents of this directory 'rm -r * /var/build/gales' so as to start afresh?
[16:00] dorion: caai, up to you if/when you remove the old build env. you could move it as a first step, e.g. to /var/build/gales-2019 ; then you could remove it once the new build is confirmed working.
Day changed to 2022-08-11
[18:18] caai: dorion: simply renaming the old build env. sounds like a good plan. i will do that
[20:12] jfw: caai: moving it aside is a way to postpone a decision at least, which may be reasonable here since through further work you'll gain more insight into its value or lack thereof.
[20:15] jfw: from my perspective though it's a large pile of auto-generated bits (gigabytes in fact) with the source deliberately maintained elsewhere and any resulting artifacts of interest likewise fished out to a permanent home, so I wipe it when needed without thinking twice.
[20:17] jfw: what's the intent of the '*' in your example 'rm' command? because I don't think it means what you want it to mean
[20:19] jfw: or it does, but the way you use it will not have the desired effect but a decidedly undesired one
Day changed to 2022-08-12
[16:18] jfw: https://www.openwall.com/lists/oss-security/2022/08/09/1 - another zlib vulnerability is out; nobody seems to have looked into when it was introduced but the Gales zlib doesn't appear to contain the affected code at all.
[16:18] sourcerer: 2022-04-05 16:49:59 (#jwrd) jfw: oh hey, there's a zlib compressor bug making the rounds with a 17-year exposure window... and the Gales version just so happens to be unaffected.
[16:27] jfw: indeed, the code in question was introduced in zlib 1.2.2.1 from 2011, described as "Add inflateGetHeader() to retrieve gzip headers".
[19:01] caai: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4847 - these are valuable insights. i am considering renaming the previous build environment. as i progress with the new build, i can reference it to verify that i am on the right path.
[19:01] sourcerer: 2022-08-11 20:12:41 (#jwrd) jfw: caai: moving it aside is a way to postpone a decision at least, which may be reasonable here since through further work you'll gain more insight into its value or lack thereof.
[19:10] caai: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4847 - upon further scrutinization, i believe i see the error. 'rm -r * /var/build/gales-2019' would recursively delete all directories, subdirectories and files under the current workding directory AND under /var/build/gales-2019 as well. in other words, if i were in the root directory, it would be equivalent to 'rm -r /*' AND 'rm -r
[19:10] sourcerer: 2022-08-11 20:12:41 (#jwrd) jfw: caai: moving it aside is a way to postpone a decision at least, which may be reasonable here since through further work you'll gain more insight into its value or lack thereof.
[19:10] caai: /var/build/gales-2019/*'. is that correct?
[19:12] caai: the intended command would be 'rm -r /var/build/gales-2019/*', right?
[21:35] jfw: caai: exactly. (you wouldn't need the * either, to remove the whole thing)
Day changed to 2022-08-13
[01:02] caai: jfw: alright. understood!
Day changed to 2022-08-15
[13:18] caai: jfw: at the end of step 1, should the last 3 lines be:
[13:19] caai: BUILD=x86_64-gales-linux-gnu
[13:19] caai: TGT=x86_64-gales-linux-musl
[13:20] caai: J=j3
[13:20] caai: ?
[13:28] dorion: caai, that'll work. for BUILD and TGT they have to be different. J is the number of cores in your CPU+1.
[14:59] jfw: needs to be J=-j3
[15:00] jfw: (the - is included in the instruction)
[15:01] jfw: for BUILD I'd go with x86_64-unknown-linux-musl, to meet the "different" requirement without fibbing about libc vendor (the last component)
Day changed to 2022-08-16
[12:42] caai: got it. thanks
Day changed to 2022-08-17
[12:35] caai: jfw: i am now on step 2.1: when i issue 'mkdir build root', i receive the following message: 'mkdir: can't create directory 'build' : Permission denied', 'mkdir: can't create directory 'root' : Permission denied'
[13:10] caai: i believe /var/build/gales was created as 'root', 'su - build'
[17:51] jfw: caai: now that you're getting into the thick of it, might be a good time to schedule the next class session which is intended simply for continuing the build process.
[17:52] jfw: some evening later this week perhaps?
[17:53] jfw: as to permission problems, in general, either it means you're not running with the intended user ID, or the owner/group/mode of the relevant files aren't set correctly to allow access to the intended user.
[17:53] jfw: to check the first you can do 'whoami' (or 'id' for more detail)
[17:55] jfw: to check the second it's of course 'ls'; but for creating, deleting or renaming things, the relevant item being modified is the containing directory, so that's what you need write permission on.
[17:56] jfw: so check the working directory is as expected then check its permissions with 'ls -ld .'
Day changed to 2022-08-18
[13:08] caai: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4874 - sounds good. how about on Aug 22 or Aug 24 at 23:00 UTC?
[13:08] sourcerer: 2022-08-17 17:51:55 (#jwrd) jfw: caai: now that you're getting into the thick of it, might be a good time to schedule the next class session which is intended simply for continuing the build process.
[13:12] caai: http://fixpoint.welshcomputing.com/2022/jwrd-logs-for-Aug-2022/#4876 - got it! i see that i created the 'gales' directory as root rather than build, a rookie mistake. i removed it (it was empty) and created it as build
[13:12] sourcerer: 2022-08-17 17:53:34 (#jwrd) jfw: as to permission problems, in general, either it means you're not running with the intended user ID, or the owner/group/mode of the relevant files aren't set correctly to allow access to the intended user.
[15:16] jfw: caai: I suppose it's an extra level: not only do you have to keep track of where you are, but sometimes also of *who* you are. turns out we're actually training undercover agents here!
[15:17] jfw: Aug 22 is good.
Day changed to 2022-08-19
[16:14] caai: jfw: indeed, it is a multifacted environment! Aug 22 it is!

No Comments »

No comments yet.

RSS feed for comments on this post. TrackBack URL

Leave a comment

Powered by MP-WP. Copyright Jacob Welsh.